PhD defence: Shape, Frustration, and Elasticity in Colloidal Liquid Crystals - A Particle Based Simulation Study

PLEASE NOTE: The candidate gives a layman's talk, therefore the livestream will start fifteen minutes earlier.
 

Matter is commonly classified by the degree of microscopic order: crystals possess positional order, whereas liquids are disordered. Liquid crystals, first observed by Reinitzer in 1888, occupy an intermediate state: they flow like liquids yet retain internal orientational (and sometimes partial positional) order, giving rise to striking optical and mechanical properties. In lyotropic, colloidal liquid crystals this ordering is governed largely by particle geometry, making structure formation primarily a problem of shape.

This PhD thesis investigates how particle shape selects liquid-crystalline phases by studying predominantly hard-particle models and using extensive Molecular Dynamics and Monte Carlo simulations, supported where relevant by experiment. The central design principle is that particles whose geometry resembles one of the fundamental nematic deformation modes; twist, bend, splay, and saddle-splay, tend to promote that same deformation at the macroscopic level, as captured by the Oseen–Frank elastic framework.

The thesis shows that hard banana-shaped particles can stabilise highly complex modulated nematic structures, including skyrmion-like vortices that either form three-dimensional filament networks or, under strong confinement, assemble into hexagonal skyrmion lattices. These predictions are corroborated by the first experimental observation of skyrmions in colloidal mixtures of bent and straight silica spherocylinders, where tuning bend angle and composition enhances the stability of modulated phases and reveals previously unobserved ordering. The work further demonstrates that saddle-splay frustration can be realised using distorted tetrahedral (“pringle-symmetry”) particles, which spontaneously generate twisted nematic phases from achiral building blocks and, in smectic regimes, produce curved-layer morphologies ranging from lamellae and hexagonal domains to gyroid-like phases, showing that geometry alone can stabilise structures often associated with complex chemistry.

Beyond hard-particle self-assembly, the thesis develops an efficient machine-learning approach grounded in a rigorous tensorial expansion to model general anisotropic interactions, enabling accurate coarse-grained simulations of complex materials. It also illustrates how external fields, gravity, and confinement can drive porous network formation even for simple rods through induced dipolar interactions. Overall, the dissertation establishes a unified, shape-based framework for understanding and engineering frustration-driven liquid-crystalline order, and outlines open theoretical and computational challenges for extending this approach to broader, thermotropic and technologically relevant systems.