An apparent double glass transition in semicrystalline polymers

Abstract

Thermal expansion and/or specific heat and/or dynamic mechanical loss data reveal the presence of two glass-like transitions in bulk crystallized polyethylene, polypropylene, polybutene-1, polypentene-1, cis-and trans-polyisoprene (natural), poly-4-methylpentene-1, isotactic polystyrene, poly(vinyl alcohol), nylon 6, the oxide polymers ‒(CH₂)_nO‒, with n = 1 to 4, polyethylene terephthalate, polyvinylidene fluoride, polyacrylonitrile, and polyvinylidene chloride. We designate the lower of these as T_g(L), which appears identical with the conventional T_g at zero crystallinity. The higher one, designated as T_g(U), is strongly increased with increasing levels of cystallinity. The differnece ΔT_g = T_g(U) − T_g(L) tends to approach zero as the fractional crystallinity, X, approaches zero. For a X of 0.5 [Ptilde] 0.1, ΔT_g is about 50°C and T_g(U)/T_g(L) is about 1.2 with temperatures in °K. The increases in coefficient of thermal expansion, (Δα)_L and (Δα)_U, at these two transitions seem to depend on crystallinity and morphology in the expected manner for polyethylene and polypropylene: for × = 0.5–0.7, (Δα)_U is stronger than (Δα)_L; for X χ 0, (Δα)_L is stronger than (Δα)_U. Such data are not available for the other listed polymers. Some atactic polymers, poly-4-methylpentene-1, and polystyrene also seem to have a double T_g, the upper of which we tentatively ascribe to the presence of Geil-Yeh types of local order. Since polyethylene, polyvinylidene fluoride, and polyvinylidene chloride exhibit the apparent double T_g, tacticity, per se, is not necessary to produce it. Special care must be exercised to distinguish T_g(U) from the crystalline phase α_c relaxation occurring at temperature T_c. It is shown that T_c for well-annealed crystalline material tends to occur at about 0.83 to 0.85 T_M where T_M is the crystalline melting point in °K. Hence T_c is very close to the temperature at which rate of bulk crystallization is a maximum. While the phenomenon of a double T_g seems clear, its origin is in doubt. We suggest that T_g(L) and T_g(U) arise from tkie presence of different types of amorphous material. For example, polymer molecules not incorporated in the crystallites and/or cilia might give rise to T_g(L). Morphological entities under greater restraint, such as tie molecules or loose loops, might give rise to T_g(U). Conversely, pseudocrystalline structures (smectic or nematic) might be responsible for T_g(U), at least in polypropylene, poly-4-methylpentene-1, and possibly in some nylons. Data available in the literature do not permit making a definite choice between different possible origins of the apparent double glass transition. Indeed, the origin may vary from polymer to polymer.