Glass Formation in the Gay - Berne nematic liquid crystal

We present the results of molecular dynamics simulations of the Gay-Berne model of liquid crystals, supercooled from the nematic phase. We find a glass transition to a metastable phase with nematic order and frozen translational and orientational degrees of freedom. For fast quench rates the local structure is nematic-like, while for slower quench rates smectic order is present as well. 61.30.-v, 61.30Cz, 64.70Pf Typeset using REVTEX 1 The study of nematic glasses [1] is richer than the corresponding study of isotropic glasses due to the presence of orientational degrees of freedom in the former systems. Assuming that we supercool a liquid crystal starting from its nematic phase (rather than the isotropic phase) we expect a nematic glass to have long-range orientational order like an equilibrated nematic as well as frozen density and director fluctuations. Orientationally this glassy phase is similar to a mixed magnetic phase where both ferromagnetic and vector spin glass order coexist [2]. However, unlike a spin glass the nematic glass is a nonequilibrium phase of matter and the molecular translational degrees of freedom freeze as well at the glass transition. In this Letter we study the formation of a nematic glass using molecular dynamics (MD). We model the liquid crystal using the Gay-Berne (GB) potential [3] which is an anisotropic Lennard-Jones potential. Previous molecular dynamics studies have indicated that the GayBerne potential exhibits a rich phase diagram [4–6] and the principal dynamical features [7–10] of real liquid crystals. The GB potential is given by, U(û1, û2, r) = 4ε(û1, û2, r)× [{ σo r − σ(û1, û2, r) + σo } 12 − { σo r − σ(û1, û2, r) + σo } 6 ] (1) where û1, û2 are unit vectors giving the orientations of the two molecules separated by the position vector r. The parameters ε(û1, û2, r) and σ(û1, û2, r) are orientation dependent and give the well depth and the intermolecular separation where U = 0 respectively. The well depth is written as, ε(û1, û2, r) = εoε (û1, û2)ε (û1, û2, r) (2) where ε(û1, û2) = (1− χ 2(û1 · û2) ) (3) and ε(û1, û2, r̂) = 1− χ 2 { (r̂ · u1 + r̂ · u2) 2 1 + χ(u1 · u2) + (r̂ · u1 − r̂ · u2) 2 1− χ(u1 · u2) }