Peters

Morphology of cold-drawn high-density polyethylene fibres W. G. Harland, M. M. Khadr and R. H. Peters Department of Polymer and Fibre Science, University of Manchester Institute of Science and Technology, Manchester M60 1QD, UK (Received 11 June 1973; revised 22 August 1973) Drawn fibres of high-density polyethylene, and their crystalline residues obtained by nitric acid digestion of the amorphous phase, exhibit single- or double-peaked melting endotherms depending on the degree of crystallinity of the fibre, the draw ratio and, in the case of some of the crystalline residues, whether or not they have been melted once or twice. The melting behaviour of the drawn fibres and their crystalline residues has been accounted for in terms of the rupture of a fraction of the original lamellar structure and the growth of a new crystalline structure. INTRODUCTION There is a considerable amount of evidence to show that the process of cold drawing high-density polyethylene (HDPE) involves rupture of the original lamellar structure and the formation of a new crystalline microfibrillar structure 1-5. At some intermediate stage of drawing at least two different crystalline entities must co-exist and they would be expected to have different melting points; nevertheless drawn HDPE products with two or more melting points have not been reported. However, Meinel et al. z,6 have observed that the crystalline residues obtained from drawn HDPE film by removing the amorphous fraction with nitric acid yield endotherms with twin peaks when melted a second time; furthermore, the position on the temperature axis of one of the twin peaks was observed to change markedly as drawing proceeded. Twin melting peaks were not observed with undrawn and drawn parent films, with crystalline residues from undrawn film, or with crystalline residues from drawn film after a first melting. Meinel and Peterlin attributed the low-temperature peak ( L T P ) to the presence of lamellae which had had their folds cut by the nitric acid treatment, and the high temperature peak ( H T P ) to the melting of a crystalline entity formed from tie molecules which had survived the nitric acid treatment. They argued that initially these two components of differing low molecular weight would constitute a molecular mixture and as such exhibit a single melting point, but would separate by a process of fractional crystallization during solidification after melting, and when remelted would exhibit separate melting points. The validity of these arguments is dependent on tie molecules surviving the nitric acid treatment, for which the evidence is slender. Meinel and Peterlin 6 considered that to a first approximation the elastic modulus of the acid-treated films should be proportional to the number of tie molecules. They claimed that for as long as the modulus could be measured, namely for up to 8h

treatment with acid, the value was practically unchanged. After further acid treatment the specimens were too brittle to make measurements with. This evidence is not satisfactory in that their data show a decrease in modulus of 13~o and, more significantly still, show that after 8 h treatment the weight loss was only 5 or 6 % whereas the samples whose melting endotherms were under consideration were derived from film that had received 50 h acid treatment and suffered a weight loss equal to its amorphous content (about 25 ~). The investigations of Meinel et al. 3, 6 were confined to two moulded samples of HDPE film, one of which had been slowly cooled and the other quench cooled to yield isotropic products of 75 and 68 % crystallinity respectively. The absence of twin melting points in the endotherms of drawn film and uncertainty as to the identity of the two crystalline fractions giving rise to twin peaks in the melting endotherms of the crystalline residues, prompted the authors of this paper to conduct a similar investigation with samples of HDPE in a lower but overlapping range of initial degree of crystallinity, and with samples also differing in initial degree of orientation. Since these two parameters are difficult to vary with films the necessary samples were obtained by extrusion as fibres. It has been found that a fibre containing about 6 0 ~ crystalline material (hereafter designated 60-fibre) gives, after drawing, a twin-peaked melting endotherm, while fibres of 67.5 and 70 % crystallinity (67-5-fibre and 70-fibre) yield only single peaks. The crystalline residues derived from the 60-fibre (60residues) yield twin peaks when melted or remelted, whereas the 67.5 and 70 residues have endotherms with twin peaks only when remelted. At very high draw ratios all fibres and crystalline residues have a single melting point. In addition to studying melting behaviour, the degree of crystallinity of the fibres and the molecular weight of the crystalline residues have been measured at various draw ratios. The discovery that HDPE of lower degree of crystallinity than that used by Meinel et al. a, 6 yields cold-

POLYMER, 1974, Vol 15, February

81

Morphology of cold-drawn HDPE fibres: W. G. Harland et aL drawn products which have melting endotherms with twin peaks, and that the crystalline residues derived from these products exhibit endotherms with corresponding twin peaks during melting and remelting, has simplified the identification of the crystalline fractions responsible for the twin-peak phenomenon. Arguments are put forward to show that in the initial stages of drawing they are the torn-off blocks of folded chains and the original lamellae; the former constituting the LTP and the latter the HTP. As drawing proceeds both components undergo modification and the changes in melting characteristics of both fibres and residues are consistent with partial destruction of the original lamellae and formation of a new crystalline structure. While the conclusions as to the identity and origins of the crystalline fractions differ from those of Meinel and Peterlin6 they are in qualitative agreement with the accepted mechanism of deformation and with models proposed for the crystalline fibrillar structure of oriented polyethylenO-5. Notwithstanding the above conclusions, consideration has also been given to the possibility that the presence of twin melting points in endotherms obtained when crystalline residues are melted a second time is connected with the presence of buried chain folds in lamellae. It is possible that fractional crystallization would occur during solidification after a first melting and the endotherm of remelting would have multiple peaks corresponding to multiple traverses of the lamellae. However, twin melting peaks are not observed in undrawn HDPE or their corresponding crystalline residues, although Ward and Williams7 have obtained gel permeation chromatograms with multiple peaks from such residues. When the residues of drawn HDPE are investigated multiple peaks in melting endotherms are formed but not in gel permeation chromatograms. The two phenomena appear to be mutually exclusive. EXPERIMENTAL

Polymer Commercial HDPE (Code No. 65.045) was supplied by Shell Plastics Laboratory, Carrington. It had a viscosity-average molecular weight of 6.0 x 104 and a melt flow index of 4-5g/10min at 190°C. Fibres Fibres were extruded at 200°C from a rod-type spinning unit. The output rate was kept constant, but the wind-up velocity and the distance below the spinneret at which the fibres were quench-cooled in water was varied. Variation of these two parameters provided fibres of different degree of crystaUinity and different birefringence as shown in Table 2. Drawing procedure Filaments (2cm lengths) were drawn at 23°C and 0.5 cm/min on an Instron machine. Samples were drawn to their natural draw ratio and to draw ratios of 10, 15 and 20. The natural draw ratio was calculated as the ratio of the cross-sectional area of fibre in the relaxed state before and after drawing. Oxidation with nitric acid Samples of drawn and undrawn fibres were treated with 95 ~o nitric acid at 80°C for 50 h. The residues were

82

POLYMER,

1974, V o l 15, F e b r u a r y

Table 1 Crystalline content ( ~ ) from AH and nitric acid digestion 70-fibre

60-fibre (An=0.0016)

60-fibre (An=0.012)

Draw ratio

AH

HNOa

AH

HNOa

AH

HNOa

Undrawn 10 15 20 25

71 74 75 76 77

71 75 75 76 76

60 77 79 80 81

62 76 78 80 81

60 74 75 76 m

61 74 75 75

washed in water until free from acid, extracted with hot acetone for 4h, and dried in vacuo at 60°C to constant weight. The main requirement of the nitric acid treatment is that it should be adequate to remove all the amorphous phase from a variety of fibres which differ in crystaUinity and orientation. Table 1 compares degree of crystaUinity calculated from heat of fusion with percentage weight of residual material. The two sets of figures are very similar and, since those derived from residual weight are inflated slightly by the presence of terminal carboxyl groups, it follows that the duration of acid treatment was more than adequate.

Molecular weight of crystalline residues Intrinsic viscosities were determined from decalin solutions at 120°C and viscosity-average molecular weights were calculated from the expressionS: [~7]=2"76 x 10-4~v) 0.7s Values at different draw ratios are given in Table 2. Owing to the presence of buried and uncut chain folds the molecular weights give only a rough indication of fold length.

Melting endotherms The endotherms were determined with a Perkin-Elmer differential calorimeter Model DSC-1. Samples of known weight (3--4mg), were heated in dry nitrogen (18 ml/min) at a rate of 8 °C/rain. In the case of crystalline residues heating was continued to 10°C above the melting point and the sample was cooled in the instrument at 8°C/rain. The thermogram for remelting was obtained from the newly crystallized sample with the same rate of heating as before. Percentage crystallinities were derived from heats of fusion and a value for crystalline polyethylene of 66 cal/g. RESULTS AND DISCUSSION

Fibres From a wide variety of experimental methods an impressive amount of evidence has been accumulated to show that cold drawing HDPE involves rupture of the original lamellar structure and the formation of crystalline microfibrils from small blocks of folded chains 1-5. The derivation by Meinel and Petedin z of a differential work density curve from a true stressstrain curve is particularly informative since it deafly demonstrates the existence of two more or less simultaneous deformation mechanisms--destruction of the original lamellar structure and deformation of the new structure. It was proposed that the second process occurs by longitudinal slip of microfibrils which a r e

Morphology of cold-drawn HDPE fibres: W. G. Harland et aL Table 2

Properties of fibres and crystalline residues 70-fibre: (280)a (9.2)b (19.5)c (0.60)d

67.5-fibre: (260)a (6.8)b

Draw ratio

Crystallinitye (~/o)

Anf

MW of residues

Crystallinity (~/o)

An

MW of residues

Undrawn 10 15 20

70.8 74.0 75.2 75.9

0.0015 0.016 0- 022 0.095

1360 1500 1540 1610

67-5 75.2 76- 8 77.9

0.0018 0.029 0.039 0.040

1160 1480 1560 1670

Rupture

77.9

.

.

.

60-fibre(low An): (220)a (5" 1)b (22-1)c (0.24)d Undrawn 10 15 20

Rupture

60.4 77.0 79.1 80.2 81.3

.

. 60-fibre(high An): (45)a (2.6)b (21-2)c (0.80)d

0.0016 0.035 0.050 0.051

1000 1640 1740 1750

60.0 74.0 74.8 76.0

0-012 0.032 0.044 0.046

--

--

80.2

--

1100 1540 1670 1720

a Filament denier (g/9000m) b Natural draw ratio c Ultimate strength based on final denier (g/denier) d Ratio of nominal tenacities at yield and rupture e Derived from heats of fusion f Measured with a polarizing microscope and a calibrated quartz-wedge compensator

connected together by relatively few fie molecules. To this interpretation of the drawing process must be added changes which occur in total crystalline content and in fold length (Table 2). Nevertheless, it is not at variance with the established concepts to suggest that during drawing, segmental and molecular mobility are so enhanced that increases in degree of crystallinity and fold length may readily occur. The data in Table 2 suggest that the increase in crystallinity is largely accounted for by the increase in fold length. It is anticipated that increases in fold length may occur both in the remainder of the original lamellae and in the smaller blocks of folded chains. It should, however, be easier for fold length increases to occur in the latter since the co-operative movement of a smaller number of folds will be required. The melting behaviour of the fibres and their crystalline residues must be related to morphological changes accompanying drawing. The melting endotherms of 60-fibres at their natural draw ratios (Figure 1) clearly show that a new crystalline species of lower melting point is being created. In order to explain the lower melting point it must be assumed that the size of the crystal units are less than in the original lamellae such that the surface free energy contribution to melting is increased, and/or the fold length has decreased. Estimates of crystal dimensions from the line broadening of wide-angle X-ray scattering, and calculations of crystalthickness from the long period obtained by small-angle X-ray scattering show that both changes may occur during drawing 9. The fraction of lower melting point is, therefore, identified with torn-off blocks from the original lamellae, and the fraction of higher melting point with the original lamellae. As drawing proceeds both peak temperatures increase but that of the LTP increases the more rapidly (Figure 1). These changes are a reversal of those occurring in the initial stages of drawing and may be attributed to a differential rate of fold growth and aggregation of small crystals into larger units (fibrils) with a lower surface free energy. Owing to lack of resolution of the two peaks it is not clear whether the whole or only a part of the original lamellar structure is destroyed by drawing.

~

a

b

129oc

130'5°C

9.5oc

131oc

131"5°C

~

131.5oC

2 5°C 132oc

133°C

J

(2C ~'

Figure 1 Melting end0therms of 60-fibres at draw ratios shown in parentheses. An undrawn state: (a) 0.0016; (b) 0.012. Peak temperatures indicated

Meinel and Peterlin 1 are of the opinion that with HDPE film destruction is complete at a draw ratio of about 10, but the fact that the height of the HTP is maintained up to draw ratios of at least 15 suggests otherwise for fibres. This view is supported by evidence derived from the endotherms of the crystalline residues which is discussed later.

P O L Y M E R , 1974, V o l 15, February

83

Morphology of cold-drawn HDPE fibres: W. G. Harland et aL a

131oc

b

130.5oc

the widths of the peaks at half height increase from 1.5 to 2"1°C and 1.5 to 3.7°C for 70- and 67.5-fibres respectively. Thus there is no reason to suppose that the morphology and melting behaviour of the drawn versions of the fibres of relatively high crystallinity are in any way anomalous. At high extents of drawing (Figures 1 and 2) only a single melting peak is observed for all fibres regardless of their original degree of crystallinity and orientation, which is consistent with the formation of a homogeneous fibrillar structure. It is interesting but not surprising that at ultimate elongation the fibres have the same melting point (135 ° + 0.5°C), a similar degree of crystallinity, and a similar tensile strength (Table 2). It is possible that if a much lower rate of drawing had been used differences would be even less.

Crystalline residues

32°C

132"5°C

Having established the identity of the two crystalline fractions responsible for the twin melting peaks of fibres, it is necessary to consider their relation to the melting endotherms of residues. By comparing Figures 1, 2 and 3 it may be seen that in terms of the existence of single or twin melting peaks at equal draw ratios, and after the first melting of the residues, the residues and fibres are strictly comparable. However, where twin peaks exist in fibres and residues there is a noticeable difference in the degree of resolution of the two peaks. a

(201

12OI

Drawn fibres of 67"5 and 70% crystallinity do not appear to have two melting points (Figure 2). Since there is no reason to suspect that a different deformation mechanism applies, the possibility that twin melting points exist at extents of deformation less than the natural draw ratio must be examined. There is no reason to expect equivalent structural changes to occur at equal extents of deformation in materials with different natural draw ratios. This situation arises first because drawing takes place in two stages--drawing to the natural draw ratio at a neck followed by uniform extension of the material which has already been drawn to its natural draw ratio, and secondly, because the rate of deformation, and hence the intensity of deformation, is enormously greater when drawing is proceeding at a neck. Thus at equal draw ratios different extents of morphological change are to be expected in 67.5-, 70-, and 60-fibres of low and high orientation, which have draw ratios of 9-2, 6.8, 5.1 and 2-6 respectively. In particular the significant morphological changes which give rise to twin melting endotherms appear at lower extents of deformation the higher the natural draw ratio. Furthermore, close inspection of Figure 2 shows evidence for the previous existence of twin melting peaks by the fact that the endotherms are wider for fibres at their natural draw ratio than they are for fibres in the undrawn state or at a draw ratio of 10. Between the undrawn state and the natural draw ratio

:84

POLYMER, 1974, Vol 15, F e b r u a r y

C 124.5oc 25:5°C

124.5 o C

/~1

Figure 2 Melting endotherms of (a) 70-fibres and (b) 67.5-fibres at drawn ratios shown in parentheses. An undrawn state: (a) 0.0015; (b) 0.0018. Peak temperatures indicated

b

124oc 124"5°C

125oc

(11)[i

<

.

(lol

~ 1260c

AI25oc

|I H ~28oc I I ,'l

l | 127"5°C• I I ,~ IIi I

~ 127°C

^127°C

ll/;

122soc/ V~

28°C

(!.s)

.IY, I

..o#'i

,24o

_-- J

2,,o

~ i-

,

A

ll.q127°C I I/I

III'

,,~o / A !

126.5oc

I//t 128'5°C jq~h128oc

/ . . . .

125oc

111 ,i

li'i 1

L.,,','~,2~k'<. 129oc

II I

~128oc

I[,e !1= I

< ...j

Figure 3 Melting endotherms of (a) 70-residues, (b) and (c)

60-residues.--, First melt; . . . . . , second melt. An of undrawn parent fibre: (a) 0.11015;(b) 0.0016;(c) 0.012. Draw ratios of parent fibres shown in parentheses, Peak temperatures

indicated

Morphology of cold-drawn HDPE fibres: W. G. Harland et aL The resolution is far better at higher draw ratios in the residues than in the fibres. Such differences of detail are possibly the result of changes in morphology occurring during the course of heating fibre samples to determine their melting endotherms. In this connection Hosemann et aL 1°, ~ have commented on the susceptibility of the crystals of drawn fibre to thicken and give rise to a higher melting point practically independent of the original crystal thickness. In principle such changes cannot occur in the residues and it is concluded that the melting endotherms of the residues give a more realistic picture of the fibre morphology. The first- and second-melting endotherms of the 60-residues are slightly different. After first melting and solidification both peak temperatures increase by about 2°C, the area of the LTP increases at the expense of the HTP, and the degree of crystallinity also falls by up to 5 ~ as observed by Meinel and Peterlin6. These changes undoubtedly result from a more effective fractionation of the molecules during recrystallization and indicate that, while the two crystalline fractions produced by the drawing process have different melting points largely owing to differences of fold length in the fibres, differences in overall crystallite size also exist. It would appear that the first-melting endotherms of the residues give the best indication of the morphological changes brought about by drawing, but unfortunately, the relative peak areas cannot be resolved accurately. A rough visual estimate would suggest that about 30 of the lamellar structure of the 60-fibres has been degraded and about 20 ~ in the case of the 60-fibre of higher initial orientation. The endotherms of the 70-residues are more difficult to interpret; a single peak is obtained from the first melting and a double peak from the second melting. Identical behaviour has been described by Meinel and Peterlin 6 for residues derived from drawn HDPE film. Their explanation, which has already been given in the introduction, is feasible if, as they claim, tie molecules and molecules derived from oxidized lamellae are involved in a fractional crystallization process during solidification following the first melting. Since this claim cannot be valid for 60-residues, which have double-peaked endotherms when first melted, and since there is no reason to suspect any radical difference between drawing 60, 67-5, and 70-fibres, their explanation is of doubtful validity. While the experimental data from this investigation offer no explanation for the phenomenon, it should be noted that geometrical differences between crystals give rise to different surface free energy contributions to melting and hence two crystalline fractions of different thickness may melt at the same temperature. Such an occurrence would, however, be fortuitous. Fortunately for present purposes an explanation is not essential. The absence of twin melting peaks for 70-fibres (Figure 2) and for the first melting of 70-residue (Figure 3) makes estimation of the fraction of lamellar structure degraded during drawing difficult. Judged from a comparison of endotherms obtained during the second melting of 60 and 70-residues there does not appear to be any significant difference. However, these endotherms are not a reliable guide and if, as claimed by Meinel and Peterlin ~, HDPE with a crystalline content of more than 70 ~ becomes less crystalline at extents of drawing below the natural draw ratio while polymer of less

than 7 0 ~ crystallinity shows only an increase, it is a reasonable deduction that the fraction of lamellae degraded in the 70-fibres is somewhat greater than for the 60-fibres. Certainly there is no evidence for complete destruction of lamellar structure in fibres as has been suggested for drawn HDPE film. According to Meinel and Peterlin 1 the destruction of the original lamellar structure is nearly completed at a draw ratio of about 10, and further deformation proceeds largely by longitudinal slip of newly formed microfibrils that are believed to be attached to one another by a small number of tie molecules which transverse quasi-amorphous boundary layers. While this mechanism may be consistent with the negligible changes which occur between draw ratios of 10 and 15 it does not assist in explaining the disappearance of the LTP at draw ratios between 15 and 20, which was not observed in the work of Meinel and Peterlin6 on drawn films. Another significant difference between the two investigations is that rupture of the films occurred at nominal stresses less than the yield stress, whereas rupture of the fibres occurred at nominal stresses appreciably greater than the yield stress (Table 2). The different behaviour of the fibres suggests that in their case a more coherent fibrillar structure is developed and undergoes further deformation by slipping of constituent blocks of folded chains until lateral cohesive strength is improved to the point where it equals that of the fibrils, or that of interfibrillar tie molecules, and the fibre ruptures. Such a mechanism would destroy the identity of the original crystalline blocks used to build the fibrils, and hence the crystalline phase would exhibit a single melting point. The experimental data and deductions also have a bearing on the investigations of Ward and Williams7 into the distribution of buried chain folds in the lamellar structure of cold-drawn bulk polyethylene. Using gel permeation chromatography to elucidate the distribution of chain lengths in crystalline residues, they detected multiple peaks corresponding to multiple chain traverses in lamellae derived from branched polymer, but obtained a single broad distribution from residues of linear polyethylene. Since the twin melting peaks discussed in this paper have been attributed to the creation of two crystalline fractions of different thickness, it is interesting to question why they were not detected in the investigation of Ward and Williams. The reason is almost certainly because the residues of both fractions contain buried chain folds and hence have distributions of chain lengths which when combined overlap to produce a single-peaked broad distribution. The same concept also serves to explain the observation of Meinel and Peterlin6 that annealing residues just below their melting point removes the LTP and leaves a broad low-temperature tail in the endotherm; annealing would eliminate the chain folding and produce an imperfect chain-extended structure with a broad melting range. Since the present work contains evidence that the extent of lamellar degradation incurred by drawing is less for 60-fibres than 70-fibres, it is probably much less still or even negligibly small, for cold-drawn branched polyethylene with a crystalline content of less than 50 ~o and a much higher compliance. Therefore, the crystalline residues from cold-drawn low-density polymer and undrawn HDPE should have a distribution of molecular weight unconfounded with that of the crystals of a

POLYMER, 1974, Vol 15, February

85

Morphology of cold-drawn HDPE tTbres: 14/. G. Harland et al. second fraction of significantly different dimensions, as in fact observed by Ward and WilliamsL Ward and Williams concluded that the absence of multiple peaks in the crystalline residues of cold-drawn HDPE was the result of large fluctuations in crystal thickness 7. It will be noted that the above explanation only differs in detail; it has been pursued to make it abundantly clear that twin peaks in melting endotherms do not have the same origin as the multiple peaks in the gel permeation chromatograms. CONCLUSIONS The appearance and disappearance of twin peaks in the melting endotherms of cold-drawn fibres of HDPE and in the melting endotherms of the corresponding residues obtained by removal of the amorphous phase by oxidation with nitric acid is consistent with a mechanism of drawing which involves degradation of a fraction of the original lamellar structure and the formation of a microfibrillar crystalline structure. The two crystalline fractions giving rise to the twin melting peaks are the

86

POLYMER, 1974, Vol 15, February

original lamellae and tom-off blocks of folded chains. In a fibre of 70 Yo crystallinity and low orientation the extent of lamellar degradation is not more than about 40Yo and is less for fibres of lower crystallinity and higher orientation. REFERENCES 1 Meinel, G. and Peterlin, A. J. Polym. Sci. (A-2) 1971, 9, 67 2 Peterlin, A. in 'Man-Made Fibres' (Eds H. Mark, S. M. Atlas and C. Cernia), Interscience,New York, 1967, Vol 1, pp 283-340 3 Meinel, G., Morossoff, N. and Peterlin, A. J. Polym. Sci. (4-2) 1970, 8, 1723 4 Hosemarm, R. J. dppl. Phys. 1963, 34, 25 5 Takayanagi, M., Imada, K. and Kajiyama, T. J. Polym. Sci. (C) 1966, 1, 263 6 Meinel, G. and Peterlin, A. J. Polym. Sci. (,4--2) 1968, 6, 587 7 Ward, L M. and Williams, T. J. MacromoL Sci. (B) 1971, 5, 693 8 Duch, E. and Kuchler, B. J. Electrochem. Soc. 1956, 60, 218 9 Glenz, W. and Peterlin, A. J. Polyra. Sci. (A-2) 1971, 9, 1243 10 Wilke, W., Vogel, W. and Hosemann, R. Kolloid-Z. Z. Polym. 1967, 237, 317 ll Hosemann, R. and Wilke, W. Makromol. Chem. 1968, 118, 23O