Styrenic TPEs - the pathway to invention.

Over the past 35 years, styrenic thermoplastic elastomers have gone from a laboratory curiosity to a major item of commerce. Their worldwide production in 1995 was estimated to be about 535,000 metric tons and this should increase to about 715,000 metric tons by the year 2000 (ref. 1). This last production estimate is equivalent to a value of about two billion dollars. The properties and applications of styrenic thermoplastic elastomers have been extensively described (refs. 2 and 3). This article addresses another question -- how did such an invention occur? Legge described much of this in his Goodyear medal address (ref. 4), and this article covers some of these developments in more detail.

Nomenclature 

In this article, A is used to represent a block of A mer units and, similarly, B to represent a block of B mer units, etc. Hyphens denote the links between individual polymer blocks. In following examples:

A-B A-B-A (A-B)n (A-B)nX

A-B represents a diblock polymer and A-B-A represents a triblock copolymer with two terminal A blocks and a B center block. Similarly, (A-B)n represents an alternating block copolymer, A-B-A-B-A- . . ., etc., whereas (A-B)nx represents a branched block copolymer with n branches (n = 2, 3, 4 ...) and a junction point x.

The first letter of the monomer unit is used to denote the polymer block. For example, a three-block copolymer, poly(styrene-b-butadiene-b-styrene) is written as S-B-S. If one of the blocks is a copolymer (e.g., ethylene-propylene rubber), the block copolymer poly(styrene-b-ethylene co-propylene-b-styrene) is written as S-EP-S.

Historical review

First, let us consider the state of research in the synthetic rubber industry in the early 1960s. A prime objective of this research was the economical polymerization of isoprene to give a polymer of high cis-1,4 structure -- that is, a synthetic version of natural rubber. It was also recognized that polybutadiene (which is based on a cheaper and more abundant monomer) should be commercially attractive. In the mid-1950s this work was stimulated by papers describing synthesis on a semi-commercial scale using both Ziegler type catalysts (ref. 5) and lithium metal catalysts (ref. 6). Workers at the Shell De-velopment Co. and at other industrial research laboratories (ref. 7) investigated lithium metal initiators for diene polymerization. They found that alkyllithiums ([R.sup.-][Li.sup.+]) are more convenient to use than the metal itself. In these reactions, with properly purified monomer and solvents, and under anaerobic conditions, there are no chain termination or chain transfer steps. Thus, when all the original monomer (A) is consumed, the polymer chain remains active.

[R.sup.-][Li.sup.+] + nA --- [is greater than] [R-A-A-A-A-A-A-A.sup.-] [Li.sup.+]

This product is termed a "living polymer" because it can initiate further polymerization of monomer (either of the same or of a different species). Thus, if a second monomer (B) is added.

[R-A-A-A-A-A-A-A.sup.-][Li.sup.+] + nB -- [is greater than]

[R-A-A-A-A-A-A-A-B-B-B-B-B-B.sup.-][Li.sup.+]

In 1957 (ref. 8), a process was described for the manufacture of polystyrene/polydiene block copolymers using alkyllithium initiators. Although at the time these polymers were not recognized as thermoplastic elastomers, this research contributed to our background knowledge. About this time, triblock copolymers were also reported in which the polymerization initiator was difunctional (refs. 9 and 10). These had the structure I-S-I. For this reason (see later) they were not thermoplastic elastomers.

About this time, Shell Chemical began to manufacture polyisoprene on a commercial scale and was considering the manufacture of polybutadiene. In support of this, work was being done on the rheology of these polymers, in the hope of resolving a serious problem in their commercial manufacture and subsequent storage. Both (especially polybutadiene) show extreme "cold flow" -- that is, during storage they lose their original shape and ooze out of their containers. In other words, even at room temperature, these polymers, although they seem to be solids, are in fact very viscous liquids. Research on this problem started with the construction and operation of a modified Mooney viscometer (one dating back to the second World War). This differed from the conventional instrument in three important ways:

* The rotor was a biconical type -- this allowed the results to be expressed in absolute units and so the data could be compared to those obtained on other instruments such as capillary rheometers.

* We could vary the rate of revolution of the rotor within a very wide range -- from about [10.sup.-5] [sec.sup.-1] to about 5 [sec.sup.-1]. At the slowest speed the rotor took about two weeks to make one complete revolution. This slow speed enabled one to measure polymer viscosities at rates approaching those encountered in cold flow.

* The temperature of the polymer being tested could be varied from about 25 [degrees] C to about 200 [degrees] C.

At low shear rates, it was found that even at room temperature, both elastomers approached Newtonian behavior. In other words, their viscosities leveled out to constant values as the shear rate approached zero. The results were published on polyisoprene (ref. 11) but not those on polybutadiene, because Krause and co-workers at Phillips Petroleum reported similar results (although obtained by different experimental techniques) (ref. 12). It was noted that just as polybutadiene showed more cold flow than polyisoprene, it also showed less variation of viscosity with shear rate (i.e., it more nearly approached Newtonian behavior). The viscosities of these polymers were correlated at low shear rates with their intrinsic viscosities, measured in toluene solutions. Intrinsic viscosity was used because at the time, it was the easiest measurable quantity that correlated with molecular weight.

Discovery

In 1960, Ralph Milkovich joined the group. He had previously worked on block copolymers produced by difunctional initiation (ref. 9) and he soon made an S-B diblock polymer of about 14% styrene content. It had a molecular weight (calculated stoichiometrically) of about 210,000. This was a tapered block copolymer (i.e., one with a compositional gradient between the two segments rather than a sharp boundary). Its intrinsic viscosity was in the expected range for this molecular weight. In the unvulcanized state, it was not a very promising elastomer. It was dry and brittle, but it had two outstanding features. First, at room temperature, its viscosity was very non-Newtonian, even at the lowest measurable shear rates. Second, at low shear rates especially, its viscosity was much higher (by factors of up to 100) than a polybutadiene of similar intrinsic viscosity (figure 1). An explanation of these discrepancies was sought. It was known that polystyrene and polybutadiene are incompatible as homopolymers and so it seemed reasonable to assume that they are similarly incompatible as segments in the same polymer molecule. If so, this S-B block copolymer should form a two-phase system. Since the polystyrene segments are only 14% of the total polymer, they should agglomerate to give discrete particles dispersed in a matrix of polybutadiene, as shown in figure 2. These discrete polystyrene particles were called "domains." In the solid state, at room temperature, these domains are hard and strong. Thus, they tie the polybutadiene segments together to give the equivalent of a highly branched and very high molecular weight elastomer. This agglomerated structure has a much higher viscosity than a polybutadiene of similar molecular weight to the individual S-B molecules. However, in a common solvent such as toluene, the structure dissociates and gives a solution with more normal properties.

[Figures 1 and 2 ILLUSTRATION OMITTED]

This development seemed to have solved the problem of cold flow, but another one remained -- that of green strength. This is the strength of an elastomer during and after being mixed with the various compounding ingredients but before vulcanization. Lack of this property makes the unvulcanized compounds difficult to process. The S-B diblock copolymer had little or no green strength. It was expected that a triblock S-B-S or S-I-S copolymer would be much better, because it would form a structure similar to that shown in figure 3. Here, the ends of each elastomer segment are in different domains, giving a polymer network with similarities to a vulcanized rubber. At room temperature, this network should have more strength than the corresponding homopolymer, because the hard polystyrene domains act as "tie points" for the ends of the elastomer chain. However, at temperatures high enough for the domains to soften (above about 100 [degrees] C), the triblock polymer should be able to flow. Like the S-B, it should be soluble in toluene and similar solvents.

[Figure 3 ILLUSTRATION OMITTED]

While some improvement in strength was expected, the extent of it was amazing. It was found that the tensile properties of unvulcanized S-B-S and S-I-S were similar (or even superior) to those of conventional vulcanizates -- that is they had high tensile strength, high elongations and rapid and almost complete recovery after elongation (figure 4). Reprocessing and remolding left these properties unchanged. In other words, these polymers could be processed as thermoplastics but at room temperature had the qualities of vulcanized rubbers. The term "thermoplastic elastomer" was coined to describe this new class of polymers. It was later found that Shollenberger and co-workers had used it several years earlier to describe segmented polyurethane elastomers (ref. 13).

[Figure 4 ILLUSTRATION OMITTED]

This explanation was given in terms of S-I-S and S-B-S block copolymers. It should also apply to other styrenic block copolymers, including alternating block copolymers such as (S-I)n and branched block copolymers such as (S-B)nX (where x represents a multifunctional junction point), since these should also form continuous network structures similar to that shown in figure 3. On the other hand, block copolymers such as S-I and B-S-B cannot form these structures, since only one end of each polydiene chain is attached to a polystyrene segment. Because of this, S-I, B-S-B and similar block copolymers are weak materials with no resemblance to conventional vulcanized rubbers (ref. 14).

When this explanation was postulated (ref. 14), it was also generalized to include all block copolymers with alternating hard and soft segments. It was specifically applied to polycarbonate/polyether and poly(dimethylsiloxane)/poly(silphenylenesiloxane) alternating block copolymers. A similar explanation has been applied to segmented polyurethanes (refs. 15-17). It is now accepted as the underlying mechanism that gives most thermoplastic elastomers their valuable properties.

Following the announcement (ref. 18) of these styrenic thermoplastic elastomers and a symposium on block copolymers (ref. 19) held at the California Institute of Technology, there was a surge of interest in thermoplastic elastomers and in multi-phase systems. In this respect, the domain theory served as a paradigm (ref. 20). Its simple explanation combined with the unequivocal structure and obvious commercial importance of the styrenic thermoplastic elastomers has spurred the development of a new field of polymer science.

It is important to note that the discovery of these thermoplastic elastomers was serendipitous and drew on much background research. No one at Shell ever set out to discover thermoplastic elastomers. The solution of the original problem (cold flow and poor green strength) was commercially unimportant, but it led to a different and very significant development. Had the research been too narrowly focused or too closely planned, this opportunity would probably have been missed.

Commercial development

Clearly, much work would be required to determine if the styrenic thermoplastic elastomers had commercial potential. Here, Legge played the vital role. As well as being the director of the original research, he was also the "product champion" for the commercial development. He was able to persuade Shell's upper management to divert the needed resources to this venture. Two fundamental questions had to be answered:

* Could these styrenic block copolymers be made economically and with reproducibility on a commercial scale?

* If so, for what applications could they be sold?

Anionic block polymerization was not an easy process to scale up from the laboratory to the plant, nor even now is it easy to make a consistent product. It requires extremely high purity solvents and monomers and any trace of air or other potential chain terminators must be excluded. For obvious reasons, the details of the successful commercial process have never been disclosed, but its developers deserve much credit.

One area of uncertainty was the patent situation. The basic patent was applied for (ref. 20) in early 1962, but it did not issue until 1966. For over four years, the possibility was faced that another company had made a similar discovery independently and filed their patent application before Shell's. Thus, the decision to continue work on this project in the meantime involved an extra element of risk.

From the beginning, it was recognized that tires were not a very promising market for styrenic thermoplastic elastomers, because these polymers lose their strength at higher temperatures. More promising applications were footwear, general purpose rubber goods, polymer modification and solution adhesives. Hot melt adhesives, which today are probably the largest single application, were not even considered then. The necessary machinery did not exist.

Very early in the commercialization process, it was realized that almost all applications would involve compounding. It was estimated that in at least 95% of the applications, the styrenic block copolymer makes up less than 50% of the final product. This state of affairs is not unusual in the rubber industry, but is uncommon in plastics. The company and its customers had to develop the necessary compounding expertise, both in the selection of equipment and ingredients. For applications in which the material is molded as a thermoplastic fully compounded products based on S-B-S block copolymers were developed and manufactured. For applications in adhesives and polymer modification, a knowledge of the general compounding principles that were shared with customers was gained (refs. 22-24).

A tradename was also needed. Here, Tom Baron, a member of our upper management, came to the rescue. He consulted a book on Greek mythology and in it found a god of strength named Kratos (the son of Uranus and Gaia). He simply changed the last letter to obtain the tradename Kraton.

Later developments

Both polybutadiene and polyisoprene have many in-chain double bonds, almost one per original monomer molecule. This limits the stability of the S-B-S and S-I-S block copolymers. A very significant advance was the development of fully saturated block copolymers, made by hydrogenation of the elastomer segment. The first was an S-EB-S (where EB represents a poly[ethylene-co-butylene] segment). Later, similar S-EP-S versions were also produced. Functionalized S-EB-S block copolymers containing pendant maleic acid/anhydride groups are a more recent development.

It was eventually realized that triblock/diblock blends (e.g., S-B-S/S-B) have some advantages over pure triblocks in such applications as adhesives and sealants. Blends with high diblock content are now made commercially. Branched block copolymers were also developed. One, an (S-I)nX, is intended for radiation cross-linking after it has been applied as a hot melt adhesive (ref. 25). Another newer application is the use of S-B-S and S-EB-S polymers in blends with asphalt. These are used in roofing and road surfacing. A very different ap-plication that grew out of this work is the use of S-EP all-blocks as viscosity index improvers in motor oils (ref. 26).

These polymers also simulated some excellent theoretical work. The thermodynamics of domain formation (the order-disorder transition), the role of the domains as quasi-reinforcing fillers, the effects of composition on morphology and properties, and the effects of inter-chain entanglements are among the topics still being studied (ref. 2).

Conclusions

Since their commercialization thirty years ago, styrenic block copolymers have shown remarkable growth, and continue to have about 50% of the total market for thermoplastic elastomers. They have been a very satisfactory and profitable venture for Shell. At least twelve other companies make them worldwide, now that the original patents have long expired. This success, was the result of a flexible -- even opportunistic -- approach to research and development, combined with a willingness to take reasonable commercial risks. A large company can (and in this case did) display this entrepreneurial spirit, but only if its upper management has the vision and patience to support it. This is not always easy, but effective research can only be conducted if the following requirements are met:

* The parent organization must show a consistent commitment to research and development (and in particular, to the financial support for it). This commitment must not fluctuate with the current profitability of the business or the state of the economy.

* The organization must offer stable conditions and long term career opportunities to its members. People cannot be expected to perform at their best if they live in constant fear of being arbitrarily fired (or to use the current euphemism, "downsized"). Nor will they take an objective approach to the long term health of an organization that they expect to leave shortly.

* The organization must be willing to take reasonable risks, including the risk of failure. It must recognize that a commercial or scientific failure is not a reflection on the quality of the work that led up to it. Failures should not be followed by recriminations, demotions and firings. The old adage -- that a man who never makes a mistake will never make anything else -- is still true.

* It is clear that many discoveries -- most notably, that of the American continent -- are made by people who are really looking for something else. By its nature, all true research (including industrial research) involves uncertainty and the outcome cannot be foreseen with any degree of accuracy. Indeed, there is little point in carrying out experiments whose results are already known. The program must often stray from the planned path as new developments suggest more promising areas to explore. Thus, the work must not be too closely controlled or too closely tied to solving immediate problems or focused on immediate profit. Management has to accept that research is sometimes driven more by intellectual curiosity than by obvious commercial applications. However, its value depends far more on its quality than on its motivation.

References

(1.) M.S. Reich, Chemical and Engineering News, August 5, 1996, p. 14.

(2a.) Thermoplastic elastomers -- a comprehensive review (N.R. Legge, G. Holden and H.E. Schroeder, eds.), Hanser & Oxford Univ. Press -- Munich/New York, (1987).

(2b.) Thermoplastic elastomers -- a comprehensive review, 2nd ed. (G. Holden, N.R. Legge, R.P. Quirk and H.E. Schroeder, eds.), Hanser & Hanser/Gardner -- Munich/Vienna/New York/ Cincinnati, (1996).

(3.) Handbook of thermoplastic elastomers, 2nd ed. (B.M. Walker & C.P. Rader, eds.), Van Nostrand Reinhold, New York, 1988.

(4.) N.R. Legge, "Thermoplastic elastomers," Charles Goodyear Medal Address, presented at the ACS Rubber Div. Meeting, Montreal, Quebec. May 26, 1987. Rubber Chem, Technol. 60, G83 (1987).

(5.) S.E. Home, Jr., et al, Ind. Eng. Chem. 48, 784 (1956).

(6.) F.W. Stavely, et al, Ind. Eng. Chem. 48, 778(1956).

(7.) L.E. Foreman, "Polymer chemistry of synthetic elastomers," part II, Kennedy & Tornquist, Eds., John Wiley & Sons, New York, p. 497 (1969).

(8.) L.M. Porter (to Shell Oil Co.), U.S. Patent 3,149,182, filed 1957 (1964).

(9.) M. Szwarc, M. Levy and R. Milkovich, J. Am. Chem. Soc. 78, 2656 (1956).

(10.) M. Szwarc, Nature 178, 1168 (1956). See also A.F. Halasa, Rubber Chem. Technol. 54, 627 (1981) and M. Szwarc, Polym. Prep. 26 (1), 198 (1985).

(11.) G. Holden, J. Appl. Polym, Sci., 9, 2911 (1965).

(12.) J.T. Graver and G. Krause, J. Polym. Sci. Pt. A, 2, 797 (1964).

(13.) C.S. Schollenberger, H. Scott and G. R. Moore, paper presented at the ACS Rubber Div. Mtg., Sept. 13, 1957, Rubber World, 137 549 (1958), Rubber Chem. Tech. 35 742 (1962).

(14.) C. Holden, E.T. Bishop and N.R. Legge, "Thermoplastic elastomers, Proc. International Rubber Conference, 1967," MacLaren and Sons, London, p. 287 (1968), J. Polym. Sci., Pt. C 26, 37 (1969).

(15.) S.L. Cooper and A.V. Tobolsky, J. Appl. Polym. Sci. 10, 1837 (1966).

(16.) S.L. Cooper and A.V. Tobolsky, Textile Res. J. 36, 800 (1966).

(17.) W.H. Charch and J.C. Shivers, Textile Res. J. 29, 536 (1959).

(18.) J.T. Bailey, E.T. Bishop, W.R. Hendricks, G. Holden and N.R. Legge "Thermoplastic elastomers," presented at a meeting of the ACS Rubber Division, October 22, 1965; Rubber Age 1966, Oct. p. 69.

(19.) J. Polym. Sci. Pt. C26, 59(1969).

(20.) T.S. Kuhn, "The structure of scientific revolutions," University of Chicago Press, Chicago (1970).

(21.) G. Holden and R. Milkovich, (to Shell Oil Co.), U.S. Patent 3,265,765,filed Jan. 1962 (1966).

(22.) Technical Bulletin SC:198-92, Shell Chemical Co., Houston, TX, 1992.

(23.) D. J. St. Clair, Rubber Chem. Technol. 55, 208 (1982).

(24.) Technical Bulletin SC:165-88, Shell Chemical Co., Houston, TX, 1985.

(25.) Technical Bulletin SC:889-86, Shell Chemical Co., Houston, TX, 1985.

(26.) Technical Bulletin SC:1064-89, Shell Chemical Co., Houston, TX, 1989.