Farrow

Crystallinity in Poly(ethylene terephthalate)." A Comparison of X-ray, Infra-red and Density Measurements G. FARROW and I. M. WARD Crystallinities have been measured in oriented poly(ethylene terephthalate) fibres by methods based on density, infra-red spectra and X-ray diffraction. The results have been compared and no correlation exists between the three; the values from infra-red spectra are high, thosd from X-ray measurement low, and those from the density fall roughly between the two. The crystallinity of a polymer fibre is by definition the relative amount of three-dimensional ordered material present in a sample and this is most directly measured by X-ray diffraction. The results reported here show that in poly(ethylene terephthalate) the infra-red measurements are only correlated with configurational changes affecting the order within individual molecules. Using the X-ray measurements to define crystallinity it i~ found that the density of non-crystalline material increases with orientation so that density measurements bctsed on the concept of constant non-crystalline density (equal to the density of amorphous material) are inevitably in error when applied to oriented samples. The results also show that highly oriented pin-drawn yarns of low draw ratio have no appreciable crystallinity which is in contradiction to the view that the limiting stress built up in the drawing process is caused by the onset of crystallinity. Although the onset of crystallinity may well increase the reinforcement, it is clear that it is not essential to it. Dynamic loss data which had previously been correlated with crystallinity measured by the infra-red method may also be interpreted in a different light.

INTRODUCTION IT rIAS long been recognized that the physical properties of textile fibres can be considerably affected by the orientation of the molecular chains and the proportions of crystalline and non-crystalline material. The measurement of the crystallinity is, therefore, an important part of the characterization and several physical methods have been proposed and developed for the determination of crystallinity in polymers and fibres. These include X-ray diffraction, density measurements, infra-red spectroscopy, nuclear magnetic resonance spectra and moisture sorption, of which the first three will be considered in this paper. The interpretation of all these methods is a matter of some discussion, but we have adopted the view that by definition the crystallinity of a fibre is directly related to the presence of three-dimensional order which may be determined directly by X-ray diffraction. Clearly, this view is open to the question of the exact degree of three-dimensional order required, i.e. the minimum number of adjacent unit cells ~aecessary to give a discrete X-ray reflection, perfection of crystallites, etc. This point is now being investi330

CRYSTALLINITY IN POLY(ETHYLENE TEREPHTHALATE)

gated quantitatively by Taylor and Robinson using the optical diffractometer. This technique exploits the analogy between the diffraction of light by a set of holes and of X-rays by a set of atoms ~. It may, therefore, be possible in the future to define the minimum amount of order detectable by X-ray diffraction. In the meantime we shall accept that, within the limits of error set by experimental technique, the ratio of the total integrated intensity of the X-ray diffraction maxima (typical of three-dimensional order) to the intensity of the diffuse halo always present in partially crystalline samples (and typical of the entirely non-crystalline material), gives a measure of the crystallinity. The work to be described in this report is the comparison of crystallinity determined by X-ray diffraction with the crystallinity determined by infrared and density measurements on poly(ethylene terephthalate) fibres subjected to different stretching and heat treatments. EXPERIMENTAL

X-ray determination o[ crystallinity The method is described in detail elsewhereL Oriented fibres are chopped and made into a randomized sample by a pelletting technique. The sample is placed on the circumference of a focusing camera (evacuated to eliminate 'air scatter') and exposed to a strictly monochromatic beam of X-rays. The X-ray diffraction lines are recorded on film which, after processing and drying, is scanned with a microdensitometer. From the resultant trace the crystallinity is determined by measuring the integrated area of the crystalline reflections and the integrated area of the non-crystalline background and comparing the two.

Infra-red measurements When amorphous poly(ethylene terephthalate) is heat-crystallized or stretched, considerable changes occur in the intensity of many of the infrared absorption bands. The infra-red spectra of fibres are most simply studied when the fibres are mounted as a grid between rock-salt plates and an immersion medium (in this case liquid paraffin) is used to reduce the amount of scattered radiation. A grid of poly(ethylene terephthalate) fibres (of say 50 denier yarn, 2 denier per filament) absorbs too intensely in the range 5-10/, for any changes in band intensities to be useful here, but in the range 10-15# there are several bands whose intensities can be conveniently measured. We have therefore followed the method of Miller and Willis 3 and measured the intensity of the band at 898 cm -1, using the intensity of the band at 875 cm -1 as a measure of the thickness of the sample. The 898 cm -1 band has been assigned to a gauche (Figure lb) configuration of the - 4 9 CHz~CH~---O--- group in poly(ethylene terephthalate) (possibly the C H ~ rocking vibration) 4. The 875 cm -x band has been assigned to the aromatic - - C H out-of-plane deformation4, and although small changes in peak intensity may occur on crystallization these are very small (,~ 10 per cent) compared with the changes observed in the intensities of absorptions assigned to vibrations of the --O---CH2--CH~--O-- group. 331

G. FARROW and I. M. WARD

The band at 898cm -1 exhibits a maximum absorption for a sample which is completely amorphous (density l'335g/cm 3) but on heatcrystallising or drawing a reduction in the absorption of this band occurs owing to a change in configuration from gauche to trans of a number of --O--CH2--CH2--O-- groups. It is known from the crystal structure of poly(ethylene terephthalate) 5 that the configuration of the

(a)

Trans

(C)

Gauche

Oxygen •Carbon OHydrogen Figure 1--(a) The crystalline trans configuration; (b) proposed additional gauche configuration, which is present in amorphous material; (c) end views

---O--CH2---CH2--O--- group in the crystalline regions is trans. It is suggested that in the amorphous regions both the gauche and the trans configurations occur. It is assumed in this method, however, that the intensity of the gauche bands alone gives a measure of the amorphous content. Thus, by measuring the reduction in intensity of the 898 cm -~ band, the reduction in the amophous content due to various treatments can be estimated. Miller and Willis remarked that their method might not necessarily measure crystallinity because it only measured the 'pure amorphous content" and there might be many molecules in an intermediate form, described as ordered amorphous material. In spite of this proviso, the data have been used to calculate the densities of 'amorphous' material in drawn yarns. In the papers of Woods and Thompson 1°.17, the measurements are used to calculate crystallinities, and as we are principally concerned with reexamining their conclusions, we shall also calculate the crystallinity. 332

CRYSTALLINITY IN POLY(ETHYLENE TEREPHTHALATE)

BireIringence Birefringences were measured using polarized light and a Berek compensator. Each fibre was cut with a razor blade at 45 ° to the fibre axis to produce a wedge of fibre. The black interference fringe, which occurs at the edge of the fibre at zero compensation, could then be followed to the centre of the fibre as the Berek compensator was moved to values corresponding to higher birefringence. This method removes the difficulty that owing to the difference in dispersion between calcite and poly(ethylene terephthalate) the correct interference fringe will in general be colouredL

Density The physical densities of the filaments were measured by observing the point to which the samples sank in a graded density column7. Some of the densities of oriented fibres which had been subsequently heat-crystallized were found to be rather high and were checked by the flotation method, using a centrifuge to increase the sensitivity. The density of amorphous poly(ethylene terephthalate) obtained by quenching rapidly from the melt is 1"335 g/cm ". The density of the crystalline material calculated from the X-ray crystal structure~ is 1,455 g/cmL If it is assumed that the density of the amorphous material remains constant, even when oriented, then by measuring the density of the sample and applying simple proportion the crystallinity of the sample can be inferred.

Preparation of samples The filaments used in this investigation were prepared by the conventional two-stage process 8. Molten polymer is first extruded and collected as noncrystalline filaments with slight molecular orientation (melt spun yarn). Filaments of high orientation are then produced by the application of continuous extension to a pre-set limit (the draw ratio). This second drawing stage, as it is commonly called, is carried out either by passing the spun yarn over a heated cylindrical pin alone or over such a pin and then over a heated plate. A machine is used which, in essence, consists of two pairs of rolls with the heating device somewhere in between. One pair of rolls feeds the undrawn yarn from the spinning machine forward at a constant rate and the second pair moves at a higher velocity, drawing the yarn over the heated surfaces. Yarns of different draw ratios (i.e. ratios of velocity of feed rolls to velocity of draw rolls) were produced; the temperature of the pin was maintained at 80°C and of the plate, when used, at 180°C. A further heat treatment of yarns was carried out in a silicone oil bath. The yarns (pin-drawn and pin- and plate-drawn) were held under tension in the oil and brought up to the required temperature slowly to avoid sudden stress, and then held at the desired temperature for half an hour. After cooling, the samples were washed with carbon tetrachloride, to remove any oil adhering to the surface, and dried. In this paper details are also given of a comparison of crystaUinities between spun yarn (0"006in. diameter) and thin poly(ethylene terephthalate) film (0"005 in. thick), of little or no birefringence, which were heat-treated at different temperatures. A silicone oil bath was again used but care had to be taken that the heat treatment and method of immersion did not produce 333

G. F A R R O W and I. M. W A R D

Table 1. Crystallinities of unoriented film (0"005 in. thick) and yarn (0"006 in. diameter) samples from X-ray, infra-red and density measurements Temp.

Density (g / cm 3)

Density

117 146 186 213 227

1'359 1"378 1"385 1"395 1"391

20 36 41 50 48

117 146 186 213 227

1"356 1'382 1"384 1-398 1"358"

(°C)

Crystallinity (%) Infra-red l

X-ray

61 55 59 73 73

29 35 37 46 48

Film

Yarn

Table 2.

18 41 31 38 35 33 41 37 38 52 46 49 19" 54* 41" *The~e results are inconsistent. Crystallinities of oriented fibres from X-ray, infra-red and measurements

density

Density Sample draw ratio

Birefringence*

Density

3"0 3'5 4"0 4"5

0'138 0"165 0'180 0"191

1'356

3"0 3'5 4'0 4"5 5"0

0" 148 0"163 0"179 0"187 0"187

(g/cm 3)

Crystallinity (%) Density Infra-red X-rw

22 28 21

1'361

1"369 1"360

/ / /

59 72 76

/ [ [

(g/cm 3)

Oriented, non-cryst. 1"356 1"359 1"359 1"352

2'0 8"0 7"0

Pin- and plate-drawn yarns

Table 3.

1'385 42 58 27 1"385 42 63 27 1'385 42 77 25 1'382 39 75 24 1'382 39 84 28 *Figures supplied by D. W. Woods. Measurements on the same oriented fibres as in Table 2 crystallized for half an hour at 212 ° under tension*

,Sample draw ratio

Density

3"0 3"5 4"0 4"5

1 '402 1"402 1"404 1 "409

(g/c m3)

Crystallinity ( %) Density Infra-red X-ray 81 77 80 85

56 56 58 62

1'354 1-354 1"355 1-354 1"348 but heat-

Density (g / cm 3)

Oriented, non-eryst.

38

1"355 1"353 1-359 1"364

40 39 41 41 41

1"359 1"365 1"365 1"365 1"356

39 41 38

Pin- and plate-drawn yarns 3"0 3"5 4-0 4"5 5"0 *Previous work on

1"408

1"412 1'412 1'412 1"406

61 64 64 64 59

75 80 86 83 86

oriented fibres, heat-treated under tension, has shown that only small changes occur in birefringence at temperatures less than 225°C. It was. therefore, assumed that this applied also to the fibre samples listed in this table.

334

CRYSTALLINITY IN POLY(ETHYLENE TEREPHTHALATE)

any preferential orientation of molecular chains. This was checked by birefringence measurements on the treated material. RESULTS AND D I S C U S S I O N T h e results a r e g i v e n in Tables 1-3. T h o s e i n Table 2 are also s h o w n in g r a p h i c a l f o r m i n Figure 2. .~100

(a)

200

+.J

BO

180

J

:E

g

u

C:

160 g( -

.Io I" ~ "°~

60

t_

/+"

40

A

x

x

t~ . . . . . .

~---

140 :~ x

----o

.......

o . . . . . .

~] . . . . . .

C

2O 0 2.5

I

3"0

I

3"5

I

I

4-0 4'5 Draw ratio

I

- 120

100

5"0 200

_~I00

i ~"

..1 t " B

7"o1111o

80 ¸

180

/÷'/.

/o

60

~ 40

160 Co~

/t"/"o / / "

t,.

140 m

/+

e,~ x

/

o

120

20 __p_. . . . . . o. . . . . .

0

25

. - ~ - . . . . . . l a. . . . . I I 3-0 3'5 4"0 4-5 Draw ratio

C

I 5"0

100

Figure 2---Crystallinity as a function of draw ratio for oriented samples o'f poly(ethylene terephthalate) yarn: (a) drawn over heated pin and heated plate; (b) drawn over heated pin. A, density, B, infra-red, C, X-ray, D, birefringence measurements

Comparison oi X-ray and infra-red measurements It can be seen that the infra-red measure of crystallinity is higher than the X-ray measure for the unoriented samples (Table 1), particularly for thin film, and very much higher for drawn yarns. In fact, a value of 50 per cent (Table 2) is obtained for a sample of draw ratio 3"0, which shows no crystallinity on the basis of the X-ray measurements. An 335

O. FARROW and I. M. WARD explanation of this divergence has already been suggested4. If the 898 cm -1 band is correctly assigned to the gauche configuration of the --O--CH2---CH2---O~ group, the intensity of this band is related to only part of the amorphous material, since the trans configuration can be present both in amorphous and crystalline material. It was assumed previously that in unoriented, heat-crystallized material the ratio of trans and gauche isomers in amorphous material remained constant and thus a decrease in the concentration of the gauche isomer gave a measure of the decrease in amorphous content. However, there is a marked difference between the crystallinities determined from infra-red measurements on spun yarn and thin film, which is not apparent in the X-ray measure, or in the density, and implies that the ratio between gauche and trans configurations does not remain constant. It is possible that surface orientation (which is hard to detect) may exist, and this would affect the ratio of the trans to the gauche configuration. This explanation could account for the difference between film and fibre where very different surface areas are involved. With drawn yarns, the drawing process will tend to produce fully extended trans configurations, without necessarily producing crystallinity. Thus it can easily be seen that the infra-~d measure of crystallinity would be very much higher than the X-ray measure. Comparison of X-ray and density measurements It has been shown previously2'9 that a reasonable correlation exists between crystallinities measured by X-rays and density for unoriented specimens of poly(ethylene terephthalate). This is again confirmed in Table 1, the only discrepancy occurring with samples that have been heatcrystallized below 120°C. No similar agreement exists between the crystallinities measured by density and X-rays for drawn yarns. It is suggested that this arises from the fact that in drawn oriented yarns it is no longer permissible to regard the non-crystalline material as having a constant density of 1.335g/cmL With drawn yarns, of draw ratio 3-0 and less, no discrete X-ray reflections appear; there is only a 'thickening' of the intensity on the equator of the X-ray photograph. This is in disagreement with the previously held viewx°, based on the infra-red method of measuring crystallinity, that the noncrystalline density decreased on drawing, owing mainly to the formation of microscopic voids. It is suggested not that voids are no longer formed at high draw ratios but that the increase in density due to the orientation of molecules (without the formation of crystallinity) is far in excess of the reduction in density caused by the formation of microscopic voids. From the X-ray measure of crystallinity and the density of the crystalline regions (calculated from the crystal structure as 1"455 g/cm3), the density of the non-crystalline material in oriented fibres can be calculated (last columns of Tables 2 and 3). In most cases the value is about 1.355 g/cmL It is a little higher for the yarns which were heat crystallized. Molecular orientation measurements Measurements of the molecular orientation in poly(ethylene terephthalate) filaments were made in a previous investigation 11 by using the dichroism 336

CRYSTALLINITY IN POLY(ETHYLENE TEREPHTHALATE) of dyestuffs extruded with the molten polymer during the spinning of the filaments. Orientation factors obtained in this way were then compared with those obtained from birefringence measurements, which depend on the polarization of the molecules. The changes occurring on stretching the filaments over a hot pin at 80°C were studied, together with the effects of subsequent shrinkage of these drawn filaments. It was concluded from this work that up to draw ratio 2-96 orientation occurred without crystallization. At higher draw ratios the orientation factors derived from birefringence and dichroism measurements did not agree, the latter values being lower. If crystallization occurs and the crystalline regions are free from dyestuff then the dichroism measurements will relate only to the orientation of the non-crystalline regions. These regions are, therefore, less oriented than the crystalline regions. This would be expected since the noncrystalline regions will contain all the entanglements and the non-ordered molecules. In addition the shrinkage in boiling water was very dependent on draw ratio: below a draw ratio of 4'0 the shrinkage was very large, suggesting that the crystallinity was very low or non-existent. The general picture which emerged from these results was that at low draw ratios a yarn of high orientation and little or no crystallinity was produced. This picture is confirmed by the present X-ray measurements. In the absence of hot-plate drawing the crystallinities (Table 2) of drawn yarns are very small.

Drawing of poly(ethylene terephthalate) filaments The essential feature of the drawing of poly(ethylene terephthalate) filaments is the sigmoid shape (dependent on the rate of straining) of the stress-strain curves. The final rise in modulus in these curves implies that reinforcement occurs and it is of some interest to consider how this arises at molecular level. It has been shown by Treloar 1~ that a network of long chain molecules will exhibit this form of stress-strain characteristics and it is usual to assume that the network is formed by chemical cross!ink.~ and entanglements. It would now appear, therefore, that the principal cause of reinforcement in poly(ethylene terephthalate) is the presence of entanglements, since chemical crosslinks do not occur. Independent evidence for the existence of such entanglements has come from the results of nuclear magnetic resonance investigations of poly(ethylene terephthalate) and related polyesters13'14. The measured second moments at 20°K and 90°K (which are a measure of the interatomic distances of the hydrogen atoms) are greater than the calculated values for wholly crystalline samples. This was attributed to the tangling of molecules in the amorphous regions and consequent very close approach of some of the hydrogen atoms. In the case of poly(ethylene terephthalate) it was previously considered that the reinforcement occurring in the drawing process arises from stressinduced crystallization1~. Although the onset of crystallization may well increase the reinforcement, it is clear that it is not essential to it. This has been confirmed by production of oriented fibres of poly(ethylene methylterephthalate) of birefringences up to A=0"06 (draw ratio 6-5). The drawn fibres exhibit drawing characteristics similar to those of poly(ethylene terephthalate) and have quite reasonable tenacities (2-2"5g/denier). If it 337

G. FARROW and I. M. WARD is assumed that the maximum birefringence, ~Xm~xof this polymer can be calculated from the polarizability of the bonds 1~, an approximate figure of Amax= 0 ~14 is obtained. In this calculation it was assumed that the molecule is planar and that the packing of the molecules corresponds to a density of l'170g/cm 3, the density of amorphous poly(ethylene methylterephthalate). It seems, however, that birefringences approaching 0.14 are unattainable because the bulky nature of the methyl side-group prevents the molecular

Figure 3--X-ray photographs of poly(ethylene methylterephthalate) yarn: (a) undrawn; (b) drawn (A = 0"06) chains from extending fully. X-ray photographs (Figure 3) of the fibre before and after drawing show a non-crystalline pattern, although in the latter case there is a 'thickening' on the equator of the X-ray film, indicating that an alignment of molecular chains has occurred, but without the production of crystallinity. The density also increased to a value of 1-199g/cm 3. Various heat treatments have been applied to the polymer in an attempt to induce crystallization, but, so far, without success. Dynamic mechanical measurements Poly(ethylene terephthalate) undergoes two dynamic mechanical loss processes, one associated with the glass-rubber transition which occurs at about 80°C in amorphous polymer and the other at .about - 4 0 ° C (at -~ 100c/s). The high-temperature dynamic loss process is considerably affected by orientation and crystallization of the polymer, as shown by Woods and Thompson 15. It was found that the temperature of maximum work loss moved to higher temperatures with increasing fibre crystallinity, as determined by the infra-red method of Miller and Willis 3. As the changes in infra-red spectra are directly related to configurational changes, and correlate more closely with the extent of molecular orientation in the fibres, it would appear that a better interpretation of the results obtained from the dynamic mechanical measurements is that the upper loss peak is affected by the cordigurational changes or the molecular orientation rather than by crystallinity per se. This result is in agreement with the conclusions eached by Ballou and Smith is in some earlier work on the measurement of the dynamic modulus of poly(ethylene terephthalate) fibres. It is intended to test this point in detail in the immediate future by making dynamic mechanical measurements on the range of samples characterized in this investigation. 338

CRYSTALLIN1TY I N P O L Y ( E T H Y L E N E T E R E P H T H A L A T E )

Recent nuclear magnetic resonance measurements on' poly(ethylene terephthalate) have shown that the n.m.r spectra are considerably affected by crystallinity and orientation 19. These changes have been related qualitatively to corresponding changes in the dynamic mechanical properties. It is intended to make further comparisons between n.m.r, spectra and X-ray crystallinity data to obtain more detailed information concerning the relationship between the dynamic mechanical properties and molecular structure. CONCLUSIONS

The measurements of crystallinity in oriented poly(ethylene terephthalate) fibres by methods based on density, infra-red spectra and X-ray diffraction have been compared. There is no correlation between the results of the three methods. By definition, the X-ray measure has been considered as the standard, and it is then concluded that (1) the infra-red method measures configurational changes, more related to the orientation of molecular chains than to crystallinity; (2) the density of the non-crystalline material can no longer be assumed constant and therefore the measurement of crystallinity in drawn fibres by density is unreliable.

We wish to thank Mr D. W. Woods for his co-operation in providing quantities of drawn filament yarns. Research Department, Fibres Division, Imperial Chemical Industries Ltd, Harrogate, Yorks (Received 7th March, 1960. Revised version received 13th April, 1960) REFERENCES 1 HANSON, A. W., LIPSON, H. and TAYLOR, C. A. Proc. roy. Soc. A. 1953, 218, 371 2 FARROW, G. and PRESION, D. 'Measurement of crystallinity in poly(ethylene terephthalate)' Brit. y. appl. Phys. In press 3 MILLER, R. G. J. and WILLIS, H. A. J. Polym. Sci. 1956, 19, 485 4 GRIME, D. and WARD, I. M. Trans. Faraday Soc. 1958, $4, 959 5 DAUBENY, R. P., BUNN, C. W. and BROWN, C. J. Proc. roy. Soc. A. 1954, 226, 531 6 KNAPP, P. E. Private communication r KOLB, H. J. and ]ZARD, E. F. J. appl. Phys. 1949, 20, 564 8 HILL, R. (Ed.) Fibres from Synthetic Polymers, Elsevier, London, 1953, Chap. 14 9 MARTEN, L. and WARD, I. M. Unpublished work 10 THOMPSON, A. B. and WOODS, D. W. Nature, Lond., 1955, 176, 78 11 PATTERSON, D. and WARD, I. M. Trans. Faraday Soc. 1957, $3, 1516 12 TRELOAR, L. R. G. The Physics of Rubber Elasticity, 2nd edn, Oxford University Press, 1958, Chap. 4 13 LAND, R., RICHARDS, R. E. and WARt), I. M. Trans. Faraday Soc. 1959, 55, 225 ~ BATEMAN, J., RICHARDS, R. E., FARROW, G. and WARD, I. M. Polymer 1960, 1, (1), 63 15 THOMPSON, A. B. J. Polym. Sci. 1959, 34, 741 16 BONN, C. W. and DAtIBENY, R. P. Trans. Faraday Soc. 1954, 50, 1173 lr THOMPSON, A. B. and WOODS, D. W. Trans. Faraday Soc. 1956, 52, 1383 ~a BALLOU, J. W. and SMrrH, J. C. J. appl. Phys. 1949, 20, 493 10 WARD, I. M. Trans. Faraday Soc. 1960, 56, 648

339