Self-assembly is the autonomous organization of components into patterns or structures without human intervention, which is common throughout nature and technology. An important reason for human's interest in self-assembly is that self-assembly provides a practical approach to synthesizing useful structures at the nano/micro scale with exciting applications such as photonics, electronics, drug delivery, sensors and tissues etc. However, optimal design of a self-assembly process to obtain nano/micro-structured material with desired properties is challenging due to the stochastic nature and highly nonlinear behavior of self-assembly, which may lead to defects and kinetic trapping. In this thesis, new optimization and control methods are investigated for self-assembly to overcome...[ Read more ]

Self-assembly is the autonomous organization of components into patterns or structures without human intervention, which is common throughout nature and technology. An important reason for human's interest in self-assembly is that self-assembly provides a practical approach to synthesizing useful structures at the nano/micro scale with exciting applications such as photonics, electronics, drug delivery, sensors and tissues etc. However, optimal design of a self-assembly process to obtain nano/micro-structured material with desired properties is challenging due to the stochastic nature and highly nonlinear behavior of self-assembly, which may lead to defects and kinetic trapping. In this thesis, new optimization and control methods are investigated for self-assembly to overcome this challenge. Two application areas are investigated, which involve structural design of DNA tiles and directed self-assembly of colloidal particles.

First, a new optimization-based approach for structural design of DNA tiles is developed. The approach is based on potential-energy minimization of a desired structure. The significance of multiple local minima is demonstrated and various algorithms for optimization are investigated to identify the global minimum, which requires model reduction. The potential energy model is reformulated by including integer variables as degrees of freedom for structural design. By minimization of the new potential energy function, an optimal structural design is identified that is consistent with experimental results reported in literature when available.

Second, novel feedback control strategies to self-assemble colloidal particles in a microfluidic device are investigated experimentally. Different electrokinetic phenomena are exploited for directing self-assembly by manipulating field frequency and voltage. An automated two-step control strategy to self-assembly colloidal particles into non-periodic structures with single-particle resolution is presented. In the first step, the local particle density is controlled at a desired level, after which a second automated control step is applied to align the particles into a defect-free line with single-particle resolution. To address the non-linearity of the system, an alternative design of the feedback controller for particle density based on gain-scheduled PID control is also developed. The results show that a wide range of set points for particle density can be attained with this controller, which makes the defect-free assembly of a wide variety of non-periodic small-scale structures attainable.