ISSN 0965-545X, Polymer Science, Ser. A, 2007, Vol. 49, No. 8, pp. 903–908. © Pleiades Publishing, Ltd., 2007. Original Russian Text © O.V. Arzhakova, A.A. Dolgova, I.V. Chernov, L.M. Yarysheva, A.L. Volynskii, N.F. Bakeev, 2007, published in Vysokomolekulyarnye Soedineniya, Ser. A, 2007, Vol. 49, No. 8, pp. 1502–1509.
STRUCTURE, PROPERTIES
The Effect of Preliminary Orientation of Polymers via Tensile Drawing at Elevated Temperature on Solvent Crazing1 O. V. Arzhakova, A. A. Dolgova, I. V. Chernov, L. M. Yarysheva, A. L. Volynskii, and N. F. Bakeev Faculty of Chemistry, Moscow State University, Leninskie gory, Moscow, 119991 Russia e-mail:
[email protected] Received December 26, 2006; Revised Manuscript Received February 21, 2007
Abstract—We studied how the preliminary orientation of an amorphous glassy PET via its uniaxial tensile drawing above the glass transition temperature affects the deformation behavior during subsequent tensile drawing in the presence of adsorptionally active environments. The tensile drawing of the preoriented PET samples with a low degree of preliminary orientation (below 100%) in the presence of liquid environments proceeds via the mechanism of solvent crazing; however, when a certain critical tensile strain is achieved (150% for PET), the ability of oriented samples to experience crazing appears to be totally suppressed. When the tensile drawing of preoriented samples is performed at a constant strain rate, the craze density in the sample increases with increasing degree of preliminary orientation; however when the test samples are stretched under creep conditions, the craze density markedly decreases. This behavior can be explained by a partial healing and smoothening of surface defects during preliminary orientation and by the effect of entanglement network. The preliminary orientation of polymers provides an efficient means for control over the craze density and the volume fraction of fibrillar polymer material in crazes. DOI: 10.1134/S0965545X07080068 1
INTRODUCTION
Crazing is known to be a specific mode of plastic deformation in polymers, in addition to necking and shearing [1–3]. The crazing of polymers can be conceived as a stress-induced polymer transition into the oriented fibrillar state within localized deformation regions, which are referred to as crazes [4–6]. In contrast to other modes of plastic deformation, crazing is accompanied by an increase in the volume of polymer samples due to development of macroscopic porosity. The inner structure of each individual craze is composed of asymmetric aggregates of macromolecules, which are oriented along the direction of tensile drawing and grouped into craze fibrils bridging the opposite craze walls. In space, such microfibrillar aggregates are separated by pores [5]. The diameter of craze fibrils and pores between them ranges from several nanometers to tens of nanometers. On the whole, according to modern concepts, the structure of a fibrillar material in crazes can be modeled by a heavily entangled web rather than by a set of parallel cylinders (fibrils) composed of oriented macromolecules that bridge the opposite craze walls [7]. In contrast to the classical dry crazing of polymers taking place during their tensile drawing in air, crazing in the presence of physically active liquid environments or the so-called adsorptionally active liquid environ-
ments (AALE)2 proceeds within a wider interval of tensile strains up to the stage of strain hardening of the material [6]. In this case, a gradual transition of the polymer into the oriented state at the craze–(bulk polymer) boundary is accompanied by the development of a marked macroscopic porosity (up to 60 vol %) within crazes [6], and the inner nanoporous structure of crazes is stabilized by the action of surface active environment. The mechanism of environmental crazing has been studied in detail for various amorphous and semicrystalline polymers. These studies have been summarized in various reviews and monographs [3–7]. Taking into account the fact that crazing presents a specific mode of plastic flow in polymers, which is provided by the specific orientation of macromolecular chains in craze fibrils, it seems interesting to study the effect of the preliminary orientation of polymer samples on the development of this mode of plastic deformation. In general case, the tensile drawing of polymers is accompanied by the orientation of macromolecular chains (molecular orientation) along the direction of stretching; as a result, such mechanical characteristics as yield stress, post-yield stress (or lower yield stress), and breaking strength of the material along the selected direction tend to increase [8–10]. This processing trick is traditionally used in technology for imparting the
1 This work was supported by a program “State Support of Leading
2 This mode of crazing is referred to as solvent crazing or wet craz-
Scientific Schools,” project no. NSh-4897.2006.3.
ing or environmental crazing.
903
904
ARZHAKOVA et al.
∆n 0.03 0.02 0.01
0
50
100
150 εpr, %
Fig. 1. Birefringence ∆n vs. the degree of preliminary orientation of PET samples at 80°ë.
desired strength to polymer articles processed as films and fibers [9]. With increasing level of preliminary orientation, the interval of plastic deformation during repeated stretching along the direction of preliminary orientation gradually decreases [11]. Furthermore, when a certain critical level of tensile strain is attained, the ability of preoriented polymer samples for further plastic flow is virtually suppressed [10]. The ability of preoriented polymer to experience deformation via the mechanism of crazing has been studied in very few works [12–14]. In particular, the craze growth in preoriented polymer films was studied during subsequent tensile drawing along different directions with respect to the axis of initial orientation [14]. However, many problems concerning the specific features of environmental crazing in oriented glassy polymers still remain open; among them is the effect of the degree of preliminary orientation of polymers on their deformation behavior in the presence of AALE and the processes of craze nucleation and development of fibrillar–porous structure. In this study, we examined the deformation behavior and specific features of environmental crazing for preoriented glassy PET, as this polymer has been studied in detail from the viewpoint of crazing. EXPERIMENTAL We studied the films of the commercial amorphous glassy PET with a thickness of 100 µm. According to DSC measurements, the glass transition temperature of PET was 75°C. The preliminary orientation of PET films was carried out via uniaxial tensile drawing at 80°ë; the strain rate was 100%/min. The degree of preliminary orientation was 50, 75, 100, and 150%. After tensile drawing, the polymer films were quickly cooled down to room
temperature in order to preserve the maximum level of orientation of polymer samples at a given tensile strain. For our further studies, the test samples were cut from the preoriented PET samples as dumbbell-shaped specimens with a gage size of 4 × 10 mm. The mechanical characteristics of the preoriented samples were studied during their subsequent tensile drawing in air and in the presence of AALE using an Instron-1122 universal tensile machine; the strain rate was 20 mm/min (or 200%/min). The samples were also studied under the action of constant load or under creep conditions. The tensile drawing of the preoriented samples was performed along the direction of preliminary orientation. As AALE, n-propanol was used. The birefringence of preoriented PET samples was investigated using an optical microscope equipped with a system of crossed polarizers. The microscopic studies of the samples and the calculation of craze density were performed with an Opton optical light microscope and a Hitachi S-520 scanning electron microscope. Prior to SEM studies, the surface of the test samples was decorated with platinum. RESULTS AND DISCUSSION Birefringence measurements allow direct estimation of the level of molecular orientation in polymers. At the present time, there is adopted knowledge that the uniaxial drawing of PET above the glass transition temperature Tg is accompanied by the orientation of polymer chains [15, 16]. As follows from Fig. 1, birefringence linearly increases with increasing degree of orientation from 0 for unoriented PET to 0.028 for preoriented polymer samples stretched by 150%. In the selected range of tensile strains, birefringence measurements agree well with the literature data and indicate that the stretching of PET at 80°ë at a given tensile strain rate is accompanied by the molecular orientation of polymer chains along the direction of tensile drawing [15]. Figure 2 presents the stress–strain curves recorded during the tensile drawing in n-propanol for the PET samples with different levels of preliminary orientation. All curves are seen to be typical of PET and show a well-pronounced yield tooth, stress decay after the yield point, plateau region corresponding to the postyield flow and, finally, the region of strain hardening and breakdown. As follows from the data shown in Fig. 2, with increasing degree of preliminary orientation, the mechanical characteristics of preoriented PET samples gradually increase (elastic modulus, yield stress, post-yield stress, and breaking strength), and the corresponding stress–strain curves are gradually shifted to higher stresses. In this case, the region of the postyield plateau narrows down. Various mechanical characteristics for the preoriented samples with different levels of preliminary orientation are summarized in the table. POLYMER SCIENCE
Series A
Vol. 49
No. 8
2007
THE EFFECT OF PRELIMINARY ORIENTATION OF POLYMERS
In the case of PET, an increase in the mechanical characteristics with increasing degree of preliminary orientation is likely related not only to the orientation of polymer chains along the direction of tensile drawing but also to partial stress-induced polymer crystallization. This process is known to be appreciably promoted when the polymer inner structure becomes ordered due to the orientation of polymer chains upon tensile drawing [17–20]. For preoriented samples with a degree of preliminary orientation ranging from 25 to 100%, stretching in the presence of AALE proceeds via the mechanism of environmental crazing, which entails the formation of numerous crazes across the whole gage length of the test samples. The only exception is provided by the sample with a maximum tensile strain of 150%. In this case, the corresponding stress–strain diagram appears to be appreciably different from those recorded for the preoriented PET samples with lower degrees of preliminary orientation. This stress–strain curve does not show the typical postyield plateau region after the yield point. At the same time, a relatively low stress decay after the yield point is immediately followed by an abrupt and intensive stress growth. This profile indicates the occurrence of strain hardening along the direction of preliminary orientation along with an increase in the mechanical characteristics due to crystallization. Fracture of the sample is observed at much lower tensile strains. In this case, the tensile drawing of the sample in the presence of AALE takes place not in the local regions via the mechanism of environmental crazing but proceeds homogeneously along the whole gage length of the test sample. Therefore, the preliminary orientation of PET by 150% has a critical effect on the mechanism of polymer deformation during the repeated stretching along the direction of preliminary orientation. A question concerning the efficiency of AALE for the preoriented PET samples with different degrees of preliminary orientation arises. To answer this question, let us compare the values of the yield stress σy and the postyield stress σpy for the preoriented samples during their tensile drawing in air and in the presence of AALE. Figure 3 presents the corresponding experi-
905
σ, MPa 125 5
75 4
3 2
1
25 0
150
300
450
600 ε, %
Fig. 2. Stress–strain diagrams of preoriented PET samples recorded during tensile drawing in n-propanol at a strain rate of 230%/min. The degree of preliminary orientation: (1) 0, (2) 50, (3) 75, (4) 100, and (5) 150%.
mental data plotted as (σair – σAALE)/σair against the degree of preliminary orientation, where σair and σAALE stand for the yield stress in air and in AALE (n-propanol), respectively. As follows from Fig. 3, with increasing degree of preliminary orientation, the efficiency of the action of AALE on the mechanical response of preoriented polymer samples markedly decreases. Let us consider the effect of preliminary orientation on the ability of polymer samples to experience environmental crazing during their tensile drawing in the presence of AALE, which is accompanied by the formation of numerous regions of localized deformation (crazes). For the samples after their stretching via the mechanism of environmental crazing, the craze density in the initial unoriented PET film with a thickness of 100 µm is 50–80 mm–1; for the preoriented PET samples with a tensile strain of 50, 75, and 100%, the craze density appears to be 110–140, 150–170, and 170–200 mm–1, respectively. Therefore, with increasing degree of preliminary orientation, the craze density (the number of
Mechanical characteristics of preoriented PET samples with different degrees of preliminary orientation by the tensile drawing of initial films at elevated temperature during their subsequent stretching in n-propanol (the strain rate is 200%/min) Degree of preliminary orientation, %
Yield stress, MPa
Postyield stress, MPa
Onset of strain hardening, %
Breaking strength, MPa
Elongation at break, %
Elastic modulus, MPa
0 50 75 100 150
35 50 45 52 112
12.5 18 25 20 98
315–340 200–220 145–165 115–130 20–35
29 50 47 50 120
575 330 296 241 81
1676 1688 1724 1735 3384
POLYMER SCIENCE
Series A
Vol. 49
No. 8
2007
906
ARZHAKOVA et al. (σair – σAALE)/σair, % 80
ε, % 60 1
60
40
2
2 40
3
20
1
4 0
50
100 εpr , %
0
15
30 Time, min
45
Fig. 3. Relative decrease in (1) yield stress and (2) postyield stress for preoriented PET samples stretched in n-propanol vs. the degree of preliminary orientation. The strain rate is 200%/min.
Fig. 4. Creep curves of preoriented PET samples in n-propanol under the action of constant stress: (1–3) 32.8 and (4) 49.2 MPa. The degree of preliminary orientation: (1) 0, (2) 50, (3) 75, and (4) 100%.
crazes per unit area) in the preoriented PET samples increases when the samples are stretched at a constant strain rate. For example, as compared with the initial sample of the unoriented PET, the craze density in the preoriented PET sample with a tensile strain of 100% increases by a factor of 4.
Therefore, the preliminary orientation of polymers below a certain critical level of orientation can be treated as an efficient means for increasing the craze density in the sample and preparation of materials with uniform distribution of porosity throughout the whole volume of the sample by stretching at a constant strain rate. To understand the mechanism of craze nucleation and to avoid the factor of stress changes during stretching at a constant strain rate, the preoriented PET samples were subjected to cold drawing in the presence of AALE under constant stress or, in other words, under creep conditions. At a constant stress level of 32.8 MPa, the craze density in the samples is 60–80 (the initial unoriented PET sample), 10–12 (the preoriented sample with a tensile strain εpr = 50%), 6–8 (εpr = 75%), and 1–2 mm–1 (εpr = 100%). The last sample with εpr = 100% was loaded at a constant stress of 49.2 MPa. In this case, the level of loading is higher because, at a lower stress of 32.8 MPa, the preoriented PET samples with ε = 100% do not show any visible flow within the time of creep experiments. Therefore, as compared with the uniaxial tensile drawing of the preoriented samples at a constant strain rate, the experiments performed under creep conditions show that the density of nucleated crazes markedly decreases. Furthermore, as follows from the above experimental evidence, with increasing degree of preliminary orientation, the process of craze nucleation and growth during tensile drawing under constant stress appears to be markedly hindered, and the craze growth proceeds at appreciably lower rates. For example, the analysis of creep curves (Fig. 4) shows that, with increasing degree of preliminary orientation from 25 to 100%, the most probable craze growth rate (the rate of craze-tip advance in the direction perpendicular to the axis of
An analysis of the experimental data on the craze density for each sample under study allows one to conclude that, with increasing degree of preliminary orientation, the rate of polymer transition into the fibrillar state in crazes markedly decreases because the number of sites of localized deformation (craze density) increases in proportion to the tensile stress in the samples stretched at a constant strain rate [6]. Seemingly, this increase in the craze density during the environmental crazing of preoriented samples with increasing degree of preliminary orientation is related to the higher stress level at which the tensile drawing proceeds [6]. As was shown earlier [6, 7], the process of craze nucleation in polymer samples can be described in terms of the Griffith model of critical nucleation conditions. As is known, the craze nucleation is primarily localized at surface defects, which exist in any real polymer material. According to the Griffith model, with the increasing stress level stored by the sample, the source of plastic deformation and further craze nucleation is provided by smaller nucleation sites with lower critical dimensions. If the preliminary orientation is assumed to have no effect on the whole set of potential nucleation sites of plastic deformation, the above behavior of preoriented polymers during their repeated tensile drawing seems to be quite expected: the higher the stress level in the sample, the more nucleation sites are involved in the mechanism of plastic deformation, and the higher the density of nucleated crazes.
POLYMER SCIENCE
Series A
Vol. 49
No. 8
2007
THE EFFECT OF PRELIMINARY ORIENTATION OF POLYMERS
tensile drawing) decreases by more than two orders of magnitude. The dramatic decrease in the craze density for the preoriented samples stretched under creep conditions can be explained as follows. First, taking into account the fact that craze nucleation is initiated by the stress concentration at weak structural sites, including surface defects, various structural inhomogeneities of the material, and foreign inclusions [4–7], one can reasonably conclude that the preliminary orientation entails a partial healing of surface defects. This conclusion is confirmed by the data reported in [21], where AFM studies showed that, with increasing degree of orientation in the polymer sample during tensile drawing, large surface defects are smoothed and healed. Furthermore, one should take into account the fact that the subsequent stretching is performed for the preoriented samples, which are characterized by strengthened structure and, hence, higher mechanical characteristics. In this case, the applied stress level (32.8 MPa) during creep tests is comparable to the yield stress only for the initial unoriented sample (35 MPa). For all other preoriented samples, the applied stress level is much lower than their yield stress. It is worth mentioning that, under the creep conditions used in this study, in accordance with the Griffith criterion, only the most dangerous defects are involved in the process of craze nucleation. To explain this behavior of the test samples, let us resort to the published data concerning changes in the inner polymer structure during orientation [10, 22, 23]. According to the traditional models [10], the structure of an amorphous polymer can be modeled as a network formed by macromolecular entanglements (the entanglement network) and can be compared to the structure of a crosslinked rubber, where entanglements between macromolecules serve as crosslinks [10]. This state of entanglements is the critical factor that governs the mechanical response of the polymer sample upon loading. The preliminary tensile drawing leads to morphological changes in the entanglement network and furthermore entails formation of a new entanglement network; the strength of this network increases with increasing macroscopic tensile strain of the sample. As the craze growth proceeds via the craze-tip advance in the direction perpendicular to the direction of tensile drawing, the viscous and plastic flow of a growing craze encounters resistance from the dense entanglement network formed during the preliminary tensile drawing. Furthermore, the transition of polymer material into the oriented fibrillar state within crazes also appears to be markedly hindered as this process proceeds via the disentanglement of the existing entanglement network [22], whose strength in the direction of tensile drawing for preoriented polymer samples is rather high. One should also take into account the fact that, in the craze–(bulk polymer) transition region, the already oriented polymer material is subjected to orienPOLYMER SCIENCE
Series A
Vol. 49
No. 8
2007
907
tation, and this factor, in turn, requires higher stresses for plastic flow. It seems probable that all the above factors are responsible for the marked retardation of craze growth rate during the tensile drawing of preoriented samples under creep conditions. Therefore, the preliminary orientation is shown to change both a set of local structural inhomogeneities in the sample and the polymer morphology. As follows from Fig. 2, with increasing degree of preliminary orientation, the length of the postyield plateau in the corresponding stress–strain curves decreases. As was found earlier, the length of this postyield plateau region is related to the natural draw ratio λ of a polymer; this parameter is controlled by the density of the entanglement network, which is typical of a given polymer [10]. At room temperature, the natural draw ratio of PET in the neck is close to four [6]. Due to the preliminary orientation of a polymer, macromolecules appear to be partially oriented along the direction of tensile drawing, and the overall draw ratio of the sample should be equal to the natural draw ratio of PET. To estimate the degree of the resultant draw ratio in neck for the preoriented PET stretched by 50, 75, and 100% and for the initial undeformed PET, the samples were stretched in air to the same tensile strain with a constant strain rate. Prior to tests, all samples were bent at their central parts in order to nucleate a neck and to provide further deformation in a strictly localized region. Then, during tensile drawing, one can observe uniform neck propagation along the sample. When all samples are stretched to the same tensile strain, the draw ratio of the polymer in the formed neck is measured. For the initial sample and preoriented samples with a degree of preliminary orientation of 50, 75, or 100%, this value is ~4–4.2, 3, 2.5, or 2, respectively. The product of the above values and the initial draw ratio (1, 1.5, 1.75, and 2) is 4.0–4.2, which is the typical natural ratio of PET. Therefore, the estimated values correlate with the tensile strains that correspond to the onset of strain hardening of the material (Fig. 2). This fact directly shows that the natural draw ratio λ of a polymer in the neck is equal to the λ of the polymer material in the fibrillar state within the volume of crazes. Hence, one can conclude that lower draw ratios of a polymer in the neck due to higher degrees of the preliminary orientation of PET are related to an increased volume fraction of polymer material in crazes because the volume fraction of fibrils in crazes is inversely proportional to the natural draw ratio of polymer. For an unoriented PET, this value is 0.25 (the natural draw ratio is 4.0); for the preoriented sample with 100% tensile strain, this parameter increases to 0.5. This conclusion has been confirmed by the direct microscopic observations in [13]. As was shown, as a result of the repeated tensile drawing of oriented PS samples along the direction of preliminary orientation,
908
ARZHAKOVA et al.
the volume fraction of craze fibrils νf is higher (νf = 0.45) than the volume fraction of craze fibrils in the initial unoriented sample (νf = 0.25). Therefore, the preliminary orientation of polymer samples makes it possible to control the density of nucleated crazes and the volume fraction of polymer fibrillar material in crazes. Hence, in this study, we investigated the effect of preliminary orientation of the polymer above the glass transition temperature on the specific features of environmental crazing. As was shown, preoriented samples preserve their ability to undergo deformation via the mechanism of crazing until a certain critical tensile strain is attained; then, crazing appears to be totally suppressed (150% for PET). Under stretching at a constant strain rate, with increasing degree of preliminary orientation, the craze density markedly increases. When the preoriented samples are stretched under creep conditions, the density of the nucleated crazes dramatically decreases, and this behavior is related to partial healing and smoothening of surface defects due to the preliminary drawing and the effect of the existing entanglement network. The preliminary orientation of polymers can be treated as an efficient means for control over the density of nucleated crazes and the volume fraction of the fibrillar polymer material in crazes. REFERENCES 1. A. J. Kinloch and R. J. Young, Fracture Behavior of Polymers (Elsevier, London, 1983). 2. J. G. Williams, Fracture Mechanics of Polymers (Halsted, New York, 1984). 3. R. P. Kambour, J. Macromol. Sci., Rev. Macromol. Chem. 7, 1 (1973).
4. H. H. Kramer, Adv. Polym. Sci. 52, 1 (1983). 5. A. L. Volynskii and N. F. Bakeev, Solvent Crazing of Polymers (Elsevier, Amsterdam, 1995). 6. R. L. Estevez, E. Tijssens, and V. Giessen, J. Mech. Phys. Solids 48, 2585 (2000). 7. J. S. Vrentas, C. M. Vrentas, and A. H. Bhombal, Polym. Eng. Sci. 24, 965 (1984). 8. S. V. Vlasov and V. N. Kuleznev, Oriented State of Polymers (Znanie, Moscow, 1987) [in Russian]. 9. M. C. Boyce, D. M. Parks, and A. S. Argon, Mech. Mater. 7, 15 (1988). 10. I. M. Ward, in Proceedings of First Symposium on Oriented Polymers, London, 1997, p. 25. 11. M. Dettenmaier and H. H. Kausch, Polym. Bull. (Berlin) 3, 571 (1980). 12. R. F. Nigel and E. J. Kramer, Polymer 22, 691 (1981). 13. V. A. Lapshin, Candidate’s Dissertation in Chemistry (Moscow, 1977). 14. K. Hong and G. Strobl, Macromolecules 39, 268 (2006). 15. J. S. Vrentas, C. M. Vrentas, A. H. Bhombal, and M. S. Lahalih, Polym. Eng. Sci. 24, 956 (1984). 16. E. B. Gowd, C. B. Ramesh, S. M. Mark, et al., Polymer 45, 6707 (2004). 17. S. A. Jabarin, Polym. Eng. Sci. 32, 1341 (1992). 18. C. M. Roland and M. F. Sonnenschein, Polym. Eng. Sci. 31, 1434 (1991). 19. P. Chandran and S. Jabarin, Adv. Polym. Technol. 12, 133 (1993). 20. T. Nishino, A. Nozawa, M. Kotera, and K. Nakamaea, Rev. Sci. Instrum. 71, 2094 (2000). 21. A. M. Donald, E. J. Kramer, and R. A. Bubeck, J. Polym. Sci., Part B: Polym. Phys. 20, 1129 (1982). 22. A. M. Donald and E. J. Kramer, Philos. Mag. 43, 857 (1981). 23. Y. Kisaragi, I. Tomohiro, and T. Yoshihiro, Int. J. Mech. Sci. 45, 1863 (2003).
POLYMER SCIENCE
Series A
Vol. 49
No. 8
2007