Paul J. Flory Stanford University Stanford, California 94305
Macromolecules Vis-a-vis the Traditions of Chemistry
Polymeric substances are distinguished at the molecular level from other forms of matter by the concatenation of atoms or groups in chain-like sequences, often of great length. The structural units that identify various polymers, natural and synthetic, include the gamut of possibilities in chemical constitution afforded by methods of synthesis, in uiuo and in uitro. Yet, the structural motif common to all of them is the sequential connection of units, of whatever description, in chains. Linear polymers consisting of a single sequence or chain command foremost attention. However, branched and cross-linked structures, and even three-dimensional networks, are included as well. Inasmuch as points of branching or of cross linking generally are separated by fairly long linear sequences of units, the covalently linked chain remains the characteristic structural feature even in so-called nonlinear polymers. Consistent with this generalization is the fact that covalently bonded two- and three-dimensional arrays typified by graphite and diamond are customarily excluded from consideration as polymers. Reflecting on the ubiquity of macromolecular materials on the one hand and the scant attention given them in the undergraduate chemistry curriculum on the other, an inquisitive student might raise questions like the following: Does the type of structure of matter occurring in polymers depend upon some principle alien to the body of chemical science? Is this broad class of substances, collectively, an aberration extraneous to the central bodv of facts and principles comprising chemistry? Or, is the rational interret tat ion of macromolecules so difficult. intrinsically, that they are intentionally excluded from the fare for indergraduates? On the contrary, can they be so readily comprehended by straightforward deduction from laws and rules laid down for small molecules as to obviate anything more than superficial mention? Do the manifold applications of polymers, in industry and in biology, disqualify them as objects of pure scientific inquiry? In response to questions such as these, it could he ~ o i n t e dout, first, that the occurrence of macromolecular bubstances is implicit in the rules of valency and chemical comhination. At a most rudimentary level, the very idea of multivalency and the capacity of certain kinds of atoms -notably carbon, sulfur, nitrogen, oxygen, silicon, and phosphorus-to enter into sequential combinations raises the possibility of chain structures. Simple examples illustrating manifestations of this possibility are commonplace: the n-alkanes, elastic sulfur, and polypeptides. If in fact multivalent atoms refused to enter into sequential combinations, or if this propensity did not manifest itself in the occurrence of macromolecules beyond some finite limit in length, it would be obligatory to introduce a qualification to this effect early in the exposition of conventional chemical doama. Thus, given the rules of valencv governing chemicai combination, if long chain molecule^ com~risinetens or even hundreds of thousands of skeletal atoms linged in sequence did not occur, then the chemical rule forbidding their existence would need to be set forth. Clearly, such a rule does not exist. Hence, if we are to ignore the implications of multivalency and consecutive bonding of atoms as manifested abundantly amongst natural and synthetic polymers, then the rationale behind ex732
/ Journal of Chemical Education
clusion of these suhstances should he explicitly made known. Logically, then, the occurrence of polymers is implicit in the elementary principles of chemical bonding. It is noteworthy that the chemical bonds in polymers are equivalent quantitatively to corresponding bonds, similarly situated, in small molecules. Within limits of measureability, their energies are identical. The same holds for bond lengths, bond angles, and force constants. Constitutive properties of macromolecules can therefore be formulated from the same information as for small molecules. The two regimes are continuous; no definable boundary separates one from the other. So much for the logical foundations of the area of science having to do with polymers, and its coherence with chemistry as a whole-if we regard chemistry as the science that deals with matter a t the molecular level. It is of interest to examine the historical record in regard to the foregoing assertions to the effect that the occurrence of macromolecular forms of matter is implicit in elementary ideas concerning valency ( I , 2). The concept of valency took form in the work of Frankland around 1852. Quadrivalency of carbon was enunciated by Kekul6 and by Kolbe in 1857. Kekul6 pointed out in 1858 that carbon has the capacity to use one of its valences to combine with another carbon atom, and thus to build chains. His textbook of 1861 has been said to mark the heginning of carbon chain theory. His celebrated ring formula for benzene appeared four years later. The graphic formula, that particularly facile device by which valences are represented by lines drawn between symbols for atoms, was introduced by Couper in 1858. It was adopted shortly thereafter by others, including CrumBrown (1861), Wurtz (1864) and, with variations, by Kekul6 and Loschmidt (1861). The graphic formula a t first was primarily a topological device denoting connectivity between atoms, and especially the integrity of groups which was a foremost concern of chemists of that era. Its geometrical significance in defining the spatial relation between atoms emerged gradually, and bas continued to evolve up to the present in stereochemistry, structural chemistry, and conformational theory. The implications of multivalency and its embodiment in the graphic formula are readily apparent in retrospect. To what degree were they evident to chemists at the time those important concepts emerged? Answers to questions of this kind must be sought by reading between lines and therefore are to some extent conjectural. The idea of polycondensation and its application to yield polymers of a degree n as expressed in formulas such as H(OCH2CH2)n-OH and in H-(OCsH&O),-OH, are recorded in the period 1860-70. Other condensation polymers were prepared, and while their molecular weights doubtlessly were low by present standards, their basic linear chain structure was clearly recognized in a number of instances ( 3 ) . However, the foundations of organic chemistry were not sufficiently secure a t that time to sustain the extension of them to polymers of high molecular weight. Moreover, physical methods for measuring molecular weights were unknown before 1885. I t is not surprising therefore that possibilities in this direction were not vigor-
ously pursued in the era that marked the dawn of structural chemistry. Interest in the synthesis and investigation of high molecular weight compounds all but vanished from the scene for the next forty to fifty years, due in large measure to two independent trends. First, the molecule and its representation by a concise graphical (or structural) formula became the prime object of all effort. By tacit assumption every pure substance worthy of investigation could he represented uniquely by such a formula. It was incumbent upon the investigator to specify a formula devoid of amhieuitv for even,. (oure) " ~, .. . chemical substance. Polymers Kenprally comprise mixtures of homulogous species und hence are described hv mol~culardistributions. Moreover, a long chain must bear end groups, determination of which would have overtaxed available analytical methods of the times. The formula could not be completed without specifying them. For these reasons, polymers could not he fitted easily into the accepted scheme. Secondly, colloid chemists of the era asserted that all colloidal substances are aggregates of smaller species. Their judgments went unchallenged in most quarters. Since polymers, both natural and synthetic, are colloids according to the quite proper definitions introduced earlier by Graham, they too must he aggregates. Some, such as W. Ostwald, relegated the molecule to a strictly suhordinate role in the hierarchy of structural entities at the suhmicroscopic level. In this respect, his views and those of his numerous following were diametrically opposed to the outlook evolving in chemistry as a whole. However, a larger conflict was avoided by tacit limitation of each view to its own domain: molecules for low molecular suhstances and colloids for large ones. By the end of the nineteenth century a considerable number of polymers actually had been prepared, often inadvertently, by bifunctional condensation. That the products were in fact polymers is clear from the properties reported. With rare exceptions, however, they were represented as cyclic monomers or dimers. Aggregation was invoked to explain their "colloidal" properties. (3). Although a numher of vinyl polymers had been stumbled upon in the course of investigations of their monomers (some of them before 1850), their constitution was unknown. Like the imagined cyclic molecules above, they were regarded as colloidal aggregates held together hy mysterious "secondary valences." The elucidation of naturally occurring polymers, viz., ruhher, rarbohydrates and proteins, presented morc suhstnnti\.e dilfirulrie~.In the period 1870-1900 earh of thew classes of pol\mers was the sut~jectof intensive in\x!;tiptions, conducted on the respective classes by separate groups of investigators working quite independently. The complexities of the repeating units comprising these polymers resisted determination until after 1900, and in this circumstance a definitive grasp of their constitution was precluded. The relationship of isoprene to ruhher and of amino acids to proteins had long been known, but the structures of the corresponding units in the respective polymers had first to be established, the former through the investigations of Harries (1904-14) and the latter through the work and inspiration of many investigators, notably Fischer (-1906) and Hofmeister (l902), cf. seq. The structural units of carbohydrates were even more recondite because of their greater complexity and their stereochemical ambiguities. The structure of the anhydroglucose unit C6H1005 was not fully established until 1928 (Haworth). Yet, the role of this unit in starch and in cellulose was recognized, and even during the 1880's formulas (C~HIOOJ),where n = 100 were suggested for "amylodext ~ s from " starch. Whether or not the units were covalently bonded was a question which did not require an explicit answer a t the time, hut i t seems clear that some authors, at least, had molecules in mind when presenting
their formulas, often supported by primitive determinations of molecular weights. All this was changed with the ascendency of the colloid apostasy in the early twentieth century. Natural polymers, like their more obscure synthetic analogs, were declared to he physical aggregates of smaller molecules held together by secondary forces, of origins unidentified. Not until the 1920's and 1-ter did their molecular nature come to he recognized ( 3 ) . The history of concepts concerning the nature of proteins, vividly portrayed by Edsall (41, offers an especially interesting paradigm of the evolution of ideas concerning macromolecular suhstances. The number of proteins isolated, purified, and crystallized before 1900 is truly impressive. In 1885 Zinnoffsky arrived at a (minimum) molecular weight of 16,730 for horse hemoglobin on the basis of careful analysis of its content of iron. Colligative methods for determination of molecular weights, originating in the work of Raoult (-1885). were a .~.o l i e dto ovalbumin. hemoglobin, and serum albumins during the succeeding 20 vears. Molecular weiehts of the order of 15.000 or -areat" er were indicated, hut the results came to he discounted on the grounds that the solutes were colloidal and theref o e beyond the purview of the laws of physical chemistry applicable to solutions of "crystalloids." The polypeptide hypothesis was offered in 1902 by Hofmeister as the only plausible bask on which to explain the chemical properties of proteins (4). Fischer subscribed to the same view and, in its support, undertook the task of synthesizing polypeptides. His preparation of a molecularly monodisperse polypeptide having a degree of polymerization of 30,stands as a monumental achievement. He did not, however, envisage polymers of greater chain length in proteins. In keeping with the outlook of the times, he preferred instead to regard proteins as aggregates of shorter polypeptide chains. Even apart from the questions of molecular weight and of secondary aggregation of short chains to make up the protein particle, the polypeptide chemical strllcture was not universally accepted for nearly 30 years after its inception. A number of chemical combinations were proposed for the primary ingredient of proteins by organic chemists during the mid-1920's in a frenzied effort to find an alternative to the polypeptide (i.e., polymeric) hypothesis. These alternatives were cyclic compounds of various sorts (diketopiperazines, a110 forms and other more esoteric heterocycles). They were presumed to be joined by aggregation in some mysterious way not specified. Finally, a plea for sanity in recognizing elementary facts of protein chemistry was eloquently voiced by Vickery and Osborne (5) in 1928. The polypeptide theory gradually gained ascendencv thereafter. the remenance to chemists of . polymeric~tructuresin substances as intimate as proteins notwithstanding. Even Vickerv and Oshorne (5) (P. 419) quailed at the-specter of polypeptide chains ha&g as many as 500-600 units. This prospect, if proven to he correct, would in their view raise "difficulties." Although the nature of the difficulties foreseen was not made clear, they may have concerned the exact specification of the sequence of units in a chain of such length. The saga of polynucleotides is fascinating also, but will not be recounted here in detail. It must suffice to point out that these polymers, exceeding in chain length any other linear macromolecules characterized to date, were thought to he tetrameric until the mid 1940's. The fact that the Watson-Crick model for double-stranded DNA was put forward only a few years thereafter (1953) is significant. The publicity it has received seems to have ohscured preceding events that were essential precursors to this important advance. It would be incorrect to place most of the blame for the prolonged delay in the acceptance of the molecular viewpoint regarding polymers on the dogma of colloid chemistry around the turn of the century and during the folVolume 50, Number 1 1 . November 1973
/
733
lowing three decades. Probably of even greater importance was the reluctance of chemists to consider seriously macromolecular formulas and to face the associated obscurities of molecular distribution and end groups. Lack of the capabilities in earlier times to resolve these aspects posed genuine difficulties. The doctrines of a former era in colloid chemistry were eventually cast aside, hut a distant attitude toward macromolecules continues to he widesuread amonest chemists even today. 'The traditional aversion of chemists to forms of matter that cannot he represented by a unique formula doubtless recited ahove. Their is rooted in the historical preo'ccupation with small molecules during the latter part of the nineteenth century and the early part of this one is readily understandable, as discussed previously elsewhere (,3~) . ,The uoint that concerns us here is the fact that the character of chemistry, nourished by spectacular successes. eelled in this era. In particular, the one-to-one association between formula -and substance was firmly entrenched by the time chemistry became a discipline. Macromolecules are not easily accommodated in this scheme without basic revisions, as pointed out above. It is especially to he noted that the basic pattern of chemistw was established prior to the appreciation of the molecul~rnature of po~ymkrsand to faGulation of princ i ~ l e sand theories describing their hehavior. The field was sibdivided along lines that largely persist today. Polymers, being in eclipse at the time, did not find a place in the syllabus, apart from some desultory entries on colloids. In more recent times, the dominance of the deductive approach to chemical science has directed attention largely toward the simplest of molecular systems. The reductionist view that all matter, animate and inanimate, is governed by the same set of fundamental laws can scarcely he challenged. But the corollary usually inferred from it to the effect that a tractable conceptual framework for understanding and interpreting the hehavior of complex systems including biological-even sociolagical-ones can he deduced from these laws, is fallacious. In recent years much of chemistry (and other branches of natural science as well) have fallen under the spell of this misleading notion. To he sure, a great deal can he learned by investigation of the simplest systems, hut comprehension of the next level in the hierarchy of sciences cannot he achieved through straightforward processes of deduction alone. As Anderson (6) has pointed out, the same creative insights, generalizations, etc., are required at each leyel. Thus, it should not he assumed that full knowledge concerning simple molecules will pave the way for comprehension of macromolecules without further creative effort, formulations of new concepts with appropriate abstraction, etc. This is not to say that information gained in one domain cannot he applied in the other. The debt of polymer science to knowledge gained from the study of small molecules is very great. It is also true, however, that investigations conducted on macromolecular systems have enriched chemistry as a whole. Our understanding of chain reaction kinetics, of free radical and ionic reaction mechanisms, and of thermodynamics of solutions has been advanced and enlarged by investigations on polymers. As other papers of this symposium testify, a wealth of illustrative material is available from current knowledge of polymers in these areas and others including chemical statistics, stereochemistry, conformational hehavior, and statistical mechanics. As a consequence of the pervasiveness of the deductive viewpoint, chemistry has come to he regarded in some quarters, and even among some chemists, as a sub-discipline of physics on the one hand and the handmaiden of hioloev on the other. Close connections with both these disciplines and a deprndenry on physics do indrwl exisr. Culri\.ation oi the ronrinuum oi knowledge between physics and chemistry has certainly been to the henefit of 734
/ Journal of Chemical Education
chemistry. But to hold that all chemistry follows deductively from physics and dismiss the matter therewith is to overlook the central role of science in erecting constructs for representation of physical reality in terms rational to the human mind. Elementary textbooks of chemistry seem to have been influenced overwhelmingly by the reductionist doctrine, or rather by its converse, "deductivism." The current fashion is to develop the subject largely from a set of rules. These are often introduced by appeal to physics, and to atomic physics in particular. The connections with these fields should certainly he emphasized at the proper stage in the student's development, hut to present chemistry as a derivative science is to conceal the historical foundations and concentual framework of the science of molecules and molecularbehavior. This viewpoint could conceivably lead ultimatelv to denial of the rightful existence of chemistry as a separate discipline. An examination of elementary texts of physics and chemistry brings to light a striking difference in pedagogy. In the former, principal topics-heat, light, mechanics, etc.-are introduced with a brief discourse on common observations and experience related rher~to.The srope of rommonplaceexperienre may then he enlarged 1)). preientation of observations under idealized conditions. i.e.. hv citation of experiments, or by prrftmnunce of demonstrations. This is followed hv zeneralizatiun and d ~ r r l w m e n of t concepts with some degree of abstraction. Most df the subjects treated in elementary chemistry texts, on the other hand, are removed from common experience. The traditional choice of subject matter offers little opportunity to commence with examples familiar to the beginning student. It is almost as if chemistry and its chemicals are from another world, i.e., from the occult world of the laboratory. Perhaps cognizance of the most commonplace materials of man's environment, indeed of man himself, if introduced appropriately as a part of the scientific scenario, would contribute a welcome note of relevance. With large numbers of students inclined to enter the biological sciences and medicine at the present time, the importance of connections between polymer science and biology should not he overlooked. This interdisciplinary area is grossly underdeveloped. It seems self evident that polymer science should furnish the medium for interrelating chemistry and hiology. The gap between these disciplines cannot he effectively bridged without recourse to a rigorous conceptual framework for dealing with macromolecules in general. The often implied view that hiopolymers are suhjects apart from the tainted materials of industrv. r as currentlv ,. and hence removed from ~ . o l.v m e science misunderstood rests on supposirions and regrertable misronrentimx. \Vohler's swtheais oiurea a rrnturv and a halr ago d;d not destroy the'hoctrine of Vitalism! ~ n f a c tbiopo, lvmers manifest all of the properties of synthetic polymers i f less refined constitution; i o he sure,-they also &hibit some important additional features for which they have been specially equipped through eons of evolution. Nevertheless, they share much in common with "non-living" macromolecules. The frequent assertion that macromolecules are intrinsically too complicated, and therefore too difficult, to he comprehended by undergraduates can he controverted on either of two counts. First, and most directly, a wealth of topical material is now available which, when put to test, demonstrates the contrary to he true. Secondly, the judgment of difficulty is necessarily subjective, being conditioned by one's point of view and hence by the state of organization of knowledge (and one's familiarity with it) in the given field. It is hardly an intrinsic characteristic of the given area of interest. Complicated suhjects have repeatedly been made simple by improvements in method and in point of view. ~~
~~
~
~
~
~
~
~
Conclusion I have endeavored to direct attention to the coherence of the scientific aspects of macromolecules with the chemistry of small molecules, and to the close connections between the two domains. These observations, augmented perhaps by the key importance of polymers both in biology and in industry, should support cogent arguments for accommodating macromolecular science within the discipline of chemistry, and hence for incorporating the subject into the core curriculum. However, the content of a discipline and its associated curriculum may more oflen he determined by precedent, i.e., by historical accident, than by logical considerations. Hence, I do not offer these remarks with any substantial expectation that the ideas they are meant to express will have much influence on the future character of chemistry, even if a consensus were to grant them validity. If, despite these forehodings, and for reasons unforeseen, macromolecules were to be brought into the chemistry curriculum in the near future, then a vigorous plea should be entered that the treatment due them he integrated into the logical exposition of the discipline. As remarked earlier, the occurrence of macromolecules should he introduced as the embodiment of the principle of multiple valency and consecutive bonding. This aim should be met irrespective of the level of sophistication adopted for treatment of the chemical bond. The subject of macromolecules should not, ns a t present, be introduced through an awkwardly inserted chapter late in the text. The opportunity to develop the suhject logically as a part ~~~
~
~~
~~
of the whole having been forfeited, the introduction of the subject so belatedly must perforce be justified by scientifically extraneous remarks on the technical importance of polymers, replete with statistics in tons per annum. Better to ignore the subject altogether than to offer it on these grounds in a course purporting to he science. By this I do not mean to deprecate the importance of the technological ramifications of the subject. These do indeed enhance interest in, and the significance of, the study of polymer science, and properly so. It is my opinion rather that the subject should he treated first from the detached standpoint of pure science, and preferably as an integral part of molecular science (chemistry) of which it is logically a part. One may subscribe unreservedlv to the view that motivations and content in pure science-are unrelated to practical applications. I t does not follow, as often is assumed, that a subject of practical relevance is automatically excluded from pure science and ineligible for abstract inquiry. Accordingly, i t is my view that polymers should be introduced in the context of the constitution of matter a t the molecular level, or not at all. Literature Cited (1) Psnmgton. J. R., "A Shoe Hiatow of Chemistw," 3rd Ed.. MacMillan. London. 19Si. (2) Lcicester, H.M., .'The Histotical Backgmund of Chemi~fry."John Wiley and Sons. New York 19FR. (3) Flow, P. J., "Principles of Polymer Chemistry: Carnell University Press. Ithsca, NewYork, 1953. Chsp.1. (4) Edsall,J.T.,Arth.Biorhem. Biophy$., Sup& 11.12119621. (51 Viekery, H. B.. and Osbarne, T.B.,PhyaioL Reos.. 8.398 119281. (6)Anderson. P. W..Scirncr. 177.393 (1972).
~~.~~~~~
Volume 50, Number 7 1 . November 1973
/
735
