(305e) Image-Based State Representation of 2D Colloidal Self-Assembly Systems.

Colloidal self-assembly refers to the spontaneous organization of nano/micro-particles into secondary structures in a solution. The property of the assembled structure depends on the physicochemical properties of the particles, as well as the geometrical and morphological properties of the assembly. Therefore, it is of paramount importance to quantify the geometrical features of the colloidal assembly for a controlled production with the desired assembly properties. However, such a characterization of the system state is challenging, due to the high dimensionality and complex dynamics of the assembly. Here, we present an image-based framework to classify the state of the colloidal self-assembly system. The framework leverages deep learning algorithms with unsupervised learning for state classification and a supervised learning-based convolutional neural network for state prediction. The neural network models are trained and tested using synthetic data from a Brownian dynamics simulation, which was validated against experimental measurements. Based on simulation results, our model outperforms the commonly used order parameters in identifying the states of the colloidal system, specifically in classifying a defective state from an ordered state. Given the image-based nature of the proposed approach, we anticipate it to be applicable to other complex systems where image acquisition is available.