Polyurethane and unsaturated polyester hybrid networks - 2

Polyurethane and unsaturated polyester hybrid networks - 2

(Parte 2 de 2)

The dynamic mechanical properties of these samples were also investigated by creating master curves of shear storage modulus and tan d vs. frequency following the time–temperature superposition (TTS) principle [30] using 1068C as the reference temperature. The shift factor, aT, is defined by the Williams–Landel–Ferry (WLF)

Fig. 3. Storage shear modulus G9 and tan d of EG12 before and after quenching.

equation as shown below:

where c01 and c02 are the shift coefficients and T0 the shift reference temperature. c01 and c02 appeared to be constant for a given chain extender within experimental error. c01 and c0 2 were 14.3 and 168.48C respectively for hybrid networks with EG incorporation, and 12.9 and 165.68C for hybrid networks with HD as chain extender. They are of the same order of magnitude as those found in the literature

[30,31]. The activation energy, Ea, of the shift factor was calculated by the Rheometrics software using an Arrhenius equation of the form:

aT ¼Aa exp ¹Ea

RT (6) where aT is the shift factor and Aa is a constant. All samples showed good correlation, with a correlation factor r 2 between 0.85 and 0.97. The average activation energy for polyurethane and unsaturated polyester hybrid networks, hEaiEG, was 282 kJ mol¹1 with EG incorporation. The average activation energy for hybrid networks with HD as chain extender, hEaiHD, was 242 kJ mol¹1. This is in the same range of activation energy found for polyurethane systems by Senich and MacKnight [26] and by Hartmann and coworkers [31,32]. Changes in the dominant relaxation frequency of the mas- ter curves, f0, have been shown to be directly related to changes in Tg determined by the maximum of tan d [3]. For a phase-mixed polymer system the relationship between log(f0) and 1/Tg is linear. Samples with a higher f0 have a lower Tg. Experimental results are given in Fig. 4 showing log(f0) plotted against 1/Tg. The b transition of tan d in d.m.a. tests was used as the Tg to be consistent with the method. All HD samples were almost superimposed because they had about the same Tg. Polyurethane and unsaturated polyester hybrid networks with low contents of EG as chain extender, i.e., before phase separation, followed a linear relationship. When the system was phase-separated, experimental results did not agree with the linear correla- tion. f0 increased as the Tg of each sample increased. The deviation is due to the inhomogenity of the phase-separated hybrid network. To clarify this point, the same tests were conducted for quenched samples with higher EG contents.

In this case, a higher Tg led to a lower f0 and agreed with the theoretical expectation. A linear relationship was found among all phase-mixed samples. The results of hybrid networks with HD as the chain extender, with EG incorporation lower than 6 wt%, or quenched samples with higher EG incorporation, could be extrapolated by a single straight line. Fig. 5 shows storage shear modulus at the rubber plateau,

G0, vs. the weight fraction of hard segments in the hybrid network. Data obtained from time–temperature superposi- tion experiments gave good confirmation of the results of d.m.a. tests. G0 decreased as the hard segment fraction increased. However, we can note that after phase separation,

G0 determined from shifted data was much larger. This means that G0 was not the ultimate rubber plateau but the intermediate rubber plateau. The TTS principle can be used only with a phase-mixed system, which was not the case for the hybrid network with high content of EG. When the tests were conducted with quenched samples, the experimental results agreed with expectation. All values can be described by a decreasing line as the weight fraction of hard segments increases. No significant difference in G0 was found between EG and HD hard segments.

3.2. Mechanical properties

Typical polyurethane block copolymers are rubbery at ambient temperature. Their mechanical properties are similar to those of elastomeric materials. The modulus is typically of the order of MPa, and the elongation is generally between 500% and 1000% at break [34,35]. The polyurethane and unsaturated polyester hybrid networks are glassy at ambient temperature, and therefore their mechanical properties are different from those of the polyurethane block copolymers. Comparison of results from the Izod impact resistance tests of unnotched specimens is shown in Fig. 6. Hybrid networks with HD incorporation showed an improvement in impact resistance of nearly 20% for increasing content of chain extender from 0 to 10% by weight. Micrographs of the unnotched Izod fracture surface of HD3 and HD10 are shown in Fig. 7(A), Fig. 7(B), respectively. As the HD content increased, the smooth fracture initiation region became smaller while the rough secondary fracture region grew. Secondary cracks were shown throughout the secondary fracture region. The secondary cracks consumed more energy than the smooth initiation region. As the amount of

HD incorporated increased, larger fibrils were pulled out of the secondary fracture surface.

On the contrary, the impact resistance decreased for all samples with EG incorporation. A bigger drop in impact strength was observed when phase-separated hard domains were formed. Since both the soft domains and hard domains exist in the hybrid network, the impact resistance is controlled by the more brittle phase, formed by hard domains. Fig. 7(C), Fig. 7(D) show micrographs of the unnotched Izod fracture surface of EG3 and EG12, respectively. The fracture surface of the polyurethane and unsaturated polyester hybrid networks with EG incorporation, before phase separation occurred, was similar to that of hybrid networks with HD as chain extender as shown in Fig. 7(C), Fig. 7(A). However, Fig. 7(C) shows fewer secondary cracks and smaller fibrils pulled out compared with the fracture surface of the HD samples. Since less energy was consumed by the secondary cracks, hybrid networks with EG as a chain extender had a lower impact resistance. Fig. 7(D) shows that the fracture initiation region was larger and the secondary fracture region was very smooth after phase separation occurred. No fibril pull-out was observed on the fracture surface, which agreed with the very low impact resistance for these samples. The chain length of the chain extender plays an important role in determining the impact resistance of hybrid networks. A more flexible, long chain extender gives higher impact strength, while a more rigid, short chain extender decreases the impact strength.

Quenched samples had a much higher impact resistance and the impact resistance increased with incorporation of the chain extender, with the exception of samples that were initially phase-separated. Results for materials with HD or EG as chain extender were about the same for a given weight fraction of hard segments, within experimental error. The fracture surface of quenched samples showed a small smooth fracture initiation region and a large rough secondary fracture region with both cracks and fibril pullout. Samples that had a two-phase structure before quenching showed a very sharp drop in impact strength. This was due to the flat, round cracks that formed from internal stresses within the materials during quenching. The higher the incorporation of EG, the more numerous the cracks. The flat round cracks were between 1 and 5 m in diameter. Stresses may appear at the interface of the soft and hard domains because of a differential thermal expansivity. The cracks developed from weak areas of the material in order to release internal stresses and thus formed the flat round geometry.

Fig. 7. Micrographs of the unnotched Izod fracture surface of: (A) HD3; (B) HD10; (C) EG3; (D) EG10.

The flexural modulus of hybrid networks vs. weight fraction of total hard segments is given in Fig. 8. The modulus of samples with HD as the chain extender decreased slightly when the HD content increased, while the modulus of samples with EG as the chain extender increased with increasing EG content. As previously discussed, this is due to the chain length of the chain extender. A significant increase in flexural modulus was observed for the samples with phase-separated hard domains. In this case, hard domains form a co-continuous phase with soft domains as described by Ophir and Wilkes [14] and give a higher flexural modulus. The quenched samples had a lower flexural modulus and the modulus decreased as the incorporation of chain extender increased for both sets of samples. Because hard segments do not form a phase-separated hard domain in quenched samples, the apparent flexural modulus of hybrid networks of quenched samples was controlled by the soft domain.

Fig. 9(a), Fig. 9(b) show the stress at yield and strain at yield as a function of the weight fraction of hard segments. Both the flexural stress and strain at yield increased slightly for samples with low incorporation of chain extender. For higher amounts of chain extender, both flexural stress and strain at yield decreased, especially after phase separation. Mekhilef and Verhoogt explained the fact that the co-continuous structure of polymer blends usually possesses weak mechanical properties by virtue of the weak interfacial interaction between the two polymers, although both components are continuous and thus could fully contribute to the properties of the blend [36]. However, quenched samples showed a decrease in yield stress and an increase in yield strength as the content of chain extender increased. Shifts in the flexural stress and strain at yield were also observed before and after quenching. Quenched samples had a lower yield stress and a higher strain than unquenched samples. Flat round cracks were found in quenched samples that had previously been phase-separated. The cracks decreased the flexural stress and stain at yield dramatically.

4. Conclusion

The thermomechanical properties of polyurethane and unsaturated polyester hybrid networks have been investigated by heat distortion analysis and dynamic mechanical analysis. Transition temperatures agreed with the results measured by modulated d.s.c. as reported in a previous study. Multiple transitions shown by d.m.a. for samples with EG contents above 6 wt% were related to the phaseseparated structure. The a transition was related to the phase-separated hard domain and the b transition was related to the soft domain. The rubber plateau of the storage shear modulus was correlated to the crosslinking density. However, theoretical predictions at temperatures lower than the glass transition temperature of the hard domains did not agree with experimental results for the phase-separated hybrid networks. The phase-separated hard domains acted as ‘virtual crosslinks’ and reduced the number-average molecular weight between crosslinks. Mechanical properties at room temperature were generally improved by the incorporation of a chain extender. HD increased the flexibility of polymer chains, resulting in higher deformation and impact resistance of the hybrid networks. Hybrid networks with EG as the chain extender were stiffer by virtue of the rigid hard domains. Hybrid networks with an EG content greater than 6 wt% showed a high

flexural modulus, but poor ultimate mechanical properties because of the formation of phase-separated hard domains.


[1] Malinconico M., Martuscelli E., Volpe MG. Int J Polym Mater 1987;1:295. [2] Martuscelli E, Musto P, Ragosta G, Scarinzi G, Bertotti E. J Polym

Sci—Phys 1993;31:619. [3] Bucknall CB, Davies P, Partridge IK. Polymer 1985;26:109.

[4] Crosbie GA, Phillips MG. J Mater Sci 1985;20:182. [5] Martuscelli E, Musto P, Ragosta G, Scarinzi G. Polymer 1996;37 (18):4025. [6] Cooper SL, Tobolsky AV. J Appl Polym Sci 1966;10:1837.

[7] Seymour RW, Estes GM, Cooper SL. Macromolecules 1970;3 (5):579. [8] Seymour RW, Cooper SL. Macromolecules 1973;6 (1):48.

[9] Cooper SL, West JC, Seymour RW. Encycl Polym Sci Technol 1976;1:521. [10] Srichatrapimuk VW, Cooper SL. J Macromol Sci—Phys 1978;B15 (2):267. [1] Van Bogart JWC, Lilaonitkul A, Cooper SL. Adv Chem Ser 1979;176:3. [12] Paik Sung CS, Hu CB, Wu CS. Macromolecules 1980;13:1.

[13] Ophir ZH, Wilkes GL. Polym Prepr, Am Chem Soc, Div Polym Chem 1978;19 (1):26. [14] Ophir ZH, Wilkes GL. In: Cooper SL, Estes GM, editors. Multiphase polymers, Adv Chem Ser, vol 176. 1976:53. [15] Valette L, Hsu CP. Polymer (in press).

[16] Flory PJ. Principles of polymer chemistry. Ithaca (NY): Cornell

University Press, 1953. [17] James HM, Guth E. J Chem Phys 1947;15:669.

[18] Patel SK, Malone S, Cohen C, Gillmor JR, Colby RH.

Macromolecules 1992;25:5241. [19] Painter PC, Shenoy SL. J Chem Phys 193;9:1409.

[20] ASTM D 256-8. Philadelphia (PA): American Society for Testing and Materials, May 1988. [21] Nielsen LE. Mechanical properties of polymers. New York: Reinhold

Publishing Corporation, 1962:Ch. 6. [2] ASTM D 790-86. Philadelphia (PA): American Society for Testing and Materials, September 1986. [23] Seefried CG Jr, Koleske JV, Critchfield FE. J Appl Polym Sci 1975;19:2503. [24] Melot D, Escaig B, Lefebvre JM, Eustache RP, Laupetre F. J Polym Sci—Phys 1994;32:249.

[25] Hesketh TR, Van Bogart JCW, Cooper SL. Polym Eng Sci 1980;20:190. [26] Senich GA, MacKnight WJ. Adv Chem Ser 1979;176:97.

[27] Pandya MV, Deshpande D, Hundiwale DG. Br Polym J 1987;19:1.

[28] Van Krevelen DW. Properties of polymers. 2nd ed. Amsterdam:

Elsevier Scientific, 1976. [29] Fedors RF. J Polym Sci 1969;C26:189.

[30] Ferry JD. Viscoelastic properties of polymers. 3rd ed. New York: John

Wiley, 1980. [31] Hartmann B, Lee GF. J Non-Cryst Solids 1991;131133:887.

[32] Hartmann B, Lee GF, Lee JD. In: Raju PK, editor. Vibro-acoustic characterization of materials and structures, Am Soc Mech Eng, NCA vol 14. 1992:21. [3] Fedderly J, Lee GF, Ferragut DJ, Hartmann B. Polym Eng Sci 1996;36 (8):1107. [34] Grillo DJ, Housel TL. In: Physical properties of polyurethanes from polyesters and other polyols, Proc. SPI 34th Polyurethane Technical/ Marketing Conference. New York: Society of the Plastics Industry, Inc., 1992:552. [35] Miller JA, Lin SB, Hwang KKS, Wu KS, Gibson PE, Cooper SL.

(Parte 2 de 2)