Double yield points in poly(tetramethylene terephthalate) and its copolymers under tensile loading.

INTRODUCTION

Double yield under tensile loading was reported by R. Seguela and F. Rietsch with heat pressed polyethylene specimens (1). I. M. Ward et al. investigated the phenomena using three polyethylene samples with different degrees of short-chain branching (2). They observed double yield points for three samples. The first yield point marked the onset of 'plastic strains' which are slowly recoverable at least in part. Deformation beyond the second yield point is effectively irrecoverable and was associated with a sharp necking of the samples. The proposed Ward and Wilding model (3), which consisted of two non-linear Maxwell elements in parallel, interpreted the phenomena mechanically.

We have found that poly(tetramethylene terephthalate), (PTMT), and its copolymers also show double yielding prior to necking. To further elucidate the relation between crystallinity and double yielding, the effect of annealing on the stress-strain curves and several mechanical and thermal properties of injection molded specimens of poly(tetramethylene terephthalate) are also discussed in this paper.

MATERIALS

Poly(tetramethylene terephthalate) and its copolymers were kindly supplied by Mr. T. Sugita and Mr. T. Katsuura of Mitsubishi Kasei Corporation. Copolymers contain 4,4[prime]-isopropylidenebis[2,6-dibromophenoxy)ethoxy-2-ethanol] as a comonomer. Bromine contents (comonomer contents) of the copolymer 1 and copolymer 2 are 6 w% (4.1 mol%) and 12 w% (8.2 mol%) respectively.

EXPERIMENTAL

Sample pellets were dried at 120 [degrees] C under nitrogen over night before injection molding with a Nissei FS75 injection molding machine. A cylinder temperature of 260 [degrees] C and mold temperature of 40 [degrees] C were utilized. A rather low mold temperature was utilized to make the specimen less crystalline. The molding cycle consisted of 15 s of injection time and 20 s of cooling time. The size of specimens and their arrangement in the mold are shown in Fig. 1.

The molded samples were then annealed at 180 [degrees] C (homopolymer and copolymer 1) or 150 [degrees] C (copolymer 2) for 0.5, 2, 8, and 24 h under reduced pressure. A lower annealing temperature was utilized for specimens of copolymer 2 due to its low Tm of 201 [degrees] C. Tm of specimens of homopolymer and copolymer 1 are 224 and 216 [degrees] C respectively. All specimens were placed under one atmosphere at 23 [degrees] C and a relative humidity of 50% for at least 24 h prior to the mechanical measurements. Mechanical properties were measured at this ambient condition.

Tensile measurements were performed according to ASTM D-638. Tensile properties were measured with a Shimadzu Autograph S-500 and S-2000 testing machine. Configuration of a test piece is shown in Fig. 1. Sample displacement was 115 mm. Samples were deformed at 5 mm/min (7.3 x [10.sup.-4] [s.sup.-1]). Five specimens were used for each measurement and the results were averaged.

Flexural measurements were performed according to ASTM D-790. Flexural properties were measured with a Shimadzu Autograph S-500 testing machine. Configuration of a test piece is shown in Fig. 1. Sample displacement was 100 mm. Samples were deformed at 3 mm/min at their center. Five specimens were used for each measurement and the results were averaged.

Izod impact strength was measured according to ASTM D-256. A Toyo-seiki universal impact tester was used for the measurements. Notched samples with 12.7 mm (1/2 in) and 3.2 mm (1/8 in) in width were used for the measurements. Configuration of test pieces is shown in Fig. 1.

[TABULAR DATA FOR TABLE 1 OMITTED]

Intrinsic viscosity was measured with an Ubbelohde viscometer. A mixture of phenol and 1,1,2,2,-tetra-chloroethane (1 + 1 in weight) was used as solvent. Samples were dissolved at 110 [degrees] C for 30 min and then cooled to room temperature to give 1 g/dl solution. The cooled solutions were clear yellow to pale yellow. Measurements were performed at 30 [degrees] C. Six successive intervals during flow were averaged to calculate the intrinsic viscosity.

Density was measured with a Mirage electronic densimeter ED-120T which measures density through the difference between weights in air and water. Five specimens for the izod impact test (3.2 mm in width) were utilized for a measurement unless otherwise specified.

A Seiko DSC 220 equipped with a SSC5200H thermal analysis station was utilized for DSC measurements. The following sequence was utilized for these measurements:

Heating 0 to 260 [degrees] C heating rate: 20 [degrees] C/min 260 [degrees] C and held for 5 min Cooling 260 to 30 [degrees] C cooling rate: 20 [degrees] C/min

RESULTS AND DISCUSSION

Results of the measurements of mechanical properties are compiled in Table 1. Intrinsic viscosity, density, and thermal properties which was measured by DSC are listed in Table 2.

Annealing of the samples affected the mechanical and thermal properties of PTMT homopolymer and copolymers. The test changed from ductile to brittle through annealing. Tensile elongation at break and izod strength decreased with annealing. At the same time, tensile and flexural modulus increased with annealing. A substantial part of these changes occurred during the first 2 h of annealing. Changes during 2 to 24 h of annealing were slower than the changes during the first 2 h.

[TABULAR DATA FOR TABLE 2 OMITTED]

Typical stress-strain curves of annealed PTMT homopolymer, copolymer 1 and copolymer 2 are shown in Figs. 2, 3, and 4 respectively. Sudden decreases in tensile stress were observed at necking. One reason for the sudden decrease of stress in these three Figures is that stress (Y-axis) is calculated using the initial width and thickness of the test pieces. Changes of sample size during tensile measurements are not taken into account. Also, stress-strain curves over 15% In strain (X-axis) of PTMT homopolymer and 10% in strain of copolymer 1 and copolymer 2 are not shown in these Figures so that the shapes of the curves before necking are more clearly demonstrated. Therefore, break points of several samples do not appear in these Figures.

Two yield points were observed before necking of these test pieces. Second yield points became larger and the first yield points became smaller with extended annealing time with all three samples. Substantial change in sample width, like that which was observed at necking, was not observed at the first yield points. It must be noted that a second yield point was not observed with copolymer 2 before annealing.

These changes are considered mostly due to the change in crystallinity. Macromolecular materials often become brittle with degradation and/or cross-linking but these phenomena are not the primary cause in this case.

Substantial degradation did not occur in this experiment. Degradation should have decreased the intrinsic viscosity. On the contrary, intrinsic viscosity increased slowly as a result of annealing because of the solid state polycondensation reaction of polyester which occurs under vacuum or inert gas atmosphere (5-8). The solid state polycondensation reaction is usually observed at 185 to 200 [degrees] C (4). Since the annealing temperature was rather low, the reaction was rather slow. The low annealing temperature of 150 [degrees] C of copolymer 2 which was 30 [degrees] C lower than copolymer 1 and PTMT homopolymer resulted in a slower increase of intrinsic viscosity of copolymer 2 as a result of annealing. Also, the slow but steady increase in intrinsic viscosity does not correspond to the pattern of changes in mechanical properties. Therefore degradation cannot explain the change in mechanical properties.

Crosslinking does not explain the phenomena either. As is shown in Table 2 and Figs. 5 to 7, the melting point (Tm: peak temperature of DSC curves) did not change with annealing. If crosslinking occurred during annealing the Tm should change with annealing or the material might undergo thermal degradation not melting. Crystallization temperature during cooling (Tcc: bottom temperature on DSC curves) and the heat of crystallization did not change with annealing. This means that the annealed specimens behave almost the same thermally as completely melted material. Also, the result that the sample solubility for intrinsic viscosity measurements didn't change supports the nonexistence of substantial amounts of crosslinking in the annealed samples. If there were substantial amounts of crosslinked material a gel or insoluable material should have been observed.

Since crosslinking and degradation did not occur substantially during annealing under reduced pressure, the main factor which affected the mechanical properties should be the change in crystallinity of the specimens. The results of density measurements and DSC measurements support this.

Crystallinity can be calculated from density or heat of fusion. Crystallinity (Xc) and bromine content ([Br.sub.o]) affect density (d) according to the following equation.

1/d = [X.sub.c]/1.404 + (1 - [X.sub.c)2/1.28(1 - [X.sub.c]) + 0.008[Br.sub.o]

Assumptions made and the derivation of the equation are described in the Appendix.

The substantial change in crystallinity occurred during the first two hours of annealing. This tendency coincides with the changes in mechanical properties and the double yielding phenomena. Since the comonomer retards the initial crystallization, the crystallinity of injection molded specimens decreases with the increase of bromine content. This is the reason why copolymer 2 showed the largest change in crystallinity during the first 2 h of annealing even with a lower annealing temperature. Since PTMT homopolymer can crystallize faster, a substantial part crystallized during the injection molding process which had about 20 s of cooling time. Therefore PTMT homopolymer would increase its crystallinity less during annealing. Copolymer 1 showed crystallization intermediate between PTMT homopolymer and copolymer 2. The tendency also coincides with the mechanical properties of annealed PTMT homopolymer and copolymers.

Crystallinity of PTMT homopolymer can be calculated according to the following equation (9).

[X.sub.c] = Heat of fusion/1.4545(J/g)

As is shown in the Appendix, the crystalline structure of PTMT homopolymer and copolymers are the same. Thus the equation should be applicable to copolymers. The calculated crystallinity values are shown in Table 2. The tendency is the same as the values calculated from density. However, the crystallinity values are larger than that calculated from density. Crystallization during measurements (heating) would make the values of the heat of fusion larger. This explains the larger crystallinity values calculated from the heat of fusion.

DSC curves during heating of annealed PTMT homopolymer, copolymer 1, and copolymer 2 are shown in Figs. 5, 6, and 7 respectively. The heat of fusion during heating increased with annealing. Also, small peaks appeared near the annealing temperature. They migrate to higher temperature with extended annealing. The higher peaks relate to stable PTMT crystallites and the lower peak relates to the smaller PTMT crystallites which evolved during annealing. This demonstrates the increase crystallinity during annealing which is considered the main cause of the changes in mechanical properties.

Since the second yield point in stress-strain curves becomes more apparent with the increase of crystallinity during annealing, the first yield point must indicate the yield in amorphous regions and the second yield point must indicate the yield of crystallites. The lack of a second yield point in the amorphous copolymer 2 specimen before annealing supports the assignment of the yield points.

Assuming that the stress-strain curves of PTMT homopolymer and copolymers can be described using the Ward and Wilding model (2, 3), the stress-strain relationship can be modeled with two parallel nonlinear Maxwell elements. We consider that the nonlinear Maxwell elements represent the behavior of the amorphous and the crystalline regions of the specimens. Since a Maxwell element under small strain behaves like a spring, a stress-strain relationship of PTMT homopolymer and copolymers should be modeled as two parallel springs. Therefore, the tensile modulus of the specimen can be described as:

[Mathematical Expression Omitted]

Here [X.sub.c]: Crystallinity (0 [less than] [X.sub.c] [less than] 1)

[E.sub.c]: Tensile modulus of amorphous regions

[E.sub.c]: Tensile modulus of crystalline regions

[E.sub.t]: Observed tensile modulus of specimens

If [X.sub.c] = 1 then [E.sub.t] = [E.sub.c]. This means that when the crystallinity and the tensile modulus are plotted and [X.sub.c] is extrapolated to 1, [E.sub.t] becomes the modulus of the crystallites. Results of our calculation are shown in Table 1, Figs. 8 and 9. The estimated tensile modulus at 100% crystallinity was 3.6 to 3.9 GPa.

As is shown in the Appendix, according to wide angle X-ray diffraction (WAXS) the crystal structure is the same in PBT homopolymer and the copolymers. Our tensile modulus of crystallites is also similar to that of PTMT homopolymer which is 3 to 5 GPa after Ward et at. (10). Also reported are a crystalline tensile modulus of 2 GPa for the a and b axes, 13 GPa for the c axis of the alpha phase, and 21 GPa for the c axis of the beta phase (11). These values agree with Ward's result since the tensile modulus of crystallites of the unoriented bulk specimens should be a weighted average of the tensile modulus of the axes. The estimated tensile modulus of crystallites is similar in the homopolymer and the copolymers, whereas the estimated tensile modulus of the amorphous regions increases with increase of bromine content [ILLUSTRATION FOR FIGURE 9 OMITTED]. These results support the segregation of comonomer in amorphous regions, which we expected from WAXS results. PTMT crystallites undergo an alpha to beta transition with tensile strain (11-13). 97% of crystallites in oriented fibers of PTMT homopolymer change their crystalline structure from alpha to beta at 12% strain (11). Since necking occurred at 12% strain in the PTMT homopolymer, PTMT crystallites in injection molded specimens must have changed to the beta structure before necking. The second yield point in PTMT homopolymer and its copolymers under tensile loading must be the result of the alpha to beta crystalline transition of PTMT crystallites.

CONCLUSION

Double yield points before necking were observed in injection molded specimens of poly(tetramethylene terephthalate) (PTMT) and PTMT copolymers which contain 4,4[prime]-isopropylidenebis[(2,6-dibromophenoxy)ethoxy-2-ethanol] under tensile loading. The first yield is associated with the deformation of amorphous regions and the second yield is associated with the alpha to beta transition of PTMT crystallites.

The first yield point became less apparent with the increase of crystallinity of specimen. The second yield point became more apparent with the increase of crystallinity. Annealing of injection molded specimens increased the crystallinity and made the second yield point on stress-strain curves more apparent. Copolymerization decreased the crystallinity and made the first yield point more apparent. Amorphous copolymer 2 did not show a second yield point. Crystallinity of the specimen were calculated from density and bromine content of the copolymer.

The tensile modulus of amorphous material and crystallites are calculated. The tens fie modulus of amorphous material increased from 1.10 to 1.70 GPa with increase of comonomer from 0 to 12 w%. The tensile modulus of crystallites was 3.64 to 3.87 GPa and was not affected by the comonomer content. This result supports our assignment of the yield points.

Effects of annealing on mechanical and thermal properties of the specimens were also measured. The specimens changed their properties from ductile to brittle during annealing. Their change during annealing was mainly attributed to the increase of crystallinity and not to thermal degradation and/or crosslinking.

APPENDIX

Crystallinity of Copolymerized Poly(Tetramethylene Terephthalate)

INTRODUCTION

Crystalline polymers increase their density with increase of crystallinity since crystallites have higher density than amorphous regions. Therefore the crystallinity of homopolymer can be calculated from density. At the same time, the copolymer has different compositions in the crystalline and amorphous regions. Thus some consideration of this segregation is necessary.

Crystallinity of poly(tetramethylene terephthalate) copolymers, copolymerized with 4,4[prime]-iso-propylidenebis[(2,6-dibromo-phenoxy)ethoxy-2-ethanol], was calculated from density and bromine content according to the following method.

MATERIALS AND SAMPLE PREPARATION

Materials used in this study are described in the main section of this paper.

Heat pressed specimens with thicknesses of 1 mm were used for the wide angle X-ray scattering (WAXS). Heat pressed specimens were used since the thickness of 1 mm was appropriate for WAXS measurement. They have lower crystallinity than injection molded specimens. Sample pellets were dried over night at 120 [degrees] C under reduced pressure. The dried pellets were placed in a mold which was made of two aluminum sheets with a spacer between them. The spacer was made from stainless steel 15 cm by 15 cm inside a 1 mm in thickness. The mold and pellets were heated in a heat press at 245 [degrees] C, then quenched in cold water. Aluminum sheets were removed by immersing the quenched sheet in a warm 10% NaOH solution to dissolve them. The heat pressed sheet was then washed with water and dried at room temperature. Copolymer 2 gave a pale yellow transparent sheet. Copolymer 1 gave a slightly opaque sheet. PTMT homopolymer gave a white opaque sheet. Sample sheets were cut into several pieces and annealed under reduced pressure. PTMT homopolymer and copolymer 1 were annealed at 200 [degrees] C. Copolymer 2 was annealed at 150 [degrees] C.

EXPERIMENTAL

Wide angle X-ray scattering was measured using heat pressed specimens. X-ray photographs were taken with a toroid camera under reduced pressure. Sample sheets were set perpendicular to the X-ray beam. Nickel filtered Cu K-alpha radiation (wave length - 1.5418[Angstrom] was used. Three Kodak DFE films were exposed for 9 h at the same time. The diffraction pattern showed no orientation perpendicular to the X-ray beam. The photograph was digitized with an Optronics P-100 system in the transmission mode. The intensity arrays were analyzed on a AEM512 color graphics terminal (12).

Density was measured with a Mirage electronic densimeter ED-120T. Transparent pellets of copolymer 1 and copolymer 2 as received were used for density measurement as the amorphous or low crystallinity samples. These sample pellets were used without drying to avoid crystallization which might occur at elevated temperature. Heat pressed sheets were cut into pieces to fit in the densimeter. The density of the heat pressed specimen was measured before annealing.

DSC measurements were made according to the method which is described in the main section of this paper.

RESULTS AND DISCUSSION

Results of density measurements are listed in Table A1. Results of WAXS measurements are shown in Figs. A1 and A2. Results of the DSC measurements are shown in Fig. A3.

In the following calculation, pellets of copolymer 1 and copolymer 2, and injection molded and heat pressed specimens of copolymer 2 are considered to be amorphous. Heat pressed specimens of copolymer 1 are considered to be low in crystallinity since they were slightly opaque and higher in density than the pellets. The amorphous materials were transparent and, in the case of copolymer 2, the heat pressed sheets showed the same density within measurement error as the pellets.

Table A1. Density of PTMT Homopolymer and Copolymers.

Br content           d                      Sample

0                 1.280        amorphous PTMT homopolymer (13, 17)
0                 1.404        crystallite of PTMT homopolymer (14)
6                 1.324        copolymer 1 pellet before molding
6                 1.340        copolymer 2 heat pressed sample
12                1.375        copolymer 2 injection molded sample
12                1.377        copolymer 2 pellet before molding
12                1.378        copolymer 2 heat pressed sample

The results of DSC measurements agree with the low crystallinity of these specimens. Results of DSC measurements of transparent copolymer 2 samples (injection molded specimens before annealing, pellet heat pressed sample) are shown in Fig. A3. These three samples have an exothermic peak just over [T.sub.g]. This exothermic peak disappears when the sample is annealed at 150 [degrees] C for 0.5 h. These results suggest that even though these samples show a heat of fusion, it is caused by the melting of crystallites generated during heating above [T.sub.g]. WAXS patterns of heat pressed copolymer 2 specimens annealed at 150 [degrees] C are; shown in Fig. A1. Without annealing, only amorphous scattering is seen. Reflections which can be attributed to PTMT crystallites were not observed.

To calculate crystallinity from density, we assume that the volume is additive. Then:

1/d = Xc/dc + 1 - Xc/da (A1)

where: d = density of the bulk sample

[X.sub.c] = crystallinity of the bulk sample

dc = density of the bulk sample

da = density of amorphous region

As is shown in Fig. A2, WAXS peaks of crystallites are at the same position as in the PTMT homopolymer, copolymer 1 and copolymer 2. Thus the crystal structure of PTMT homopolymer and copolymers are considered to be identical. Therefore, we assume that the comonomer cannot crystallize and is concentrated in the amorphous region. Then:

Br = [BR.sub.0]/1 - Xc (A2)

where:

Br = Br content in amorphous region

[Br.sub.0] = Br content in the bulk sample

(copolymer 1 = 6%, copolymer 2 = 12%)

Thus:

da = [da.sub.0] + [dBr.sub.0]/1 - Xc (A3)

where: [da.sub.0] = density of amorphous homopolymer

b = constant

The constant b correlates Br content and the density of the amorphous region. The b value was calculated according to the density values of amorphous samples which are listed in Table A1. The result is shown in Fig. A4.

b = 0.0008

Therefore inserting Eq A3 into Eq A1:

1/d = Xc/dc + 1 - Xc/da0 + [dBr.sub.0]/1 - Xc

= Xc/dc + [(1 - Xc).sup.2]/[da.sub.0](1 - Xc) + b[Br.sub.0] (A4)

Density of crystallite and amorphous regions of PTMT homopolymer (dc and [da.sub.0]) has been reported by several groups (13-17). The values are listed in Table A2. The following values were used in this calculation.

dc = 1.404 after Ref. 14 [da.sub.0] = 1.280 after Ref. 13, 17

Therefore,

1/d = Xc/1.404 + [(1 - Xc).sup.2]/1.28(1 - Xc) + b[Br.sub.0] (A5)

Equation A5 is applicable until the amorphous region becomes a homopolymer of the comonomer and terephthalic acid. Since the formula of the homopolymer is C31H31O8Br4, the bromine content is 37.6%. Thus, in the case of copolymer 1 the upper limit of the Xc value is

37.6 = 6/1 - Xc Xc = 0.894

In the same way, Eq A5 yields an upper limit Xc of 0.68 for copolymer 2.

The relation between crystallinity and density, calculated with Eq A5, is shown in Table 2 and Fig. A5.

Table A2. Density of Crystallite and Amorphous Region of PTMT
Homopolymer (11).

      Density (G/[cm.sup.3])
crystallites        amorphous region            References

1.396                     1.28             Boya (13)
1.404                        -             Tadokoro (14)
1.406                        -             Mencik (15)
1.433                        -             Alter & Bonart (16)
1.403                     1.28             Bornschlegl & Bonart (17)

REFERENCES

1. R. Seguela and F. Rietsch, J. Mater. Sci., Letters, 9, 46 (1990).

2. N. W. Brooks, R. A. Duckett, and I. M. Ward, Polymer, 33, 1872 (1992).

3. M. A. Wilding and I. M. Ward, Polymer, 22, 870 (1981).

4. H. Suganuma, H. Ikeuchi, Y. Yamamoto, A. Hirai, and M. Tanaka, Sen-i Gakkaisi, 43, 538 (1987).

5. L. H. Buxbaum, J. Appl. Polym. Sci., Appl. Polym. Symp., 35, 59 (1979).

6. H. D. Dinse and E. Tucek, Acta Polymerica, 31, 108 (980).

7. B. Forruntato, F. Pilati, and P. Manaresi, Polymer, 22, 655 (1981).

8. C. Gostoli, F. Pilati, G. C. Sarti, and B. DiGiacomo, J. Appl Polym. Sci., 29, 2873 (1984).

9. I. Kirshenbaum, J. Polym. Sci., 3, 1869 (1965).

10. J. Roebuck, R. Jakeways, and I. M. Ward, Polymer, 33, 227 (1992).

11. Kazuo Yugi, Ed., Handbook of Saturated Polyester Resins, 295, Nikkann Kogyo Shinbun, Tokyo (1989).

12. R. P. Grasso, B. C. Perry, J. L. Koenig, and J. B. Lando, Macromolecules, 22, 1267 (1989).

13. C. A. Boya and J. R. Overton, Bull. Am. Phys. Soc., Ser. II, 10, 352 (1974).

14. H. Tadokoro et al., Macromolecules, 9, 226 (1976).

15. Z. Mencik, J. Polym. Sci., Polym. Phys. Ed., 13, 2173 (1975).

16. V. Alter and R. Bonart, Colloid Polym. Sci., 254, 348 (1976).

17. E. Bornschlegl and R. Bonart, Colloid Polym. Sci., 258, 319 (1980).