Conditions for melt crystallization without thermal degradation and equilibrium melting temperature of atactic poly(vinyl alcohol)

Optimum conditions for melt crystallization without thermal degradation for atactic poly(vinyl alcohol) (PVA) were found for the first time. Thermal degradation of PVA in the melt has been too serious to permit crystallization from the melt. However, it was found that thermal degradation can be suppressed by selecting in appropriate solvent for film preparation, such as water without dimethyl sulfoxide (DMSO), by careful degassing before melting and by melting under high vacuum. As a result, the sample could be held in the melt at 250 degreesC for 30 min and in the supercooled melt at 220degreesC for 480 min without thermal degradation. Therefore we could carry out the melt crystallization at a high crystallization temperature (T-c) in the range of 208 to 221 degreesC and study the structures and morphologies of melt-crystallized PVA samples for the first time. The irregularly stacked lamellae were observed irrespective of T-c. The thick lamellae with high T-m, 24 nm and 236.1 degreesC, respectively, were obtained from the melt crystallization at T-c = 220 degreesC, which indicates that the chain sliding diffusion is accelerated a little with increase of T-c. However, the distribution of lamellar thickness was sharp, even for the sample crystallized at high T-c, which indicated that the chain sliding diffusion is relatively suppressed by the intermolecular hydrogen. bonds. Hence, the lamellar thickening was suppressed during the crystallization, as compared with polyethylene (PE) and isotactic polypropylene (iPP). The equilibrium melting temperature (T-m(o)) of PVA was determined to be 249degreesC, applying an improved Gibbs-Thomson plot that was recently proposed by Yamada et al. The Hoffman-Weeks plot accidentally gave a similar T-m(o) = 248degreesC. It was confirmed that the lamellar thickness is inversely proportional to the degree of supercooling in the case of PVA crystals, which is the reason why both plots gave similar T-m(o).