THE GLASS-TRANSITION TEMPERATURE OF RANDOM COPOLYMERS .2. EXTENSION OF THE GORDON-TAYLOR EQUATION FOR ASYMMETRIC T-G VS COMPOSITION CURVES

It is obvious that asymmetric T-g vs composition curves of random copo lymers cannot be adapted by the Gordon-Taylor equation based on volume additivity, even if instead of the model-specific additivity paramete r KG-T = (rho(1)/rho(2))(Delta alpha(2)/Delta alpha(1)) a curve fittin g parameter, K, is considered. Therefore contributions of diad- and tr iad-sequences to the copolymer T-g were considered. Assuming, by lack of copolymerization kinetics data, that the diad- and triad-sequence d istribution can be related in a first approximation to the respective weight fractions of the copolymer components a concentration power equ ation is deduced for adapting the composition dependence of the glass temperature of copolymers. The two fitting parameters, K-1 and K-2, of this concentration power equation characterize the contributions to t he copolymer T-g of the hetero-diads and -triads, respectively. Values of K-1 > 0 are generally typical for positive deviations of the copol ymer T-g from additivity, suggesting stiffening effects induced by att ractive interactions between the different repeating units of the copo lymer. Negative values of K indicate negative deviations due to repuls ions within the hetero-diads. The corresponding T-g vs composition cur ves may exhibit maxima or minima, respectively. The different influenc es on the copolymer T-g of hetero-triads, are characterized by values of K-2 different from zero. Depending on whether K-2 is smaller or lar ger than K-1 the results are either pronounced asymmetric or S-shaped T-g vs composition curves, suggesting that the influence on T-g of the hetero-triads is dependent on their composition, taking into account the hetero-triads may contain either two repeating units of the one or of the second component. Although fundamental relationships between t he fitting parameters and a molecular property are not yet available, it is shown, that in a first approximation the values of K-1 are relat ed to differences between the solubility parameters of the copolymer c omponents. The T-g vs composition behaviour is analysed for copolymers of acrylates and methacrylates among each other and with styrene, acr ylonitrile and vinyl chloride, respectively. (C) 1997 Elsevier Science Ltd.