An outlook on progress in polypropylene-based polymer technology

An outlook on progress in polypropylene-based polymer technology

Prog. Polym. Sci., Vol. 16, 303-329, 1991 Printed in Great Britain. All rights reserved. AN OUTLOOK POLYPROPYLENE-BASED 0079--6700/91 $0.00 + .50 © ...

1MB Sizes 2 Downloads 12 Views

Prog. Polym. Sci., Vol. 16, 303-329, 1991 Printed in Great Britain. All rights reserved.

AN OUTLOOK POLYPROPYLENE-BASED

0079--6700/91 $0.00 + .50 © 1991 Pergamon Press plc

ON PROGRESS IN POLYMER TECHNOLOGY

T. SIMONAZZI,* G. CECCHIN and S. MAZZULLO

Himont Italia, Centro Ricerche G. Natta, 44100 Ferrara, Italy

1. I N T R O D U C T I O N

Professor Giulio Natta first synthesized crystalline, structurally regular isotactic polypropylene in 1954 at Politecnico di Milano. Today, polypropylene has evolved from Natta's original polymer into a wide variety of polyolefinic materials having a broad range of properties. This achievement is the culmination of years of research focused on understanding the basic principles of initiator systems and polymerization mechanisms. Success in understanding these very complicated phenomena makes it possible to control the molecular structure and, consequently, the properties of the polymer. The knowledge and control of the polymerization mechanism also determines the process technology and its economics. This relationship between the nature of the initiator system, the polymerization process and the final properties of the polymers has been widely recognized by researchers and companies involved in polypropylene. The history of these developments is Well documented, N° starting from the original discovery made by Professor Natta and co-workers (foremost among whom was Professor Pino) to the most advanced current production processes. 2. I N I T I A T O R

SYSTEM

DEVELOPMENT

The present Ziegler-Natta initiator systems for propylene polymerization are able to achieve five major objectives. These are: (1) a polymerization activity so high that removal of initiator residues is not required while all necessary and desired product qualities are maintained; (2) the ability to produce a polymer with the necessary and desired molecular weight distribution; (3) the ability to prepare polymers with a controlled degree of stereoregularity, either maximum or minimum, and without requiring costly operations to remove undesired polymeric fractions; (4) the ability to obtain, in polymerization, a granular *Author to whom correspondence should be addressed. Editors' note: It was our editorial policy from the beginning, in putting together this issue, to refer to the metal complexes used in Ziegler-Natta polymerization uniformly as "initiators" or "initiator systems" rather than as "catalysts", and so the former term has been used in this paper and in the others in this issue. 303

304

T. SIMONAZZ! et al.

polymer with a controlled morphology, particle size distribution and compactness, so as to achieve the maximum savings, simplicity and versatility, both in terms of production process and product properties; and (5) the capability of producing heterophasic or, better, multi-phasic products by direct synthesis. Over the years our research has dedicated major effort to the development of such an attractive initiator system. We have moved through the very first generation of TiCl3-based systems to the most recent super-active third generation initiator systems based on MgCl2-supported TIC14 and electron donors, which have attained the targeted objectives. 2.1. Growth mechanisms in the polymerization process

Comprehending and controlling the polymer growth mechanism are the key points which have made it possible to obtain polymeric alloys directly in synthesis. The initiator particle-to-polymer replication phenomenon has been known for a long time and has also been exploited industrially. This phenomenon occurs when the polymer begins to grow, not only on the surface of the external crystals, but also on the surfaces of the crystals inside the initiator granule, progressively causing the granule to expand (Fig. 1). This process is particularly critical in the first phases, because too-fast growth could cause the initiator particle to physically explode, thus preventing regular replication. To achieve similar performance, it is necessary to act not only on the nature of the active center, and thus on the polymerization kinetics. It is also necessary to act on the morphology and structure of the initiator granule itself in order to give it a porous structure consisting of crystals and properly-sized, homogeneously dispersed, primary particles allowing the monomer equal access to the active centers. ~ The mechanism is the same, even when the initiator particle is not spherical but rather has an irregular structure. In this case, final products with a "flake-type" morphology are obtained. 12 Depending on the initiator system architecture (Fig. 2), polymerization conditions and the nature, number and distribution of active centers, a proper balance of these parameters will give, during polymerization, a progressive expansion of the various layers of the initiator particle according to Scheme (A) of Fig. 2, that is a replication controlled by monomer diffusion. Alternatively, it may give a more random expansion of all the initiator system crystallites indicated in Scheme (B) of Fig. 2, that is a replication controlled by the polymerization kinetics. ~3 A simplified equation able to explain both mechanisms is the following: 9't4-19 dy dt

_

KpCv(t)[M] 1 "~- z o- - y Tp

y(0)

=

0

POLYPROPYLENE-BASED POLYMER TECHNOLOGY

FIG. 1. Example of polymerization with perfect replication of an initiator system particle. FIG. 2. Polymer growth mechanisms.

305

306

T. SIMONAZZI et aL

where [M] = monomer concentration in the reaction medium (mol[1), Kp = specific polymerization rate (1/mol sec), Cp(t) = active centers concentration (molactinit)/(moltotinit), tD = 1/(pp/q2)Dp[M] diffusion time through the initiator/polymer layer (sec), Zp = 1/Kv Cp[M] polymerization time on the initiator surface (sec), y = qp/qc polymerization yield (gpolym/ginit). When we have very low polymerization yield values (y), or in the case when the diffusion time to is quite small in respect to the polymerization time zp, the growth mechanism is primarily under kinetic control. In such cases, the denominator is approximately equal to one and the growth mechanism is then considered as purely under kinetic control. dy ,,. kp Cp [M] dt Due to the fact that all active centers are equally accessible to the monomer, the replication phenomenon is dictated by the distribution of the active centers on the support. The growth mechanism is, at the other extreme, primarily diffusion controlled whenever the diffusion time to is significantly higher than the polymerization time %. In this case the denominator becomes significantly higher than one and the polymerization rate becomes inversely proportional to the yield. dy... 1 dt - ZD " y

2.1.1. Diffusion-controlled mechanism - Should the diffusive mechanism prevail, a discrete mechanism, governed by support tensile strength, causes the support to crumble from the growing polymer. In other words, the growing polymer promotes stress on the support and, once ultimate support tensile strength has been exceeded by the stress, there is a collapse of the first layer of support. Traces of this collapse can be seen in the polymer onion structure (Fig. 3). A mathematical model able to explain the mechanism behind initiator particle shattering and its associated stress field can, therefore, be seen in Fig. 4. Pushed by the concentration gradient, the monomer reaches the active centers and transforms into polymer, forcing the preexisting polymer away. In this manner, a pressure P0 is generated within the initiator layer reached by the monomer. The stress distribution generated as a consequence of this pressure is indicated on the right in Fig. 4: (a) A radial stress trr is generated within the support. This counterbalances the pressure P0: fir = --Po" (b) The tangent stresses trx, tensile in nature, generate the radial stress. As a function of the radius R the tangent stresses take on the following value: a 3 2R 3 + b 3 a T -- tr~ = try, = P0 2R 3 b 3 _ a 3 •

POLYPROPYLENE-BASEDPOLYMER TECHNOLOGY

307

FIG. 3. Onion structure during polymer growth.

These tangent stresses require some remarks: (1) The support tensile stresses are greater on the internal surface (R = a) than on the external surface (R = b) and thus: trx(a )

1 2a 3 -t- b 3

aT(b )

3

a3

(2) The smaller the initiator volume affected by the pressure (a 3 is small), the greater the internal tensile stress will be (aT(a) >> trT(b)). %

b

olymer Initiator core

%

FtG. 4. Schematic mechanism of initiator particle disaggregation.

308

T. SIMONAZZI et aL

(3) Also, the critical internal stress is expressed: 1 2a 3 + b 3 aT(a) = P02 b 3 - a 3" For layer thicknesses within such a range that b[a <<, 1.587, the tensile stress on the support is greater than the pressure P0. Therefore, a thin layer increases the disintegration tensile stress of the support itself. 2.1.2. Kinetically-controlledmechanism - Figure 5 shows a comparison between an initiator system particle before and after the very first polymerization step (pre-polymer). The uniform polymer growing throughout the active sites can be better observed in Fig. 6. This figure shows what remains of the original initiator system structure after the delicate prepolymer extraction by xylene. Finally, in Fig. 7 we can see the completion of the replication process. As polymerization continues, all the empty space is progressively filled. If, during polymerization, the monomer is replaced with another monomer or with a mixture of monomers it is then possible to produce a second, different polymer intimately dispersed among the various pre-existing voids found within the initially generated polymer (Fig. 8). Each of the granules contains different phases, regularly dispersed within it. The real polymerization reactor is no more the steel vessel containing the monomers but, instead, each of the individual granules where the polymerization kinetics is controlled. Exploitation of these "reactor granules" may have a tremendous impact on achieving advanced alloys and composite materials. When the reactivity-controlled (kinetically-controlled) mechanism prevails, any active center is homogeneously and simultaneously involved in starting polymerization. The growing polymer surrounding the simple initiator system crystallites determines the progressive expansion of the initiator system particle itself. Consequently, any variations in the distribution of active sites within an initiator system particle leads to variations in the amount of reaction in that area, and the formation of an irregular particle that does not replicate the original initiator particle morphology.

2.1.3. Crystalline nature o f reactor products - A spherically shaped polymer particle produced directly in synthesis shows a less organized structure in comparison to the conventional pellets. This is seen particularly in lower levels of crystallinity and a lack of spherulites. As a consequence, the energy required to melt a polymer with such a structure will be lower. Besides, the deformation of the spheres under the combination of heat and pressure is easier and, consequently, melt shearing starts earlier in extrusion devices) z This behavior, combined with appropriate screw design, can result in a series of processing advantages: (1) faster development of pressure resulting in an increased output (5-20%);

POLYPROPYLENE-BASED POLYMER TECHNOLOGY

FIG. 5. Comparison between initiator system particle, and prepolymer, structures. FIG. 6. Structure of initiator particle after prepolymer extraction.

309

310

T. SIMONAZZI et

al.

FIG. 7. Comparison of prepolymer ancl final polymer structures.

FIG. 8. Impact polypropylene copolymers obtained by synthesis ( x 8000).

POLYPROPYLENE-BASED POLYMER TECHNOLOGY

311

(2) more prolonged mixing in the melt state which leads to an improved dispersion of any pigments or auxiliary additives directly fed via dry blending; (3) a reduction in the amount of time the product is in contact with air at high temperatures, thus limiting the degradation due to processing steps; (4) lower energy consumption at equivalent output and melt quality; (5) lower extrudate melt temperature. Besides the above-mentioned advantages in the processing steps, one can also consider additional advantages based on the direct use of a polymer which has not undergone pelletizing. Such advantages are: (1) oxidation of the products is avoided, thus resulting in a lower sensitivity to degradation; (2) high molecular weight fractions and the original molecular weight distribution are maintained due to the absence of prior shearing. (3) The additive package need not provide the level of protection against degradation that is normally required for pelletization. Therefore, it can be precisely tailored for the processing and application requirements at hand. Consequently, it is possible to reduce the quantity of additives, and product purity will then be higher. A much more interesting area is that of the new grades of products, those which have become available from the Spheripol process. These "unique" products require greater study into applications so as to determine their capabilities in a field previously unavailable to polypropylene polymers. Some brilliant results have already been achieved in this field. As an example, some ultra-high melt flow grades (MFR = 400 and 800) have given excellent performance in melt spinning technology, making it possible to produce very low denier fiber at an increased output. Another particular field of investigation is opened up by the ability of the process to control the porosity of the polymer. Such porosity can lead to a wide range of densities which can be tailored to the final use (bulk density range 0.5-0.3 g/cm3). Starting from a low density "porous" product (0.38 g/cm3), and using a rather inexpensive cold-mix technology, it is possible to produce a series of masterbatches having a high pigment or additive concentration (over 20%), from both liquid and solid additives, while maintaining the typical free flowing behavior of the original product. Finally, the new initiator system-process combination allows for the production of grades with higher contents of rubber or other polyolefinic materials and characterized by either lower or higher melt flows than those achieved with less advanced processes or blending technologies. 2.2. Control o f molecular parameters 2.2.1. Initiator system activity and stereospecificity -

It has generally been

312

T. SIMONAZZI et

al.

accepted that propylene polymerization with high yield initiator systems, based on MgCl2-supported TiCI4 and Lewis bases, both internal and external, is kinetically controlled. It has been assumed also that the high productivity of such systems is due both to a greater concentration of active centers and to an increase in the propagation constant when compared to conventional initiator systems.2°The previous generation of high yield initiator systems were, however, characterized by a rapid decline in polymerization activity with time and by the well known inverse correlation between polymer productivity and stereoregularity. Such phenomena can be attributed to the presence of at least two families of active species on the initiator surface. These species differ in stereospecificity, "acidity" and stability. The non stereospecific centers would be selectively deactivated by the Lewis base while the isospecific centers would, in turn, be more readily deactivated with time than the others. Such behavior follows a chemical mechanism which most likely implies an over-reduction in the transition metal of the initiating center. 9 On the contrary, the new generation of high yield initiator systems make it possible to simultaneously obtain high polymer productivity and stereoregularity within a broad A1 alkyl/donor ratio (Fig. 9). Moreover, these initiator systems give rather regular polymerization kinetics (Fig. 10). The above properties are of fundamental importance, particularly in the case

100

1500

A

-ao~a~ "O

.c_ O

O

ca

13-

o

0.1

0.2

0.2

0.3

7O 0.5

D/AI molar ratio

FIG. 9. Effect of Donor/Al alkyl ratio on polymer productivity and stereoregularity. Polymerization in liquid propylene, 70°C, 2 hours v, • superactive systems; v, o conventional high yield systems.

POLYPROPYLENE-BASED POLYMER TECHNOLOGY

313

"1 / _

,op

i 3o]

0

. ~ , j ~ . ....o. lI

I

1

I

.le~

2 Polymerization time (hours)

I

3

4

FIG. 10. Kinetics of polymerization of superactive Ziegler-Natta initiator systems. Polymerization in liquid propylene; 70°C; ~A" A1/Donor = 60mol., ~t" • AI/ Donor = 20mol.

of synthesis of the so-called sequential copolymers. This is not only because they allow for the formation of the various polymeric fractions, but also because, as we shall see further on, they constitute a number of the necessary conditions for optimal distribution of the individual fractions within the polymer granule. These results have been obtained thanks to systematic research which has led to the identification of appropriate families or, better still, pairs of internal and external Lewis bases to be combined with the MgCI2 • TIC14/A1R3 system. In this regard, the key point has proved to be the replacement of the aromatic esters (the internal and external bases of the previous high yield system) with phthalic acid diesters (internal base) and alkoxysilanes (external base). 21 The action mechanisms of the Lewis bases are still a subject open to debate due to the complexity of the equilibria involved, which are, in turn, the result of a delicate balance between electronic and steric factors. Nonetheless, there is sufficient experimental evidence to attribute a more complex role to the external base than the simple role of the selective killer of non stereospecific active centers. The increase in isotactic polymer productivity l° and the increase in the stereospecificity of the first insertion step 22 in the presence of the external base would, in fact, indicate that at least some isospecific centers are "base-associated." These, in turn, would be derived from the transformation of non stereospecific centers into isospecific centers or from the activation of the pre-existing isospecific centers. It has been demonstrated that the isotactic activation effect of the external base is proportional to its ability to exchange with the internal base as well as

314

T. SIMONAZZI et al.

to its ability to fix itself to the initiator particle surface. The extent of this phenomenon depends on the nature of the internal base/external base pair. It has been found that it is enhanced when this pair is, for example, made up of diisobutyl phthalate (DIBP) and phenyltriethoxysilane (PES) rather than ethyl benzoate and methyl para-toluate. With regard to the role the internal base plays, it is known that it is progressively extracted from the catalytic surface by the AI alkyl and that, as previously mentioned, it undergoes an exchange reaction with the external base when it is present in the system. However, the stereospecificity of the first insertion step would appear to indicate that the internal base also plays a direct role in the formation of some types of isospecific centers. 23 Another effect of the nature of the internal/external base pair is its evident influence on the kinetic profile of polymerization. The greater stability of the isospecific sites associated with the diisobutyl phthalate/silane system as compared to the ethyl benzoate/methyl para-toluate system constitutes a further indication that the base participates directly in the formation of isospecific active centers, thus affecting the kinetic decay constant. Figure 11 shows the trend in productivity and isotactic index for polypropylene obtained with the MgCI2 • TiCIa/A1R3 diphenyldimethoxysilane(DPMS) initiator system as a function of temperature. As can be noted, both these parameters reach a maximum at around 70-80°C. In this case, therefore, no evident increase in the decay constant for isospecific initiator centers corresponds to an increase in the propagation constant. Furthermore, system selectivity is kept high as a

~ ¢,.,

0

--96 "10

.c

0

--

20

0 --II

y 50

e

I

60

~

92:0

1 ~ 0

I

70

0

I

80

90

~4

Temperature (°C) F1G. 11. Effect of temperature on polymer productivity and ster¢orcgularity. Polymerization in liquid propylene; 2 hr; Al/Donor = 20 tool.

POLYPROPYLENE-BASED POLYMER TECHNOLOGY

315

100

80

2.

0 0

o

C,

o

o J

60

g40 o

..,--/e

- - 98

/

/* 20

/ /

- - 94

I

2

I

4

I

6

a2

Time (h) FIG. 12. System selectivity and productivity vs time. Polymerization in liquid propylene; 70°C; Al/Donor = 20mol.

function of polymerization time (Fig. 12). In contrast, the earlier high yield systems based on the use of aromatic esters achieve a maximum yield at around 50-60°C and the isospecific centers are rapidly deactivated at temperatures above 60°C. This observation further confirms the different nature of the active species involved in the respective initiator complexes. Obviously, the high intrinsic stereospecificity of the newer high yield initiator systems based on phthalic acid diesters and alkoxy silanes in no way compromises our ability to achieve a reduction in homopolymer crystallinity whenever desired by acting on the AI alkyl/donor ratio and/or on the polymerization temperature. This objective obviously can be reached also by copolymerizing the propylene with appropriate comonomers such as ethylene, 1-butene, etc. 2.2.2. S t a t i s t i c a l c o p o l y m e r i z a t i o n - The simultaneous copolymerization of propylene with one or more comonomers, such as ethylene and/or 1-butene, at comonomer levels generally lower than 10mole percent, makes it possible to achieve a controlled reduction in the length of the stereoregular sequences of isotactic polypropylene and thus to decrease its crystallinity. The resulting products are more flexible, are more transparent, and have a lower melting point than the homopolymer. Such materials are particularly suitable for use in applications such as low temperature heat welded films. The introduction of comonomers into the isotactic macromolecular chain of polypropylene through copolymerization causes a fundamental variation in structural, rheological and mechanical properties of the polymer. In fact, the

316

T. SIMONAZZI et

al.

following parameters are affected: (a) the inter- and intramolecular interactions in the melt and solid states (with a variation in rheological properties and glass transition temperature); (b) the possibility of packing of regular crystalline domains (with a variation in crystalline form and melt temperature); (c) the crystallization kinetics (giving rise to variations in the effects of transformation conditions). Insertion of a second monomer - for example, ethylene - in the polypropylene chain causes a lowered melting temperature, proportional to the total ethylene content and its distribution along the polypropylene chain. Upon slow cooling from the melt, the resulting polymers crystallize into a mixture of • and 7 crystals. It has been demonstrated that the ~ form content increases with an increase in the ethylene content. 24"25The structure of the ? form crystalline cell of polypropylene, previously investigated by Natta and Corradini,26 is still the object of in-depth study.27"28 Zannetti and co-workers29found a linear correlation between the amount of crystalline y form present and the number of short isotactic polypropylene segments. These results support the assertion that y form development is determined by the interruption of the isotactic sequence. For random propylene/ ethylene copolymers synthesized with superactive initiator systems, 13C-NMR analysis (Fig. 13) has shown3° that no inversions exist in propylene units and that the polypropylene sequences, separated from methylene groups, show a high degree of stereoregularity.

tCp

tCp

CH 3PPPP

tCpp E

'PPE

|

PPEP

I

I- - EPEPE II, pPEPE

EPE

J



I

40

I

30 B in p.p.m

I

20

FIG. 13. Typical BC NMR spectrum of a random ethylene-propylene copolymer.

POLYPROPYLENE-BASED

5

--

o * zx o

PEP EPE EPE EPE

triad triad triad triad

POLYMER TECHNOLOGY

37.80 30.72 19.90 33.10

317

. / / 0

==4 .

"

~a ._z2 t-

1

1

2

3

4

5

6

7

8

9

10

wt.-% c2 FIG. 14. I n t e n s i t y o f t3C N M R

peaks vs composition of a random ethylene-propylene copolymer.

It is worth noting that the peak areas corresponding to the triads PEP and EPP observed in the ~3C N M R spectrum grow linearly with an increase in ethylene content (Fig. 14). However, at the same time, all peak areas corresponding to triads with two ethylene units also grow. The amount of crystallinity due to the 7 form and the melting temperature depend not only on the total ethylene content, but also on the distribution of ethylene units along the chain and the consequent nature of the short segments. Defects characterized by the sequence - - C H 2 - C H 2 - C H ( C H 3 ) - CH 2 - CH 2 - CH 2 (EPE triad) seem to have a greater influence than defects characterized by the sequences - CH(CH3) - C H 2 - C H 2 - C H 2 - CH(CH3) - CH2 - (the PEP triad) and the EPP triad - - C H 2 - C H 2 - C H ( C H 3 ) - c n 2 - C H ( C H 3 ) - C H 2 (Figs 15 and 16). In the case of ethylene and butene terpolymerization, the effect is much more marked; in fact, at the same c o m o n o m e r molar content, the percentage of the crystalline y form is higher and consequently the resulting melting point is lower with respect to similar bipolymers. 3t The presence of a c o m o n o m e r in the polypropylene chain also affects the overall crystallization rate constant K, the Avrami exponent ~/, and the half-time of crystallization t0.5.3° In the case of r a n d o m copolymers, both great sequential (intramolecular) homogeneity and compositional (intermolecular) homogeneity are required. These properties depend on the nature of the comonomers and on the initiator system. Compositional homogeneity, linked to the homogeneity of the active centers on the initiator system particle surface, minimizes the formation of by-products having a composition and/or structure other than that of the desired product, and in particular a m o r p h o u s fractions relatively rich in the

318

T. SIMONAZZI et

al.

7°r

o

2O 10

O'

1

2

3

4

5

6

C13NMR peak intensity

FIG. 15. ~-form index vs 13C NMR peak intensity of triads of a random ethylenepropylene copolymer.

comonomer. Sequential homogeneity, linked to the product of the reactivity ratios of the comonomer/initiator system, guarantees optimal distribution of the comonomers along the macromolecular chain, thus ensuring that they are better used in reducing crystallinity and melting point.

v445 I44o

435

43o0

--

A

1

2

i

3

i

4

O

I

5

6

C13NMR peak intensity FIG. 16. Melting temperature vs ~3C NMR peak intensity of triads of a random ethylene-propylene copolymer.

POLYPROPYLENE-BASED POLYMER TECHNOLOGY

319

It is known that the heterogeneous Ziegler-Natta initiator systems generally contain families of active species having different relative reactivities for propylene and ethylene. In particular, the non stereospecific centers are more reactive with ethylene than are the isospecific centers. 32 This has been confirmed by copolymerizing propylene and ethylene (< 10%tool) with initiator systems having different stereospecificities. The amount of amorphous copolymer (relatively rich in ethylene), soluble in xylene at room temperature, decreases as the system isospecificity increases. In the case of random propylene-l-butene copolymers, a greater compositional homogeneity was observed by fractionating. This is most likely due to the fact that propylene and butene are basically polymerized on the same active centers. A new and highly interesting random copolymer has recently been obtained by copolymerization of propylene and 1,3-butadiene. 33'~ In the presence of HIMONT super-active initiator systems, copolymerization of propylene and 1,3-butadiene in liquid propylene allows the production of random copolymers having different microstructures depending on reaction conditions and initiator system composition. In particular, the insertion of the butadiene units can be controlled either by changing the polymerization temperature (Fig. 17) or the aluminum alkyl (Ai)-to-electron donor (D) ratio (Fig. 18) so as to obtain copolymers where the butadiene is predominantly linked in position 1,2. 8

O

_~

•1,2 1,4

6 0

E ~,

4

.e:

2

I

I

I

40

60

80

Temperature(*C)

FIG. 17. Effect of polymerizationtemperature on propylene-butadienecopolymer structure. Polymerization in liquid monomers; butadiene 25%wt, A1/Donor 7.5mol, 4hr.

320

T. SIMONAZZI et

al.

._c "1o

ol, 2 ol,4

I

I

5

10

20

AI/D ratio (mole) FIG. 18. Effect of A1/Donor ratio on propylene-butadiene copolymer structure. Polymerization in liquid monomers; Butadiene 25% wt, A1/Donor 7.5 mol, 70°C, 4hr.

This represents a novelty in the field of Ziegler-Natta polymerization since only copolymers characterized by a predominant 1,4 addition of the butadiene units have been reported thus far. 2.2.3. S e q u e n t i a l c o p o l y m e r i z a t i o n - Sequential ethylene-propylene copolymers have often been marketed under the name of "block" copolymers since it was felt that their structure was made up essentially of heteroblocks. However, both experimental and theoretical evidence no longer supports this theory. In fact, at the very minimum, the system must be able to satisfy the following conditions to develop heteroblocks: (1) presence of active centers able to polymerize both monomers Ml and M2; (2) presence of active centers which are sufficiently stable in time (living initiator system); (3) a sufficiently long macromolecule lifetime (living polymer). Of the above conditions, the first is generally satisfied, the second only in part, and the third proves to be the exception rather than the rule. In fact, even in the hypothetical case of a living initiator system, the well known chain transfer mechanism intrinsic to Ziegler-Natta systems is such that it significantly limits the average lifetime of the growing macromolecule. In effect, contrary to what has been reported in the literature, it has recently been demonstrated35that, even operating under optimal conditions, it is impossible to obtain well defined block copolymers. A rare example of a block copolymer obtained with a Ziegler-Natta initiator system is the one obtained by Doi 36 using sequential polymerization of propylene and ethylene with the syndiospecific initiator system V(acac)3/(A1Et2Cl/anisole at - 78°C. In general,

POLYPROPYLENE-BASED POLYMER TECHNOLOGY

321

the "block" copolymers obtained with Ziegler-Natta initiator systems under normal polymerization conditions are, most likely, a mixture of homopolymers in which the presence of true and proper "blocks", if any, is unlikely. It is, therefore, clear that the term "block copolymer" should be appropriately modified. For example, today, scientific literature and patents ever increasingly use special terms to denote the polypropylene/(ethylene-propylene) type of copolymers obtained with sequential polymerization of the propylene and ethylene-propylene mixture. Such terms reflect the morphology (i.e. heterophasic copolymers), means of preparation (i.el sequential copolymers, reaction blends, chemical blends, in situ copolymers) or, mechanical properties (high impact copolymers) of the polypropylene obtained. Besides being based on theoretical conditions, their blend nature is given by the following experimental evidence: (1) possibility of dividing the raw copolymer into its individual components; 37 (2) similarities in morphology and dynamic-mechanical properties with corresponding materials prepared by mechanical mixing of the individual components. 3s,39 High impact copolymers are, in fact, heterophasic polymers where polypropylene (the major component) is the continuous phase, and the elastomeric phase (ethylene-propylene rubber) is uniformly dispersed within the matrix. Tailoring the initiator system is important not only for the molecular structure, but also for the location of the various polymeric phases. In heterophasic copolymers, it is essential that the rubbery phase be homogeneously dispersed and that its size be controlled in order to achieve an optimal stiffness-impact balance. To obtain such a product, initiator system performance must be optimized with a tailor-made process able to initially produce a homopolymeric phase with an elevated isotacticity and then to uniformly polymerize the desired elastomeric material within the matrix. In this regard, not only is the morphology of the initiator system pellet important but so too is its kinetic behavior. In fact, if at the end of the homopolymerization stage activity is not sufficiently high, as in the case of the previous generation of initiator systems, the system must be reactivated by adding fresh aluminum alkyl. However this brings about several important factors which cause uncontrolled growth of the rubber fraction: 4° (1) First, the newly generated polymerization centers are not distributed homogeneously along the polymer granule (they are located mainly on the shell) and among the various polymer particles. This problem stems from the difficulty the fresh aluminum alkyl has in permeating the polymer granule and in uniformly spreading over the different particles. This is accompanied by corresponding fluctuations in distribution of the rubber since the overall system reactivity depends on the actual A1/Donor ratio. (2) Second, the inhomogeneous distribution of the aluminum alkyl is most likely accompanied by other drawbacks, such as undesired and uncontrolled

322

T. SIMONAZZI et al.

changes in the nature of the rubber fraction, since it is known that the Al/Donor ratio affects both bipolymer composition and molecular weight. Similar effects could be caused by hot spots arising from the high reactivity associated with the occurrence of high local Al/Donor ratios. Today, with the Spheripol process, the new high yield initiator system, based on Ti salts supported on active MgCI2, is able to meet the challenge if the proper Lewis base is employed. In other words, it is possible to obtain a highly crystalline polypropylene phase with isotacticity values reaching 99% and an elastomeric phase homogeneously distributed within the former. This results in the best balance between stiffness and impact resistance. The versatility of the new initiator system and process is so high that it is possible to obtain a whole range of elastomeric and plastomeric phases including, optionally, a polyethylene fraction (low blush copolymers). Due to the fact that the elastomeric phase grows on the same initiator system crystals which previously produced the homopolymeric phase, a morphological structure nearly approaching the one desired is obtained directly in synthesis. Beyond the evident advantages in terms of energy savings (the need for homogenization being minimized), direct synthesis makes it possible to obtain a particular elastomeric phase structure, for example, fractions with a certain crystallinity which improve the compatibility and property balance of the final product. 41 The amount of elastomeric phase, which continues to form in the interstices of the polymeric matrix, increases as the copolymerization process proceeds (Fig. 8). Performance of the new initiator system was so greatly improved, with regard to copolymerization, that now it is possible to obtain polymers which can no longer be considered merely as highly impact-resistant polypropylene, but rather as "free-flowing" elastomers with spherical morphology. Of particular interest, both from the theoretical and practical points of view, are understanding and controlling the "architecture" of the previously mentioned low blush, or low whitening, copolymers. The whitening phenomenon is due to cavitational effects and begins after a particular strain level has been reached within the copolymer. The addition of polyethylene decreases the whitening effect, shifting its beginning to higher longitudinal deformation values. One of the roles of the polyethylene is to reduce residual stresses. In fact, polyethylene is inserted within the heterophase due to interface tension phenomena. During the melt cooling process, the volumetric contraction of the polypropylene matrix due to crystallization results in a compression of the heterophase, made up of rubber and molten polyethylene. The subsequent polyethylene crystallization, with consequent volumetric contraction, promotes a progressive decrease in the heterophase stress, t2 2.2.4. Molecular weight and molecular weight distribution - The molecular weight of the (co)polymer can easily be controlled by using hydrogen as a

POLYPROPYLENE-BASED POLYMER TECHNOLOGY

323

15

10

5

I

I

0.1

0.2

[H2I

0.5

I

I

0,3

0.4

0.5

FIG. 19. Effect of gas phase hydrogen concentration on polymer molecular weight. Polymerization in liquid propylene, 70°C, A1/Donor = 20 mol.

regulator. Similar to what has been found for earlier Ziegler-Natta initiator systems, high yield systems based on phthalic acid esters and silanes also show a linear correlation between the inverse of the average viscosimetric molecular weight of the polypropylene and the square root of the hydrogen concentration in the reaction medium (Fig. 19). This is in agreement with a mechanism which, for chain transfer, calls for an interaction between polymerization centers and atomic hydrogen after dissociative adsorption of the hydrogen molecule on the initiator surface.42 It has been experimentally found that, in terms of regulation of the polymer molecular weight, the system's sensitivity to hydrogen depends on the nature of the external base. This is a factor which also has an effect on the molecular weight distribution of the polymer. Assuming that, as is most likely, this phenomenon takes place under strict kinetic control, it can be suggested that the external base also affects the kinetic constant governing the transfer of active centers and, possibly, their distribution. Obviously, should one wish to obtain a significant broadening of the molecular weight distribution, it is possible to operate in two or more stages with different hydrogen concentrations. The role of the hydrogen is not, however, limited to that of a simple molecular weight regulator. For the Ziegler-Natta initiator systems it is known that hydrogen can have an activating or deactivating effect depending on the nature of the system and of the monomer. 42 In the case of systems containing phthalic acid esters and alkoxysilanes, hydrogen has been found to produce a reversible activating effect in propylene polymerization. The extent of this activation is markedly higher than that which was found with earlier systems based on aromatic esters.

324

T. SIMONAZZI et al.

The mechanism responsible for this phenomenon has not as yet been identified although it is most likely controlled by chemical rather than physical factors. As a working hypothesis, the increase in initiator system activity with an increase in hydrogen concentration may be attributed, for example, to an overall increase in the concentration of active centers linked to an oxidative addition of the hydrogen to the transition metal. Alternatively, it may be due to the removal from the active site, by chain transfer, of any macromolecules having terminal units added in the 2,1 rather than 1,2 fashion and thus sterically more cumbersome.43 2.3. Initiator system and process technology synergism The advent of the high yield/high stereospecificity initiator system has had a major impact on the complexity of a polypropylene plant, reducing it to a few processing stages. Three costly plant sections have been eliminated, namely: amorphous recovery facilities; initiator residue removal facilities; product extrusion facilities. The process has evolved in parallel with initiator development so that it can exploit the enormous potential of the initiator system. In order to make production of either very highly crystalline or amorphous materials or of either mono- or heterophasic polymers a matter of choice, Himont has recognized that a two-step hybrid process is the only alternative (Spheripol process). Research on the initiator system has supplied all the elements for the definition of an extremely simple and versatile process. This process is now reduced in practice to one stage of polymerization for homopolymers or random copolymers - a loop in liquid monomer phase - possibly connected to one or two gas phase reactors for heterophasic copolymers8'a4 (Fig. 20). The base criteria leading to the development of this process are reported below.

Propylen ~ Chemic~-~1 als~ ~ itiator &

z

~.

~_ '

Ethylene FIG. 20. Spheripol process flow diagram.

z~>polypropylene Spherical tostorage

POLYPROPYLENE-BASED POLYMER TECHNOLOGY

325

2.3.1. General criteria - A process which uses a parattinic hydrocarbon (a diluent) as the reaction medium, and which can be further simplified with the introduction of the new initiator system, is still complex and shows the limitations of the original technology. The required initiator system mileage (g of initiator particles per kg of product) can be reached only with large reaction volumes, thus reducing operating flexibility. Furthermore, the range of copolymers which can be produced still proves limited since the tendency of the diluent to extract the more amorphous fractions from the copolymer penalizes both heat exchange in the reactor and subsequent operations downstream (product stickiness). It seems evident that the total exploitation of all the new potentials offered by the new initiator system requires adopting technology based on using the monomer (liquid or gas) as the reaction medium. Whatever the case may be, the approach must take into consideration the fact that each growing particle is, in effect, a microreactor. As such, it must be placed under such conditions that from the very first instant - it can exchange both heat and material, availing itself of an ideal monomer component ratio. Furthermore, while homopolymers and random copolymers require a single reaction stage, obtaining heterophase copolymers implies an additional reaction step. The different polymerization sequences do not necessarily require adoption of the same technology. 2.3.2. Homopolymer (and random copolymer) reaction technology - All other conditions being equal, for a stable hourly production within the practical application range, kinetic considerations indicate that the mileage which can be achieved in a continuously agitated reactor is largely proportional to the product of monomer fugacity and polymer hold-up within the reactor. In order to take advantage of all the opportunities the market has to offer in terms of product range while, at the same time, limiting storage, modern plants must all too frequently change their production campaign. Limiting plant hold-up in order to minimize the amount of transition product is a significant economic factor. 2.3.3. Gas phase polymerization - This type of polymerization makes it possible

to obtain a dry solid polymer directly as it comes out of the reactor. On the other hand, critical points are the proper distribution of the initiator system components and the ability to fully control polymer particle growth. If one wishes to guarantee homogeneity and product structural morphology, these factors indicate technological problems which will have to be overcome. The reaction temperature is no longer linked to the physical properties of the reaction medium. Nonetheless, above the temperature range where use of liquid monomer is operable, the effectiveness of the gas phase process is rather limited. Therefore, in order to maintain adequate polymer yield per gram of initiator system with an acceptable polymer hold-up, one must operate at a pressure

326

T. SIMONAZZI et al.

(23-35 bar) close to the vapor pressure of liquid monomer at the reaction temperature. The most attractive technology is that of a fluid bed reactor which provides for external recirculation of the cooled reaction gas. This simultaneously removes the polymerization heat and agitates the polymer bed. It is worth noting that the fluidizing flow must be maintained well above the minimum required to suspend the polymer spheres to as to prevent any "hot spots". Therefore operating pressures and bed size have a significant effect on controlling the process. Furthermore, the size of the reactor must take into account the volume required for solid/gas disengagement. This is a volume which is not used by the reaction but which significantly contributes to investment costs. 2.3.4. P o l y m e r i z a t i o n in l i q u i d m o n o m e r - The strong points of this type of polymerization are: (1) high polymerization rate; (2) excellent heat exchange on the growing individual polymer particles; (3) initiator system which is homogeneously dispersed throughout the entire reactor (the coinitiators - the aluminum alkyl and the Lewis base - are completely soluble in the reaction medium); (4) low mixing energy. On the other hand, use of the liquid medium leads to evaporation of the unconverted monomer upon leaving the reactor, which also requires the same recycling. The most interesting technology is made up of a loop reactor. In this apparatus polymer suspension and agitation are guaranteed by an adequate circulation of the reaction slurry. The reactor operates in a completely full state, thus avoiding areas which lay wet and dry, that could give rise to uncontrolled polymer growth. The particular geometry, coupled with a high circulation rate, makes removal of the reaction heat rather easy and thus the reactor is relatively insensitive to polymer particle size. It is, therefore, particularly suited to operating at high concentrations (up to 60% polymer by weight). Hence such a reactor can achieve absolutely the highest specific productivity (even above 300 kg/hr of polypropylene per m 3 of reactor) and this with relatively modest energy consumption (110 MJ per 1000 kg of polymer produced). The disadvantages linked to the liquid medium are minimized. Evaporation of the monomer exiting from the reactor requires a rather low energy consumption (less than 100 kg of steam per 1000 kg of product). When polymer/monomer separation is performed in a certain pressure range, the monomer can be condensed with the cooling water. Thus, recycling the monomer which was not converted in the reactor accounts for only a small portion of the investment costs. Furthermore, adopting an effective, reliable reaction "killer" decreases the degree of danger of the reactor, bringing it to the levels of a modest liquid propylene storage tank.

POLYPROPYLENE-BASED POLYMER TECHNOLOGY

327

2.3.5. Reaction technologies for heterophasic copolymers - The most important group of polypropylene resins comprises the high resilience copolymers in which polypropylene makes up the continuous phase, which an elastomeric phase (ethylene-propylene rubber) adequately dispersed throughout the matrix. A further reaction unit is thus required, in this case, downstream from the homopolymerization reaction. The initiator system favors the growth of elastomeric islands of appropriate size dispersed within the homopolymeric phase. The morphological structure obtained directly in synthesis approaches the ideal structure required. Maintaining this structure is a marked advantage in applications. The monomers in the liquid medium partially extract the created elastomeric phase. A gaseous reaction medium, on the other hand, makes it possible to prevent undesired redistribution and the reaction product retains the flowability typical of a homopolymer. In this case the gas phase proves to be the most ideal approach. In fact, there are no problems in terms of distribution of the initiator system components (derived from the initiators within the polymer particles in the homopolymerization stage). The fluidized bed reactor proves to be particularly well suited. The required conversion may be performed with reactors of rather modest volume and with relatively low operating pressures (reduced energy consumption), while the gas flow required for fluidizing ensures a constant monomeric composition at all points. 3. INDUSTRIAL RELEVANCE OF HIMONT TECHNOLOGY The Ziegler-Natta initiator system developments and the parallel evolution of the process plants - previously described - led to a breakthrough in the production of both polypropylene homopolymers and random and impact copolymers. Due to the advantages in terms of product quality, process flexibility and operating economies, the Spheripol process - introduced in 1983 - has been the most successful technology ever developed. Announced expansions equal to approximately 60% of the world's existing capacity are based on the Spheripol process. Yet the innovative spirit of Natta and Pino continues to thrive in 1990. Present developments resulting from the Himont technology will draw further advantages from the ability to copolymerize different monomers and to control particle morphology, specific surface area and porosity. In fact, with the new revolutionary Ziegler-Natta initiator systems, it is even possible to produce modified polymers which offer unique potential for achieving new products. These take advantage of both the initiator reactivity and the possibility of tailoring molecular weight, molecular weight distribution, composition and the distribution of the different monomers throughout the chain over very broad ranges, from crystalline polymers to elastomeric products.

328

T. SIMONAZZI et al.

The presence of unsaturation along the polyolefin backbone is of great interest since the reactivity of the carbon--carbon double bonds, or that of the allylic hydrogen, offers unique potential with which to obtain new products or to impart further properties. The unsaturated polyolefins can, in particular, easily be chemically retailored and grafted, drastically changing their chemical nature and compatibility with other polymers. Such would make it possible to create new alloys of polyolefins and engineering plastics. REFERENCES 1. 2. 3. 4. 5. 6.

7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

G. NATTA,P. PINO and G. MAZZANTI,U.S. Pat. 3,112,300, Nov. 26, 1963; It. priority June 8, 1954. Ital. Pat. 566,242 (Sept. 26, 1955) Montecatini. hal. Pat. 612,695 (Feb. 24, 1958) Montecatini. Ft. Pat. 1,130,231 (Nov. 3, 1972) Solvay. N.B. CHUMAEVSKUet al. Makromolek. Chem. 177, 747 (1976). (a) U.S. Pat. 4,107,413 (Aug. 15, 1978) Montecatini; (b) U.S. Pat. 4,107,414 (Aug. 15, 1978) Montecatini; (c) U.S. Pat. 4,107,319 (Aug. 15, 1978) Montecatini. Y.I. YERMAKOV,B. N. KUZNETSOVand V. A. ZAKrlAROV,Catalysis by Supported Complexes, Elsevier Publ., New York (1981). R. Dl DRUSCO,Hydrocarb. Proc. 113 (November, 1984). P. GALLI, L. LUCIANIand G. CECCmN, Angew. Makromolek. Chem. 94, 63 (1981). P. GALLI, P. C. BARBi~and L. NORISTI,Angew. Makromolek. Chem. 120, 73 (1984). T. SIMONAZZI,G. CECCHINand A. MAYR, VIII Convegno AIM, Milan (October, 1987). P. GALU, T. SIMONAZZland D. DELDUCA,Acta Polym. (Berlin) 39, 81-90 (1988). P. GALH, G. CECCHINand T. SIMONAZZI,Frontiers in Macromolecular Science, IUPA C 32nd International Symposium, Kyoto (1988). P. GALU and L. NORlSTI, 5th Eur. Plast. Caoutch. Conf., 1, A5/1-A5/5 (Soc. Chim. Ind.) Paris (1978). J.W. BEGLEY,J. Polym. Sci. AI, 319-336 (1966). R.L. LAWRENCEand M. G. CHIOVETTA,Polymer Reaction Engineering (K. H. Reichert et al. Eds), p. 78, Munich (1983). S. FLOYDet al. J. Appl. Polym. Sci. 33, 1021 (1987). M.A. FERREROand M. G. CHIOVETTA,Polym. Eng. Sci. 27, 1436 (1987). S.P. TIMOSHENKOand J. N. GOODIER, Theory of Elasticity, I.S.E. (1970). P.C. BARB~,G. CECCmN and L. NORISTI,Adv. Polym. Sci. 81, 1 (1987). Eur. Pat. 45977, S. Parodi et al. Montedison, invs. M.C. SACCHIet al. Macromolecules 23, 383 (1990). (a) P. C. BARB~,L. NORlSTIand G. BARUZZI,Relations between homogeneous and heterogeneous catalysis, 6th Int. Syrup., Pisa (Sept. 1989). (b) M. C. SACCmet al. Symp. on Recent Developments in Olefin Polymerization Catalysts, Tokyo (23-25 October, 1989). P. CAVALLO,E. MARTUSCELUand M. PRACELLA,Polymer 18, 42 (1977). E. MARTUSCELLI,M. PRACELLAand L. CRlSl'Ir~O,Polymer 24, 693 (1983). G. NATTAand P. CORRADINI,Nuovo Chim., Suppl. 15, 40 (1960). S. BRUCKNER,S. V. MEmLEand W. Pozzlo, IX Convegno Italiano AIM, Vol. I, pp. 305-308, Bologna (1989). E. ORI, P. SGARZI,G. PAGANETTOand E. MARCHETTI,I X Convegno Italiano AIM, Vol. II, pp. 425-429, Bologna (1989). G. GUIDETTI,P. BUSl, I. GIULIANELLIand R. ZANETTI,Eur. Polym. J. 19, 757 (1983). M. AVELLA, E. MARTUSCELLI,T. SIMONAZZlet al. Makromolek. Chem. 187, 1927-1943 (1986). P. GALLI,T. SIMONAZZIand P. C. BARB~,VI Convegno Italiano di Scienza delle Macromolecole (AIM), Pisa (October 1983). Y.V. KISSIN, lsospecific Polymerization ofOlefins, Springer-Verlag, New York (1985).

POLYPROPYLENE-BASED POLYMER TECHNOLOGY 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

329

G. CECCHIN and T. SIMONAZZI, Macromolecules 89, Oxford (September, 1989). U.S. Pat. 4,607,077, Cecchin et al., Himont Inc. V. BUSlCOet al. Makromolek. Chem., Rapid Commun. 5, 737 (1984). Y. DOl, S. UEKI and T. KEII, Makromolek. Chem., Rapid Commun. 3, 225 (1982). R. FRANZESE, unpublished internal data. T. TAKAHASHIet al. J. Macromolek. Sci. Phys. B22(3), 425 (1983). J. KARGER-KOCSlS and L. KISS, Polym. Eng. Sci. 27(24), 254 (1987). G. CECCH1N and C. COMETTO, Euchem Conference, Lahti, Finland (August, 1986). L. D'ORAZIO, R. GRECO, C. MANCARELLA,E. MARTUSCELLI,G. RAGOSTAand C. SILVESTRE, Polym. Eng. Sci. 22, 536 (1982). T. KEn, Kinetics of Ziegler-Natta Polymerization, Kodansha, Tokyo (1979). U. GIANNINI,private communication. R. RINALDI, Chim. Ind. (Milan) 3, 234 (1990).