J O U R N A L OF M A T E R I A L S S C I E N C E L E T T E R S 12 (1993) 598-601
Double endothermic and exothermic transitions during heating and cooling of thermoreversible polyacrylonitrile gels Z. BASHIR
Courtaulds Research, 72 Lockhurst Lane, Coventry CV6 5RS, UK
Polyacrylonitrile [-CHz-CH(CN)-]n (PAN) does not dissolve in propylene carbonate (PC) at room temperature. It is soluble in PC only above about 130-150°C. Upon cooling very dilute solutions (<1% PAN) from elevated temperatures, "single crystals" of the polymer are formed . In fact, PC is one of the few solvents from which PAN single crystals can be grown [2-4]. In recent work the behaviour of concentrated P A N - P C solutions has received attention. In the first investigation of the series it was shown that cooling a transparent 20% P A N - P C solution led to a cream-coloured gel . X-ray diffraction experiments on these gels indicated the presence of crystallites . Thus, gelation of concentrated P A N - P C solutions was attributed to partial crystallization of the polymer upon cooling , leading to a solvent-swollen network. It is well known that when dry PAN powder is heated in a differential scanning calorimeter (DSC) there is no sign of melting [6, 7]. Instead, an intense exotherm occurs at about 320 °C, which has been attributed to various chemical reactions such as cyclization and cross-linking [6, 7]. Hence, it has been stated in the literature that PAN decomposes before it melts . However, in the previous work on 20% P A N - P C gels , DSC studies revealed a gel "melting" endotherm upon heating; upon cooling the solution an exothermic transition attributable to crystallization occurred. Thus, a diluent such as PC depresses the melting point of PAN, so that transitions with measurable latent heats become observable. The only other DSC study that has suggested that PAN can show dissolution endotherms and crystallization exotherms is that by Frushour, using water as a plasticizer [9, 10], Thus, from the results of these two separate investigations [5, 9, 10] it is clear that although PAN is considered to have only twodimensional rather than a true (three-dimensional) crystalline order, it can show genuine first-order thermodynamic transitions with associated latent heats. Whereas 20% PAN gels could be made easily by dissolving the PAN powder in hot PC and then cooling, it is much more difficult to form solutions with higher PAN content. In a subsequent study an alternative technique was found for preparing concentrated solutions . PAN powder produced by slurry polymerization was very porous and it was found that if liquid PC was added while mechanically grinding the powder, the PC could be adsorbed in 598
the manner that a porous sponge soaks up liquid . The PC-plasticized powders could then be compression moulded at elevated temperatures, thus "melting" the PAN; upon cooling, a rubbery gel film was obtained. In this way, solvent-containing gel films with PAN:PC compositions of 30:70, 40:60, 50:50, 60:40 and 70:30 were prepared . A detailed X-ray study was conducted on the gel films . It was found that the diffraction pattern of the PC-containing films had several new peaks compared with that of the original dry PAN powder. Furthermore, the new diffraction pattern was obtained only when the solvent was present. Hence, it was postulated that the solvent was co-crystallizing with the polymer . Although the DSC behaviour of 20% P A N - P C gels was reported in the first paper , the thermal behaviour of the gel films with higher PAN concentrations was not discussed in the subsequent investigation . It is thus the purpose of this letter to report the thermal behaviour of 30:70, 40:60, 50:50, 60:40 and 70:30 PAN:PC gel films, as there appear to be some unusual features which were not observed with the 20% P A N - P C gels. The details of sample preparation may be found in the previous X-ray diffraction study on the gel films . The thermal behaviour of these gel films is shown in Figs 1-3. Certain PAN:PC compositions (30:70, 40:60, 50:50 and 60:40) showed double endotherms during heating and double exotherms during cooling (Figs 1-3). This was not found previously with 20% P A N - P C gels, where only single endotherms and exotherms were recorded in the heat-cool cycle . Similarly, Frushour's DSC studies on water-plasticized PAN showed only single peaks [9, 10]. The behaviour of the gel films is illustrated best with the 50:50 composition. Double endotherms and double exotherms, as shown in Fig. 1, were sometimes obtained. Alternatively, a double endotherm upon heating and a single exotherm upon cooling occurred as indicated in Fig. 2. The second alternative, a single endotherm and a double exotherm, could also be obtained as shown in Fig. 3. Each type showed reproducibility over several heat-cool cycles, but it was not clear why samples cut from the same gel film gave the different traces shown in Figs 1-3. On the other hand, compositions with PAN content <30% and those with >70% manifested only single endotherms and exotherms in every trial of several. Hence, it is possible that there is a concentration range in which the multiple peak 0261-8028
Figurel The (a) heating and (b) cooling curves of a 50:50 PAN:PC gel film for heating and cooling rate 10 °Cmin 1. There were two endotherms on the heating run and two exotherms on the cooling curve. This was reproducible over several heat-cool cycles.
Figure2 The (a) heating and (b) cooling curves of a 50:50 PAN:PC gel film for heating and cooling rate 10 °Cmin-1 . There were two endotherms on the heating run and one exotherm on the cooling curve. This was reproducible over several heat-cool cycles.
behaviour is shown. Although the cause of these double endotherms and exotherms has not been established, various possibilities are discussed below with a view to future clarification. Multiple melting-endotherms are commonly found in the melting of crystallizable polymers . A frequent reason is that the material recrystallizes (causing lamellar thickening) and remelts during the DSC experiment. This can lead to several shoulders, or even to two well-separated melting peaks . In one study it was shown that dual melting-endo-
Figure3 The (a) heating and (b) cooling curves of a 50:50 P A N : P C gel film for heating and cooling rate 10 ° C m i n -1. There was one endotherm on the heating run and two exotherms on the cooling curve. This was reproducible over several heat-cool cycles.
therms in drawn nylon 6 were caused by melting, recrystallization and remelting of an originally single population of crystals . For poly(butylene terephthalate) three melting endotherms have been observed; Nichols and Robertson attributed the high-temperature peak to crystals formed during the scanning process, and the medium- and low-temperature peaks to crystals originally present in the material . Note that there was only one crystallization exotherm at all heating rates when the melt was cooled . Multiple peaks can also arise for morphological reasons. For example, melt-processed polyethylene with a shish-kebab morphology (composed of oriented fibrils with lamellar overgrowths) shows several melting peaks on the first heating run. This was attributed to the extended-chain material melting at a higher temperature than the lamellar overgrowths [15, 16]. Once this morphology has been destroyed by the first melting, the material recrystallizes in spherulitic form and, upon subsequent melting, it shows only one peak. In spherulitic poly(ether ether ketone) the presence of two melting peaks has been attributed to two different populations of crystals located within the spherulites, rather than to reorganization during melting . Another example is the multiple melting peaks in poly(butylene terephthalate) cited above: this polymer can crystallize into two different types of spherulites, and some authors have attributed each kind of spherulite to a different melting peak ; thus, with this polymer multiple melting peaks may arise due to crystal-size effects  or for morphological reasons , or both reasons. Polymorphism in the crystalline state has been cited as a reason for multiple melting peaks in some polymers such as isotactic polypropylene [19, 20]. 599
Mesophase formation can be yet another cause for multiple peaks. One class exhibiting mesophase formation is the thermotropic liquid-crystalline polymers with rigid mesogens in the chain backbone . In certain flexible-chain polymers, Grebowicz and Wunderlich have classified the existence of a "condis" mesophase . An example is trans-l,4polybutadiene which, when heated, shows too wellseparated endotherms at 74 and 140 °C; the first is attributed to the crystal-condis mesophase transition and the second to the condis mesophaseisotropic transition. Upon cooling, two exotherms are observed at 117 and 51 °C, which are due to the same transitions in reverse order . Annealing effects can, of course, be superimposed over all other possible factors cited. In the instances where double melting-endotherms have been observed, and attributed to annealing effects during the DSC experiment, only one crystallization exotherm is observed when the sample is cooled after the melting . Furthermore, when annealing effects are present during the heating scan, sometimes there is a crystallization exotherm before the melting endotherm [14, 17]. In the case of the P A N - P C gel films, the interesting feature is that double peaks can occur not only during the heating, but also during the cooling cycle. In no case did a crystallization exotherm precede a melting endotherm during a heating cycle (Figs 1-3). Although Fig. 2 is typical in form for cases in which annealing during the experiment occurs [14, 17]. Figs 1 and 3 suggest that reasons other than annealing must also be present in the case of the P A N - P C gel films. An additional complication is introduced in the P A N - P C gel films by the presence of solvent. The formation of shish-kebabs can be discounted, as the polymer was not subjected to any significant elongational flow during compression moulding. Although P A N - P C solutions gel when cooled, due to the partial crystallization of the polymer , one possibility that could occur is liquid-liquid (L-L) demixing, preceding the crystallization: that is, two consecutive phase transitions. Stoks and co-workers have proposed that this occurs during the gelation of poly(vinyl alcohol)-ethylene glycol (PVAL-EG) solutions [23, 24]. Thus, upon cooling a P V A L - E G solution, a fast L - L demixing occurred first, which led to concentrated microdroplets in a dilute matrix, and this was followed by rapid crystallization of the concentrated domains [23, 24]. However, Stoks et al.  only mentioned single gel melting and crystallization peaks with P V A L - E G gels. It could be argued that a similar L - L demixing and crystallization occurs during the cooling of the P A N - P C solutions. The higher-temperature exotherm in the cooling curve of Fig. 1 or Fig. 3 could then be attributed to the L - L demixing (assuming that this was exothermic) and the lower-temperature exotherm to the crystallization. However, the heat exchanged in L - L demixing should be much less than that during crystallization . Figs 1 and 3 show that in the cooling curves, the higher-tern600
perature peaks always had a larger area and hence a greater enthalpy. Thus, it is unlikely that the two peaks in the cooling curve were due to L - L demixing followed by crystallization. It should also be noted that the freezing point of pure PC is -55 °C, so crystallization or melting of the solvent would not be responsible for the double peaks in the temperature range shown in Figs 1-3. The other possible causes for the behaviour shown in Figs 1-3 are polymorphism in the crystalline state or mesophase formation. In the previous roomtemperature X-ray diffraction study on PAN-PC gel films  it was suggested that there was the possibility of crystallization with solvent present in the lattice. Subsequent work has also supported this conclusion . Hence, there are grounds for believing that there may be interesting phase transformations associated with the thermal transitions in these P A N - P C gel films. In order to resolve the cause for this unusual melting and crystallization behaviour in P A N - P C gel films, it would be necessary to conduct a simultaneous DSC and X-ray experiment. It is hoped that such experiments would clarify whether there are two crystalline polymorphs, or a mesophase and crystalline phase that are formed, and whether L - L demixing occurs before crystallization.
Acknowledgements The author thanks Mr R. Litchfield for running the samples in the DSC experiments and Dr D. M. Price for discussions.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
13. 14. 15. 16. 17. 18.
v . F . H O L L A N D , S. B. MITCHELL, W. L. H U N T E R and P. H. L I N D E N M E Y E R , J. Polym. Sci. 62 (1962) 145. R. C H I A N G , J. H. R H O D E S and V. F. H O L L A N D , ibid. A3 (1965) 479. W. F E S T E R , in "Polymer handbook" 3rd Edn, edited by J. Brandrup and E. H. Immergut (Wiley, 1989) p. V/58. F. K U M A M A R U , T. K A J I Y A M A and M. T A K A Y A NAGI, J. Cryst. Growth 48 (1980) 202. Z. BASHIR, J. Polym. Sci. Phys. inpress. N. GRASSIE and R. M c G U C H A N , Eur. Polym. J. 6 (1970) 1277. Z. B A S H I R , Carbon 29 (1991) 1081. G . H . O L I V g and S. OLIVIa, Adv. Polym. Sci. 32 (1979) 123. B . G . F R U S H O U R , Polym. Bull. 4 (1981) 305. Idem, ibid. 7 (1982) 1. Z. BASHIR, Polymer inpress. J. RUNT and I. R. H A R R I S O N , in "Methods of experimental physics", Vol. 16, Pt B, edited by R. A. Fava (Academic Press, 1980) p. 287. M. T O D O K I and T. K A W A G U C H I , J. Polym. Sci. Phys. Ed. 15 (1977) 1067. M . E . NICHOLS and R. E. R O B E R T S O N , ibid. 30 (1992) 755. Z. B A S H I R , J. A. O D E L L and A. K E L L E R , J. Mater. Sci. 21 (1986) 3993. Z. BASHIR and A. K E L L E R , Colloid Polym. Sci. 267 (1989) 116. D. C. BASSETT, R. H. OLLEY and 1. A. M. AL R A H E I L , Polymer 29 (1988) 1745. R . S . STEIN and A. MISRA, J. Polym. Sci., Phys. Ed. 18 (1980) 327.
19. 20. 21. 22. 23.
J . L . KARDOS, A. W. CHRISTIANSEN and E. BAER,
J. Polym. Sci. 4(A2) (1966) 777. K.D. PAE, ibid. 6(A2) (1968) 657. J. GREBOWICZ and B. WUNDERLICH, J. Polym. Sci., Phys. Ed. 21 (1983) 141. Idem, in "Advances in polymer science", Vols 61 and 62 (Springer-Verlag, Berlin, 1984). W. STOKS, H. BERGHMANS, P. MOLDENAERS and J. MEWIS, Brit. Polym. J. 20 (1988) 361.