Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

View Article Online / Journal Homepage / Table of Contents for this issue

POLYMERS AND POLYFUNCTIONALITY.

H. CAROTHERS.
BY WALLACE
Received 12th July, 1935.

1. Definitions.
The importance of polymerisations is indicated by the title of the
present program, and it is a curious fact that little agreement exists
concerning the accepted meaning of this term. Text-books, dictionaries,
and even recent publications by original investigators,l generally state
that a monomer and its polymer have identical compositions, and a
corollary is that polymerisation consists in pure self-addition, and is
peculiar to unsaturated compounds. But it appears that not one of these
conditions is necessarily satisfied by reactions that are universally re-
cognised to be polymerisations. Moreover there is little relation between
actual usage and text-book definitions.
An accepted polymerisation is the transformation of formaldehyde
into polyoxymethylene. The process under some conditions perhaps is
x CH20 --+ (CH,O),,
and under others
x CH,=O + H2O + HO-(CH2--O),-H.
In the latter case for moderately small values of x the composition of the
polymer will be very different from that of formaldehyde : a t large values
of x the difference may be questionably detectable, but the question, of
course, is one of fact not definition.
Turning to the mechanism of the process, there are various con-
ceivable possibilities, such as :
I . The polymerisation of perfectly pure and dry formaldehyde might
involve opening of the carbonyl bond giving free radicals -CH2-0-
whose mutual combination would result in a long chain.
2. In the presence of traces of water, we might have first

HOH + CH,=O + HO-CH,-OH,

a manifestation of formaldehyde’s very strong tendency to add ROH.
The product here is also ROH, and reaction therefore proceeds
HO-CH2-OH + CH,=O + HO-CH2-O-CH2-OH,
and further similar steps finally lead to a very long m ~ l e c u l e . ~
3. In aqueous solutions formaldehyde is largely present as hydrates
HOCH,OH, HOCH,OCH,OH, etc.
E.g., Staudinger, D i e hochmolekularen organischen Verbindungen, Julius
Springer, Berlin, 1932, p. 10.
[Link], loc. cit., p. 255.
Cf. Staudinger, ref. 1, p. 10.
39
 View Article Online

40 POLYMERS AND POLYFUNCTIONALITY

Polymer is formed by adding strong acid to such solutions. The reaction

might be
Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

-+
HOCH,O]~~CH,OI~]CH,O H HOCHzOCHzOCH20-
ie., an intermolecular condensation or dehydration.
Further peculiarities are illustrated by polyesters where all of the
following transformations can be realised in one case or a n ~ t h e r . ~

,*+I o-R-co I
L t c
A
HORCOOH\
Heat 1
. . . -O-R-CO-O-R-CO-O-R-CO-O-R-CO-. ..
B
In many cases the acid can be converted into a polyester under condi-
tions where the monomeric lactone C is known to be stable. Moreover,
it is then often possible by titration to show that a terminal carboxyl
group is present in the ester product in the expected amount. The
reaction involved is clearly one of condensation (not self-addition of an
unsaturated or cyclic intermediate). Nevertheless, it seems scarcely
reasonable or practical to argue that the product is not a polymer. For
one thing, such polyesters can generally be transformed under appropriate
conditions to the corresponding cyclic monomers (lactones), and the latter
can be reconverted to products practically indistinguishable from B.
In this case the reaction must consist essentially in self-addition. We can
hardly say that a compound of the type B is a polymer if formed by one
method, but not if formed by another. Moreover, there are some cases
(e.g., 6-lactones and lactide) where transformation between B and C
occurs so readily that it is difficult or impossible to know which one is
the product first formed from the acid. Similar confusion occurs with
many reactions involving formaldehyde : it is often impossible to know
whether products arise by direct intermolecular dehydration of a methylol
compound or by self-addition of an intermediate methylene compound.
Further illustrations might be offered, but these perhaps are sufficient
to show that : the composition of high polymers is not necessarily iden-
tical with that of the supposed parent monomers ; polymerisations that
appear superficially to involve only self-addition may involve prelim-
inary hetero additions, and then actually proceed as condensations (or
the reverse may be true) ; polymerisations are in any event not peculiar
to unsaturated compounds (keeping in mind, for example, cyclic anhy-
drides, esters, acetals, and certain ethers and imines).
It is obvious that definitions need to be revised. It will not suffice
to insist merely that usage is frequently wrong. The unfortunate effsct
of existing discrepancies becomes especially apparent when one reads
that a resinification reaction (between an aldehyde and a phenol) doubt-
less proceeds through an unsaturated monomeric intermediate, because
polymerisations are peculiar to unsaturated compounds.
Polymerisations are a special class of reactions, but just how are they
distinguished from other intermolecular reactions ? Conventionally,
they involve the mutual combination of a number of similar molecules,
Carothers, Dorough, and Van Natta, J. A m , Chem. SOC.,1932, 54, 761 ;
Carothers and Hill, ibid., 1559 ; Hill and Carothers, ibid., 1933, 55, 5031. 5043 ;
Spanagel and Carothers, ibid., 1935, 57, 929.
 View Article Online

W. H. CAROTHERS 41

but this is not sufficient : the formation of diethyl ether is not polymerisa-
tion, and besides we already find references to mixed or hetero-poly-
merisations where the participating molecules are not all alike.
Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

In the writer’s opinion, the essential peculiarity of polymerisations is

realised in the statement that they are intermolecular combinations
(conventionally, self-combinations) that are functionally capable of
proceeding indefinitely (or leading to molecules of infinite size). This
may, in fact, be taken as a definition. By way of illustration, referring
to the formaldehyde reactions outlined above, it is obvious that, regard-
less of which of the suggested mechanisms if any may be correct, the
functional possibilities are such that any given number of formaldehyde
molecules might be combined into a single molecule. The same possi-
bility exists generally with
(a) unsaturated compounds,
(b) cyclic compounds and
(G) polyfunctional compounds generally as, for example, x-R-y where
x and y are capable of mutual reaction.
The compounds formally capable of polmerisation then are all poly-
functional compounds. Practically the functions must be such as to
permit mutual combination, and polymerisation will then fail only in
those relatively rare cases where reaction is exclusively intramolecular.
A double bond or a reactive ring will count as a double function.
We may note then that polymerisations do not involve a single
unique type of reactivity : they are for the most part merely ordinary
reactions made manifold by polyfunctionality, and thus made capable
of indefinite continuation in one, two, or three dimensions. But the
functions and the mode of their action are generally the same as those
already familiar in simple uni-functional compounds (though addition
polymerisations of unsaturated compounds certainly present some
special peculiarities).

2. The Peculiar Significance of Polymers ; Granular Polymers

and Polymerisations in Living Organisms.
The most important peculiarity of high polymers is that they alone
among organic materials manifest to a significant degree such mechanical
properties as strength, elasticity, toughness, pliability, and hardness6
Weight for weight cellulose and silk are stronger than steel; rubber
exhibits a combined strength and elastic extensibility that is not even
remotely approached by anything in the inorganic world while diamond
is harder than any other material. The practical uses of high polymers
depend almost entirely on these mechanical properties : our clothing
and furniture and much of our shelter are made of such materials. The
names cellulose, wool, rubber, and silk suggest a t once the great im-
portance of the non-chemical uses of natural high polymers.
Probably the bulk of the organic matter in living beings is made up
of high polymers. The necessity for this lies in the fact that living
organisms must have physical form and coherence, and polymers are the
only organic materials capable of supplying these properties. The
variability of living matter also requires a high degree of structural
5 Cf,Meyer and Mark, Der Aufbau der hochpolymeren organaschen Natztrstoffe,
Akademische Verlags gesellschaft, Leipzig, 1930.
2 *
 View Article Online

42 POLYMERS AND POLYFUNCTIONALITY

complexity, and the possibilities of high polymers in this connection are

indicated by Fischer's well-known calculation that 20 different amino
acids may form 2*3 X I O ~ *different polypeptides of 20 units. Another
Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

pertinent fact is that the physical properties of high polymers are pro-
foundly affected by their physical history : the melting-points of certain
polyesters can be reduced several degrees by the mere application of
stress and their strength in the direction of stress is a t the same time
increased many fold. Finally, reactions of polymerisation also appear
to be uniquely adapted to the chemistry of vital growth because they
are the only reactions that are capable of indefinite structural propaga-
tion in space.
Curious analogies with vital growth are found in granular polymerisa-
tions which have as yet received very little published attention, although
a few observations on chloroprene in this connection have already been
described.g
If we place a*fewgrams of liquid chloroprene in the bottom of a long
tube together with a little catechol to inhibit polymerisation, chill the
tube, pump i t out thoroughly and seal the top a t thejamp, the tube may
be allowed to stand upright for weeks without the appearance of any
change. But if we concentrate a strong beam of light on a small spot
a t the top of the tube, a minute speck of white solid will appear within
a short time. The light may then be removed : the speck will grow to a
cauliflower-like mass a t the expense of the vapours and thence of the
liquid monomer until the latter is exhausted. The polymer on super-
ficial inspection seems to be a mass of globules or cellular crystals, but
X-ray examination reveals no crystalline pattern. The growth of the
mass apparently occurs a t different rates in different directions, Some-
times lateral growth is more vigorous than'vertical, and the mass will
then burst the walls of a very heavy tube, although the total volume of
polymer is only a fraction of the available space. Any fragment of the
granular mass when placed in fresh chloroprene liquid or vapour will act
as seed around which further growth occurs. The viability of the mass
as seed will usually persist on standing in laboratory air for several days,
but finally it is lost. Similar observations can be made with butadiene
and other materials. The analogies of this process with d t a l growth
are obvious. The process may be labelled heterogeneous autocatalysis,
but the mechanism of the catalytic effect is obscure. Reactive centres
in the sense of free radicals are scarcely admissible, since the viability
persists in air, and for various reasons neither adsorption nor solution
seems to offer an adequate explanation.
Many different types of synthetic polymers can be made, but no
naturally occurring polymerisation has yet been exactly simulated. Why
should we not be able to synthesize cellulose, or proteins, or rubber?
We can make innumerable members of each of these types : polyacetals,
polyamides, and polyprenes. We can even make polyprenes that equal
natural rubber in strength and elastic extensibility and are superior to it
in many other respects, but the method (ie., the polymerisation of
chloroprene) is as yet subject only to empirical methods of control and
not to rational methods based on completely elucidated theoretical
principles. Besides, this goal has not yet been approached starting from
isoprene, and it is not at all certain that nature builds up rubber directly
from isoprene : she might, for example, start with methyl-n-propyl
6 Carothers, Williams, Collins and Kirby, J. A m . Chem. SOC.,1931, 53, 4214.
 View Article Online

W. H. CAROTHERS 43
ketone and carry out a selective condensation polymerisation completely
evading the diene intermediate.
Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

- C = l ~ 2 1 C H C H 2 C H 2 C \ ~ l C H C H 2 C H 2 etc.
Cl~
I I- I-
CH3 CH3 CH3
The synthesis of proteins raises a different set of questions. ct-amino
acids (or their esters) can be polymerised directly in vitro, but the
product is usually the cyclic dimer (diketopiperazine). But in the
organism, cyclisation apparently does not occur : reaction is exclusively
intermolecular and a linear polyamide results. Generally speaking,
cyclisation and chain formation represent two paths, the choice between
which is determined almost completely by the nature of the initial
reactant (especially its unit length).'
The conditions in vitro presumably are the normal ones of chance.
If both reactions follow the same mechanism, we can say that in a com-
pound of the type NH,CH,CONHCH,COOH, intramolecular approach
of the terminal groups is more frequent than intermolecular (which
means cyclisation in vitro), but in the organism the probabilities are
reversed and chain formation occurs. How can this b e ? Temperature
and moderate dilution have small control over such matters, and perhaps
then in the wrong direction. But if reaction is preceded by adsorption
a t an interface, as it might be biologically, the molecule is no longer free
to assume its spatially probable configuration. Its head, tail, and middle
are fastened to a surface and the only terminal approaches then possible
may be intermolecular. This picture is not here proposed as a solution
of the mechanism of protein synthesis: it is introduced rather as an
interestingly conceivable possibility, but especially for the purpose of
emphasising the following points : ring formation and chain formation
are the two alternatively possible results of every bifunctional reaction ;
complete control over these alternatives is possible only by bringing the
ends of the molecules together or separating them a t will; the only
apparent prospect of achieving this result lies in the use of surface
forces ; reactions in vivo probably occur largely a t surfaces ; and in any
event the effect of surfaces on bifunctional reactions presents an almost
completely unexplored field (except for the interesting preliminary work
of Freundlich and of Salomon).8
3. Polyfunctionality .
Polymerisations generally may be divided into those that are bi-
functional and those of higher orders. We may take an hydroxy acid
as an example of the bifunctional type :
HORCOOH + HORCOOH __.+ HORCOORCOOH + H20.
The two reacting molecules are monomers; the first product is a dimer,
since it contains two structural units. At this point, half of the initial
functional groups have disappeared ; similarly, it is evident that the
formation of two ester linkages (trimer) will involve 67 per cent. reaction,
a tetramer will correspond with 75 per cent. reaction, etc. In general, if
7 Carothers, J. A m . Chem. SOC.,1929, 51,2548 ; Hill and Carothers, ibid.,

1933,559 5033-
8 Helv. Chim. Acta, 1934, 17,88.
 View Article Online

44 POLYMERS AND POLYFUNCTIONALITY

p is the degree (fraction) of reaction and x is the average degree of
polymerisation,
Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

p = 1 - : .
X
The following values are further illustrative :-
p o 0.5 0.8 0.9 0.99 0'999
X I 2 5 I0 2O
0S g 5 I00 1000

The average molecular weight rises very steeply with increasing degree
of reaction beyond 0.95, and hence it is not impossible (as some authors
maintain) to obtain very large molecules from condensation polymerisa-
tions. If p should reach one the molecular weight would become
infinite.
As a hypothetical example of a more highly polyfunctional reaction,
we may take an acid R(COOH), and consider the result of intermolecular
anhydride formation. We assume that the structure and conditions
are such that no intramolecular reaction occurs a t any stage. Repre-
senting the acid by A4 the anhydride would be
. . . - A2 - A2 - A 2 - A 2 - A 2 - A ' - . . .,
etc., where the superscripts indicate that each unit bears two unreacted
carboxyls. Hence when all of the molecules have been combined into
one (and no further intermolecular reaction therefore can occur) only 50
per cent. of the initially present functional groups have disappeared, i.e.,
p = 0.5. The formation of branching chains of course does not affect
this conclusion.
A general equation relating degree of reaction, polymerisation, and
functionality can be developed.
Let f = degree of functionality (ie., number of functional
groups per monomer molecule).
N o = number of monomer molecules initially present.
Then No .f = number of functions initially present.
N = number of molecules after reaction has occurred.
2(Nn- N ) = number of functions lost.
2(Ni-
N , = fraction of functions lost = p = extent of reaction.
No *fAT
1v
Obviously, ;?;io = average degree of polymerisation = x.
Hence
(2) T-
2
X-=
f
2
p = degree of reaction.

This equation has interesting applications. If we heat together a

dibasic acid and a glycol, a linear polyester is formed. As the reaction,
measured by the disappearance of acid progresses, the average molecular
weight rises, presumably in accordance with equation (2). Reaction
becomes exceedingly slow after it is 99 per cent. complete, and the
calculated molecular weight approaches 15,000 to 20,000. But no
matter how far the reaction is carried under normal conditions, the
product is both fusible and soluble.
If in a similar manner a dibasic acid is heated with glycerol, one
observes a t first that the viscosity of the mixture changes only slowly as
reaction proceeds, and then it rises to infinity while much of the acid
 View Article Online

W. H. CAROTHERS 45
remains still unreacted. The product is completely infusible and
insoluble.
Obviously, bifunctional reactions can yield only linear polymers ;
Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

and there is evidence among such polymers that solubility and fusibility
are not lost even a t very high molecular weights. But reactions of
higher ordeis always present the possibility of spreading out in three
dimensions.
Returning to the anhydride symbolised above as
. . . - A2 - A2 - A2 - A2 - A2 - A2 - . . .
let us suppose that the chain is 100 units long, and bears 200 carboxyl
groups. Such a molecule could easily be soluble and fusible. If further
intermolecular reaction occurs, the chain must become cross-linked.
Obviously, when 2 carboxyls have been lost from each chain, all the
chains will be locked into a single molecule. This involves the loss of
I per cent. of the remaining carboxyls : and it will certainly be accom-
panied by the disappearance of fusibility and solubility: a gram of
matter cannot exhibit the kinetic behaviour of a single molecule.
Referring again to formula (2) if x is very large the second term dis-
appears, and we have p = 2 / f which tells a t what degree of reaction the
molecular weight will become infinite, or where, in polyfunctional re-
actions, gelation will occur and intermolecular reaction cease.
2
For a bifunctional reaction p = - - I, and since this value never can
f-
actually be reached, gelation will not occur. For a trifunctional reaction,
the limit will lie at p = 3, for a tetrafunctional reaction a t 8. We
may note in passing that for this reaction at x = 100,p = 49.5 per cent. ;
thus the average molecular weight will suddenly change from a moderate
to a colossal value with very little change in the extent of reaction : and
so far as the utility of the formula is concerned, it is not important
whether the transition in properties occurs at a molecular weight of 105
or 1020.
In attempting to apply formula ( 2 ) the chemistry of the situation
must, of course, be taken into account. A double bond will count as a
double function only if reaction is exclusively intermolecular a t every
stage. Multiple function compounds often do not exercise all their
functions together. In the formation of polyprenes from dienes, only
the I , 4-positions of the latter are at first called into play. The remaining
double bond of each unit functions under slightly different conditions to
cause cross-linking. Acetylene would be a tetra-functional compound
and probably behaves as such in the formation of cuprene. In the
formation of divinyl a ~ e t y l e n e CH,=CH-C
,~ =C-CH=CH,, only half
of its unsaturation is involved. Th-is material again in the early stages
of its thermal polymerisation behaves like a bifunctional compound
yielding, for example,1°
CH,=CH-C C-CH - CH2
I I
I I
CH2-CH-C =C-CH==:CH,
but gelation usually occurs long before all of the divinyl acetylene has
reacted.
Nieuwland, Calcott, Downing and Carter, J. Am. Chem. Soc., 1931~53,4197.
and Carothers, J. Am. Chem. Soc., 1934,56, 1167.
lo Cupery
 View Article Online

46 POLYMERS AND POLYFUNCTIONALITY

Perhaps a majority of the reactions to which equation (2) would be
applied are of the A-B type involving two reactants of complementary
function. I t is evident that if each reactant is bifunctional, A2B2, the
Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

reaction as a whole will be in effect bifunctional : A3B3is trifunctional,

etc. But if A and B have a different numb'er of equivalents per molecule,
the situation is more complicated. We can distinguish two different
possibilities :
(a)Suppose that on the average every coupled molecule of A involves
also the coupling of an equivalent number of molecules of B ; then the
degree of functionality is the average number of functions per molecule
of the two reactants when they are taken in equivalent amount. Thus
in the reaction of glycerol and phthalic anhydride (ASB2) we have to
take z moles of glycerol and 3 of the acid, or 5 altogether, containing
2
12 equivalents and f = '9" = 2.4. Then a t x = co, p = - and the
2'4
limit of the reaction'will be Q (83.3 per cent.). This, in fact, represents
the maximum amount of reaction that can occur before gelation under
any distribution of combinations, provided only, of course, that the
reaction is all [Link]
(b) Another extreme possibility is the following : Suppose the glycerol
behaves a t first as though i t were bifunctional yielding a chain
-A'- B - A'- B - A ' - B - A ' - B -
( ' indicates unreacted hydroxyl). Then all of the molecules of glycerol
could be combined into one with the loss of only 3 of the functional
groups (of course Q of the phthalic anhydride remains entirely un-
reacted). Here f = 3 and 66.7 per cent. is the minimum extent of
reaction compatible with gelation. Experimental values found by
various investigators for the extent of reaction at gelation generally lie
between 75 and 80 per cent.12
Reactions of polymerisation involving degrees of functionality run-
ning into the hundreds or thousands are undoubtedly of great import-
ance. The vulcanisation of rubber is an example. The preponderance
of evidence lies in favour of the view that vulcanisation is the result of
cross-linking through reaction at the double bonds. A conservative
estimate of the number of double bonds in the average rubber molecule
is 5000. Suppose the mechanism of vulcanisation is as follows: At
some double bond :C=C + S __+ :C - C: . Sulphides of this struc-
\ /
S
ture are known to be unstable, and further reaction might occur with a
similar group in a neighbouring molecule,
:c - c:
\ /
S :c - c:
+ :c - c: - /
S
\
S
\ / \ /
S :c - c:
11 A similar calculation for the glyptal reaction has already been made by
Bozza, Giorn. chim. ind. applicata, 1932, 14, 400.
12 Bozza, loc. cif.11; Kienle and Hovey, J . Am. Chem. Soc., 1929,51, 509.
 View Article Online

W. H. CAROTHERS 47
A t a minimum then all the rubber molecules could be locked together
when only 0.04 per cent. of the double bonds had disappeared, or
about 0-02 per cent. by weight of sulphur had reacted. A good estimate
Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

for a probable maximum is- difficult to make. Experimentally, the

minimum amount of sulphur required for the beginning of vulcanisation l3
is about 0.15 per cent. Staudinger and Heuer have shownI4 that as
little as 0.01 per cent. of divinyl benzene in styrene will lead to the
formation of an insoluble polymer.

4. Molecular Weights and Molecular Weight Distributions.

The peculiarities of high polymers are due to the fact that their
molecules are very large. Estimates of molecular weights are therefore
of the utmost importance, but they are very difficult to obtain. Direct
osmotic methods present great practical difficulties. Staudinger has
proposed to infer molecular weights from the viscosities of polymer
solutions. There appears to be no doubt that, other things being equal,
the viscosity of such solutions will increase with increasing molecular
weight, and viscosity methods are unquestionably of great value as a
means of providing rough estimates concerning the relative order of
magnitude of molecular weights in a given polymeric series. But the
absolute values inferred from viscosities are subject to a very large
factor of uncertainty. An initial difficulty lies in the fact that no
generally valid relation between viscosity and concentration in solutions
of linear polymers has yet been established. Presumably some simple
and general relation will be found to exist in solutions that are sufficiently
dilute but possibly only a t dilutions requiring a higher degree of experi-
mental precision than is now generally available. (In theory, of course,
extrapolation to infinite dilution is always possible, but practically such
extrapolations may also involve large uncertainties.)
Aside from this point, there is one theoretical complication that has
as yet received relatively little attention. High polymers are generally
mixtures covering a wide range of molecular species. The properties of
a polymeric mixture will depend not only upon the average molecular
weight, but on the manner in which the different molecular species are
distributed about the average. Moreover, as Kraemer and Lansing l5
have recently shown, it is possible to define various types of averages :
specifically, for example, there are the number average M, =
and the weight average M , = ZlfiMt, where fi is the fractional weight of
the constituent of molecular weight M i in the mixture, and the summation
is applied to all the constituents. The number average is the con-
ventional one: it is this that is inferred from osmotic methods or
chekical measurements of end groups. But all the proposed viscosity
relations would give M,.
If a material is homogeneous, M n and M , will be identical : but if
we are dealing with a mixture they will generally be different. If we
have 10 molecules of weight 100 and 5 of 1000,M , is 400 but M , is 851.
Thus a correlating factor for molecular weight vs. viscosity arrived a t by
l3 Bruni, Rev. gen. Caoutchouc, 1931,8, No. 75, 19 ; Stevens and Stevens,
J. [Link]. Ind., 1932,51,44T.
l4 Ber., 1934,67,1164. See also ibid., 1935,68,1618.
J . Physic. Chem., 1935,39, 165;J . A m . Chem. SOC.,1935,57,1369. See
also Kern, Ber., 1935,68, 1439,and Staudinger, Helv. Chim. A d a , 1929,12, 941.
Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

'c
48

L
*9 a m .
3
.

,
'

I
R
.
,
I
I
1
POLYMERS AND POLYFUNCTIONALITY

1
I
View Article Online
 View Article Online

W. H. CAROTHERS 19
that we are dealing with a wide distribution of species. Certain relations
are most briefly shown by means of an example.
If we assume that the weight of the mer or unit is IOO, and reaction
Published on 01 January 1936. Downloaded by National Tsing Hua University on 4/28/2024 [Link] AM.

has been carried to 98 per cent. of completion ( p = 0.98), then M , the

number average molecular weight is 5000, and the maximum in the
distribution curve also occurs a t this point. But the nature of the curve
is such that only 26 per cent. by weight of the material has a molecular
weight below 5000, while 74 per cent. has a weight at or above 5000.
Moreover, 40 per cent. of the material has a molecular weight above
10,000. Finally, the weight average molecular weight is 9900 or nearly
twice the number average.
Mechanical properties, such as toughness and tensile strength,
certainly require the presence of very large molecules. But it is not a t
all certain that these properties will be improved by the elimination from
a material of all fractions having low or moderately high molecular
weights. In rubber and cellulose plastics the deliberate addition of
small foreign molecules (plasticisers, lubricants, etc.) often has a very
favourable effect on properties. It is probable that average molecular
weights per se will furnish an entirely inadequate basis for the inter-
pretation of those properties which are chiefly of interest in dealing with
high polymers. The study of molecular weight distributions and their
relation to properties thus presents a field which should ultimately
be of great importance to the subject of polymerisation.