Nematic Order Drives Phase Separation in Polydisperse Liquid Crystalline Polymers F. Elias, S. M. Clarke, R. Peck, and E. M. Terentjev* Cavendish Laboratory, University of Cambridge, Madingley Road, Cambridge CB3 0HE, U.K. Received October 4, 1999; Revised Manuscript Received January 11, 2000 ABSTRACT: In a polydisperse thermotropic main-chain liquid crystalline polymer, we observe the process of thermally induced nematic-nematic phase separation between short and long polymer chains. We study the dynamical features of this system, in particular the evolution of Schlieren textures formed by disclinations surrounding areas of relatively uniform director. We analyze the dependence of domain size on temperature and the nematic order parameter, and the evolution of textures with the waiting time at a temperature well below the nematic transition T ni. We find that before phase separation the coarsening proceeds toward the uniform state, with characteristic size of the texture increasing as t 1/4 . When the system is phase-separated, the texture in the regions with long chains is frozen at an equilibrium value , a reversible function of temperature, while in the short-chain regions the coarsening accelerates. This behavior is interpreted in terms of a miscibility gap that is proportional to the degree of nematic order, which is different for the different lengths of the nematic polymers. 1. Introduction Schlieren textures in nematic liquid crystals 1 are due to the spatial variations of the director field n(r). They are visualized through the optical contrast between the birefringent regions with different orientations of this axis. A typical Schlieren texture has a variety of topological defects of the orientational order (disclina- tions) that match the director field between the domains with different orientation of n. The average distance between disclinations represents the size of such cor- related regions, within which the nematic director is more or less aligned, and is a characteristic length scale of the texture. Although there is no abrupt boundary separating such regions with different average director alignments, they are often referred to as domains, the average distance between disclinations being called the domain size . The presence of a texture in the director field is energetically unfavorable and is penalized in the ne- matic by the Frank elastic energy density 1 / 2 K(n) 2 . 1 As a result, if unconstrained, the textures always tend to relax toward the equilibrium uniform director align- ment. Such a process is often viewed as the growth of correlated domains, or of the average distance between disclinations: the coarsening of the characteristic length scale (t). The evolution of nematic textures has been the subject of extensive research and interesting analo- gies. 2 Theoretical, as well as experimental, results of this research describe the interaction of disclinations of opposite sign, annihilating each other as the system approaches its lowest elastic energy equilibrium. The argument for such a law is simple 3,4 and based on representing a disclination as a line under tension τ. Assuming the texture is characterized by a single length scale , the disclination line density should then scale as F -2 . Balancing the tension force per unit length, τ/, against the viscous friction force, ηυ, one finds the characteristic velocity of disclination movement υ ) τ/η. The rate of energy loss per unit volume is then W ˙ ) Fυτ/and is equal to the reduction in elastic tension energy density d/dt(Fτ). One thus obtains the line density F (η/τ)t -1 , or the characteristic size of coarsen- ing domain texture t 1/2 . Evidently, the dynamics of coarsening is determined by the laws of friction applied, for instance, to moving disclinations. Hence, the high viscosity (and, perhaps even more relevantsviscoelasticity) of liquid crystalline polymers (LCP) should make the coarsening of Schlieren textures much slower than that in low-molecular weight nematics. Indeed, this has been seen in many materials, particularly side-chain LCPs (see, for instance, ref 5), where the reorientation of mesogenic groups is re- stricted by connection to the polymer backbone. In main- chain liquid crystalline polymers (MCLCP), rodlike mesogenic groups separated by flexible spacers form the chains. In this case, the nematic director rotation is determined by the dynamics of polymer backbone itself. 6,7 The Schlieren textures in MCLCP and, in particular, the process of texture coarsening have been extensively studied over recent years. 8-12 It appears that the coarsening occurs via the annihilation of disclina- tions of opposite sign, as in other nematic systems. It has been reported that the average domain size scales as t 0.35 , 10,11 noticeably slower than in a liquid nematic discussed above. However, in all MCLCPs studied, the textures seem to evolve toward a stable pattern, which is far from a uniform director that may have been expected to be the equilibrium. In other words, the characteristic domain size initially increases, as the coarsening dynamics would require, but then it appears to saturate at a constant value, of order of several micrometers, 12 In a previous study, 13 we have shown that in a MCLCP of high molecular weight, this final texture is an equilibrium state of the system, the domain size being a reversible function of the temper- ature. Such behavior is similar to the case of nematic elastomers, where the chains are permanently cross- linked in the network (see the review in ref 14 for details). There, the polydomain texture has been shown to represent a thermodynamic equilibrium in which the demand to minimize the Frank elastic energy is bal- anced by the quenched random-anisotropy effect of network cross-links. 15 It has been shown 16 that the characteristic domain size is a reversible function of temperature, increasing toward the isotropic phase, T f T ni . In the case of non-cross-linked MCLCP, we have attributed the existence of equilibrium textures to the 2060 Macromolecules 2000, 33, 2060-2068 10.1021/ma9916786 CCC: $19.00 © 2000 American Chemical Society Published on Web 03/02/2000