Deviations from the mean field predictions for the phase behaviour of random copolymers melts

– We investigate the phase behaviour of random copolymers melts via large scale Monte Carlo simulations. We observe macrophase separation into A and B–rich phases as predicted by mean field theory only for systems with a very large correlation λ of blocks along the polymer chains, far away from the Lifshitz point. For smaller values of λ, we find that a locally segregated, disordered microemulsion–like structure gradually forms as the temperature decreases. As we increase the number of blocks in the polymers, the region of macrophase separation further shrinks. The results of our Monte Carlo simulation are in agreement with a Ginzburg criterium, which suggests that mean field theory becomes worse as the number of blocks in polymers increases. Introduction. – In polymeric systems one often encounters phase equilibria in mixtures with a large number of components. For instance, polydispersity (i.e. a continuous distribution of chain lengths) alters the phase equilibrium between a dense polymer phase and its vapour. Random AB copolymers constitute another example of such mixtures, each sequence of A’s and B’s represents a different component. The phase behaviour of random copolymers has attracted abiding interest, because these systems are produced commercially in large amounts and exhibit a rich phase behaviour. Mean field calculations [1, 2] predict a disordered phase, a macrophase separated phase and a disordered microphase separated phase, as a function of the incompatibility χ of the constituents A and B and the correlation λ along the polymers. We present the first Monte Carlo study of the phase behaviour of random copolymers melts and compare our results to the mean field predictions. For rather short chain lengths we observe all three phases predicted by the mean field theory, but the region of stability of the macrophase separation is much smaller than predicted. Increasing the chain length, we reduce the region of macrophase separation and increase the deviation from the mean field prediction. Our simulation results indicate that random copolymer melts are one of the very few dense polymer systems for which mean field predictions do not become accurate, even if the chains strongly interdigitate and interact with many neighbours. This finding is corroborated by the Ginzburg criterium [3].