The Crystallization of Polyethylene

Terephthalate by Organic Liquids

W. R, MOOREand R. P. SHELDON

The development o[ crystallinity in essentially amorphous polyethylene

terephthalate film following immersion in a range o[ organic liquids at 25°C
ha~ been studied. The rate o[ development o[ crystallinity varies markedly
with the liquid and although molecular size o[ the liquid may be a contributory
[actor it is not the only one.~ Both the equitibrium degree of crystallinity and
swelling o[ the film vary with the liquid and in a similar manner and it would
seem that polarity, type and solubility parameter of the liquid are important
[actors governing both equilibrium crystalllnity and swelling. It is suggested
that solvation of the polymer occurs and that interaction o[ solvated polymer
and liquid leads to conditions [avourable to crystalliza(ion.

THE ability of some organic liquids to induce crystallization of a potentially

crystallizable but otherwise essentially amorphous polymer has been known
for some time but little quantitative information is available. Spence1, using
films of cellulose triacetate, found that the better the solvent the sharper
was the X-ray diagram, suggesting a greater degree of crystallinity. Baker
et al. 2 drew similm conclusions from work on cellulose triacetate and
tributyrate. Kolb and Izard 3 showed that some organic and inorganic liquids
would induce crystallization of polyethylene terephthalate at temperatures
well below those at which rates of crystallization are appreciable in the
absence of liquid. A recent study4 of the effects of a homologous series of
methyl ketones on the crystallization of polyethylene terephthalate suggested
that, if allowance were made for any absorbed liquid, apparent densities
could be corrected to true densities which were independent of the amount
of liquid absorbed. The rate of crystallization appeared to decrease with
increasing size of the ketone. There was some evidence that not only the
rate of crystallization but also the equilibrium degree of crystallinity might
depend on the nature of the liquid. In a further study of factors affecting
the crystallization of polyethylene terephth'alate in liquids this work has
been extended to cover a range of organic liquids of differing chemical types
and properties.

EXPERIMENTAL
The polyethylene terephthalate used was essentially amorphous film, 0.008
in. in thickness. Preliminary measurements showed it to have a density of
l'340g/cm ~ at 25°C, corresponding to a degree of crystallinity of 4-2 per
cent. The following liquids were used: hexane, carbon tetrachloride, ethanol,
n-butanol, benzyl alcohol, acrylonitrile, dioxan, ethyl formate, m-cresol,
aniline, nitromethane, nitroethane, acetic acid, dimethyl phthalate, benzene,
toluene, nonyl methyl ketone, n-amyl methyl ketone, ethyl methyl ketone
and acetone. The best available grades were further purified, dried by
appropriate agents and fractionally distilled before use.
315
 W. R. MOORE, A N D R. P. SHELDON

Known weights, approximately 0"07 g, of polymer were immersed in each

of the liquids at 25 ° ~+0'01 °C for varying periods of time. Surface liquid
was then removed by lightly blotting with filter paper and the sample stored
in an evacuated vacuum desiccator for at least 24 hours. It was then
weighed, all weighings being made with a semi-micro balance reading to
0-00001 g. With less volatile liquids such as dimethyl phthalate a surface
washing with ethanol preceded blotting. After reweighing the density of
the sample was determined by immersion in 10 cm3 of carbon tetrachloride
at 25 ° +0"01°C and addition, with agitation, of dry ethanol from a micro
burette until the sample neither rose nor sank in the mixture. The density
of the sample was then obtained from a graph showing the variation of
density with composition of the carbon tetrachloride-ethanol mixtures.
Preliminary experiments showed that such mixtures did not induce crystal-
lization, at least within 24 hours. The method was quite reproducible and
densities could be estimated to 0-001. With m-cresol some solution of the
polymer occurred and dissolved polymer was precipitated by addition of
ethanol and recovered by centrifuging and drying in a vacuum oven at 50°C,
the weight of recovered material being used to correct the apparent density
of the undissolved material to the true density. The influence of this pro-
cedure on the results will be discussed later.
Swelling of the polymer in the different liquids was obtained from the
weight of liquid absorbed when density had reached its equilibrium value
and expressed as a percentage of the original volume of the polymer. With
volatile liquids such as acetone and benzene weighings were made as a
function of time and followed by extrapolation to zero time to obtain the
true weight.
As pointed out previously* some liquid is retained by the polymer after
evacuation and the density da obtained by the flotation method is not that
of the polymer d~ but of the polymer plus retained liquid.
d, = (mp + m31(% + ~3
where m~ and ~ are the mass and volume of the polymer and mz and vz
those of the liquid. Assuming, from previous work" and direct measure-
ments of swelling in a density bottle, that the liquid is uncompressed, then
da = (roT,+ m3 / (m~/d~ + todd3
from which the density of the polymer d~ is given by
d,=d,/t(m,/rr~) (1 - d : / d3 + 11

RESULTS
Plots of density against time of immersion are shown in Figure 1. An
equilibrium density is reached with each liquid and these densities are
collected in Table 1 together with corresponding percentage crystaUinities
x~ obtained from
x,-- 100 (d~ -dA) / (de -dA)
where dA is the density of the completely amorphous polymer, taken5 to be
1-335 and de that of the completely crystalline, taken 6 as 1"455. Values for
316
 THE CRYSTALLIZATION OF POLYETHYLENE TEREPHTHALATE

C"E 1.42 Ethyl forma,te Acrylonitrile

-- s ~ Dioxan

,
12
I
Time
+
×o--e~
l
24
,
h
"t"

1'42[- Aniline Nitromethane Nitroethane

1"401~ o .:. ",, x ~ x-

1"38[~ ~ I ~ ~ ~o-
12 Time 24 h
1.40[- x ~ Benzene ~ Toluene

I.--I-"
D
1'34g I I I
1 Time 2 3 day
.x-
1"401- Benzyl alcohol ~x~ ~-'x" Dimethyl phthalate o-.

1'36i-~
1 " 3 M I I I I
0 I 2 3 4
Ti me day

Figure / - - R a t e of crystallization of polyethylene terephthalate in

various liquid media at 25 °C

Table 1
Equilibrium values of density, swelling and degree of crystallinity

Liquid 5250 Density Swelling Crystallinity

(cal/cma)~ (g/cc) (% vol.) (% vol.)

Hexane 7.3 1" 340 Nil 4' 2

Nonyl methyl ketone 7"9 1 "340 Nil 4"2
n-Amyl methyl ketone 8-45 1 "358 6"3 19"2
Carbon tetrachloride 8"6 l "340 Nil 4"2
Toluene 8"9 1"381 14'7 38-1
Benzene 9"15 I "390 18"8 45'8
Ethyl methyl ketone 9" 15 1 "395 18"7 50'0
Ethyl formate 9-40 1.390 20.0 45-8
Acetone 9"75 1 "405 30"8 58 "3
Dioxan I0"05 1 •396 21-0 50-8
Acrylonitri!e 10"5 1 "384 12-6 40"8
Dimethyl phthalate 10-6 1 "385 9"9 41 "7
Nitroethane 11"1 1-389 17-7 45-0
Benzyl alcohol 11-3 1-396 22 "3 50-8
n-Butanol 11 "4 1" 340 Nil 4' 2
Aniline 11 "5 1.400 36"8 54-2
m-Cresol 11 "9 1 "413 125 63
Nitromethane 12"6 1 "394 14"3 49-2
Acetic acid 12"9 1 -390 13' 1 45"8
Ethanol 13-2 1 "340 Nil 4"2

317
 W. R. M O O R E A N D R. P. SHELDON

the ketones are somewhat lower than those previously reported" and the
reason for this is not clear. As mentioned previously, some polymer, pre-
sumably of low molecular weight, was soluble in m-cresol. By taking into
account the weights of precipitated and undissolved polymer and also a
presumably proportional loss of m-cresol from the sample by extraction
with ethanol, corrections were applied to the density and swelling of the
undissolved polymer. Since the soluble fraction may behave differently
from the insoluble and very low molecular weight polymer may not be
precipitated by ethanol the corrected value given in Table 1 should be
regarded as approximate. The development of crystallinity was always
accompanied by increasing opacity of the film and, when absorbed solvent
was removed, by increased brittleness. The latter effect has been previously
noted 7 after prolonged heating in an aqueous dye bath at 100°C. Values of
equilibrium swelling are also given in Table 1.

DISCUSSION
Figure 1 shows that the rate of increase of density and hence of crystallinity
varies markedly with the liquid. In some cases crystallization is rapid while
in others it occurs at a rate convenient for measurement. Figure 1 also
shows that although molecular size may be a factor affecting the rate of
crystallization it is not the only one. Dioxan, ethyl formate, acrylonitrile,
m-cresol, aniline, nitromethane and nitroethane induce much more rapid
crystallization than benzene, toluene, acetic acid, benzyl alcohol and
dimethyl phthalate. Kinetic studies of this induced crystallization will be
discussed in a later communication and are not further considered here,
but it may be noted that the rate of disappearance of available amorphous
content at any given time seems to follow a first order equation.
The equilibrium values of density and crystallinity also vary with the
liquid. The non-polar hexane and carbon tetrachloride do not induce
crystallinity and neither do ethanol and n-butanol. The polarizable aromatic
hydrocarbons benzene and toluene do so, however, and so does benzyl
alcohol. Other liquids inducing crystallinity are polar but there seems to
be no obvious relationship between the equilibrium density of the polymer
and the polarity of the liquid. Results for acetone and n-amyl methyl ketone
and for nitromethane and nitroethane suggest that molecular size may be a
factor with liquids of the same type but this is not general.
Table 1 shows that, in general, the greater the equilibrium density the
greater the swelling of the polymer although density and swelling are not
linearly related. Swelling may be regarded as a measure of the interaction
between polymer and liquid and the development of crystaUinity may
depend on such interaction. Figures 2 and 3 show equilibrium density and
swelling respectively as functions of the solubility parameter 3 of the liquid.
3 =[(L, - R T ) / V ] }
where Le is the molar latent heat of vaporization of the liquid and V is
molar volume, both at the absolute temperature T. Values of 3, which are
given in Table 1, were taken from published values 8' 9 or calculated using
values of Le obtained by use of the Hildebrand rule 1°. The difference
318
 THE CRYSTALLIZATION OF POLYETHYLENE TEREPHTHALATE

between the solubility parameters of a liquid and a not too polar amorphous
polymer can be related to the heat change in their mixing and the interaction
of polymer and liquid will therefore depend, in part, on the solubility para-
meter of the latter.

"E 1"42
..90

1-40

Figure 2 - - Equilibrium
-~ i.3B density as a function of
C
Q solubility parameter of
liquid
1-36

oJ I I I I
7 8 9 10 11 12 13
Solubility parameter (cal/cm3) ~h

120

I00

80
~C
60- Figure 3--Equilibrium swel-
03 ling as a function of solu-
>o 4 0 - bility parameter of liquid

20
0
7 8 9 10 11 12 13
Solubility parameter (cal/cm3)Y2

F i g u r e 2 shows two maxima at 3 values of approximately 9 7 and 12-0

and similar maxima are seen in F i g u r e 3. The first maximum is associated
with liquids such as ketones and esters which may be regarded as basic in
the Lewis sense. Benzene and toluene, which are associated with this region
of the plots, can also be regarded as basic u. The second maximum is
associated with liquids of acidic type such as m-cresol, acetic acid and
nitromethane. The existence of two maxima may be a consequence of the
presence of basic carbonyl groups in the polymer and acidic hydrogen atoms
in CH~ groups adjacent to oxygen atoms. Giles et al. 1~ have suggested that
ester groups in polyethylene terephthalate may act as proton donors. Liquids
containing acidic groups should solvate basic polymer groups and those
containing basic groups should solvate acidic ones. Alternatively, dipolar
attraction may lead to solvation by both types of liquid.
319
 W. R. MOORE AND R. P. SHELDON

Swelling will occur in liquids whose solubility parameters do not differ

too greatly from that of the solvated polymer and be a maximum when
these solubility parameters are equal or nearly so. The solubility parameter
of the solvated polymer will presumably vary with the solvating liquid but
variations for the polymer solvated by liquids of similar type may not be
large. It has been pointed out that in eases of polymers containing both
acidic and basic groups there will be two ranges of solubility parameter
within which swelling or solution can occur, corresponding to the polymer
solvated by basic and acidic liquids respectively 13. Behaviour of this type
has been observed with secondary cellulose acetate TM. Dependence of induced
crystallinity on solubility parameter of liquid has also been noted for
crystallizable polymethyl methacrylate 1~.
This correlation of equilibrium density and swelling with solubility para-
meter of liquid should perhaps be regarded as a n approximation and other
factors may be operative. A number of the liquids used--ethanol, butanol,
benzyl alcohol, m-cresol and acetic acid--are known to be associated.
Latent heats of vaporization will often include energy required to overcome
forces causing association and it is perhaps doubtful to what extent solu-
bility parameters obtained from such latent heats can be related to forces
between associated molecules in the liquid state. Association might account
for the lack of swelling and density increase in ethanol and butanol. The
marked effect of benzyl alcohol, with almost the same solubility parameter
as butanol, may be due to the influence of the phenyl group. The points
for aniline in Figures 2 and 3 fall amongst those for acidic liquids although
aniline is generally regarded as basic. There is some evidence, however, that
the hydrogen atoms of the NH~ group can take part in hydrogen bonding
to basic carbonyl groups 1~.
Reasons for the correlation between swelling and development of crystal-
linity are not entirely clear. It is possible that the solvated polymer imbibes
the solvating liquid in swelling and that the imbibed liquid acts as a
plasticizer or molecular lubricant, reducing internal stresses and permitting
movement of the polymer chains at the temperature of the experiments
and the adoption of more favourable conditions for the development of
crystallinity. Here it may be relevant to note that the amount of dimethyl
phthalate absorbed at equilibrium reduced the glass temperature of the
polymer, determined by a dilatometric method, from 67 ° to 62°C. The
value of 62 ° refers, of course, to the polymer after crystallization. The glass
temperature of the amorphous polymer containing dimethyl phthalate may
be much lower.
If this picture of the action of liquids is correct it follows that the actual
degree of crystallinity attainable may also depend on such aspects of the
nature of the polymer as its molecular weight and previous thermal history.
This may be supported by the fact that although aniline and m-cresol both
cause high degrees of swelling the equilibrium level of crystallinity is not
very different from lhat resulting from the use of other liquids, such as
acetone, which cause smaller degrees of swelling.

We are grateful to Imperial Chemical Industries Ltd, Plastics Division,

for the gift of polyethylene terephthalate film.
320
 THE CRYSTALLIZATION OF POLYETHYLENE TEREPHTHALATE

P o l y m e r R e s e a r c h Laboratories,
D e p a r t m e n t o f C h e m i c a l Technology,
Institute o f Technology, Brad[ord, Y o r k s .
( R e c e i v e d M a r c h 1961)

REFERENCES
1 SPENCE, J. J. phys Chem. 1941, 45, 401
2 BAKER, W. O., FULLER, C. S. and PAPE, N. R. J. Amer. chem. Soc. 1942, 64, 776
KOLB, J. and IZARD, E. F. J. appl. Phys. 1949, 20, 571
a MOORE, W, R., RICHARDS, D. O. and. SHELDON, R. P, J. Text. lnst. 1960, 51, T438
5 THOMPSON, A. B. and WOODS, D. W. Nature, Lond. 1955, 176, 78
BUNN, C. W. in Fibres jrom Synthetic Polymers edited by R. HILL. Elsevier:
Amsterdam, 1953
r SHELDON, R. P. Unpublished Results
8 MOORE, W. R., EPSTEIN. J. A., BROWN, A. M. and TIDSWELL B. M. J. Polym.
Sci. 1957, 23, 23
9 BURRELL, H. lnterchemical Review, 1955, Spring, 3
~°H1LDF.aR.'e~D, J. L." and SCOTT, R. L. Solubility of Non-Electrolytes, 3rd ed.
Reinhold : New York, 1950
11 BARKER, J. A. and SMITFI, F. J. chem. Phys. 1954, 22, 375
15 ALLINGHAM, M. M., GILES, C. H. and NEUSTADTER, E. L. Disc. Faraday Soc. 1954,
16, 92; ARSHID, F. M., GILES, C. H. and JAIn, S. K. J. chem. Soc. 1956, 260, 1272
~3 MOORE, W. R. J. Soc. Dy. Col. 1958, 74, 500
14 MOORE, W. R. and RUSSELL, J. J. Colloid Sci. 1953, 8, 243
~5 Aus. Patent No. 36684
~ MOORE, W. R. and RUSSELL, J. J. appl. Chem. 1954, 4, 369

321