Colloidal crystals via transport-informed processes and neural networks
Abstract
Colloidal crystals via transport-informed processes and neural networks Colloidal and particulate systems as a major class of soft and structured materials are ubiquitous in nature and industry alike: consumer products, multi-phasic oil and gas, drug delivery carriers, flowable electrodes for energy storage, and many more are just a few examples. The tunability of interparticle interactions between colloids at the microscale promises an extraordinary potential for development of novel materials based on specific applications and their corresponding functionalities; however, a major roadblock in the quest to rapid [particulate] soft materials with desired functionalities is missing the link between particle-level and the macroscopic properties. The scientific goal of this research is to provide systematic synthetic routes for targeted design of colloidal structures into novel layer-by-layer fabricated materials, by developing the process-structure-properties relationship in model attractive colloidal systems. Upon understanding of this process-structure-properties relationship in attractive colloids a series of novel processing protocols will be designed and tested in order to fabricate ordered arrays of colloidal crystals. Phase diagrams will be devised, predicting the resulting structures and properties based on particle characteristics, and details of processing conditions. In order to be able to thoroughly interrogate the process-structure-property relationship one requires to fully understand the multi-scale physics of the particle assembly in and far from equilibrium. For years and years, researchers have made efforts in order to come up with new pathways for targeted assembly of nanoparticles into structures, with one motivation in mind: these can be seen as the building blocks (atoms and molecules) for mesoscale synthesis for next generation materials with superior properties. However, development of these synthetic meso-structures is lagging behind new scientific understanding of advanced materials, due to lack of a fundamental understanding of process-structure-property relationship in nanoparticles with different characteristics. This is only possible through bridging the microscale dynamics/physics and macroscale properties, via mesoscale structures. The goal is to identify the interplay of complex many-body effects in play at all scale before establishing the processing and synthetic routes to targeted assembly. In addition to activities proposed in order to underpin the physics of self-assembly in attractive particulate systems, efforts will be made in order to employ physics-informed deep learning of the high fidelity simulation data, in order to accelerate the material design and discovery, by fully exploring the phase space of system and component variables with regards to order parameter and properties. Additionally, an understanding of the meta-constitutive equations for complex Thixotropic Elasto Visco Plastic (TEVP) fluids will be developed using the rheological measurements. This will be done by designing a deep neural network (DNN) consisting of two separate neural networks, one based on high-fidelity data from experiments and simulations, and the other one based on low-fidelity synthetic data using constitutive laws, and penalizing the learning algorithm based on the underlying physics. Results in this phase will be transformative since this will be the first use of machine learning algorithms in development of constitutive equations for complex fluids, and for predictive tools for particulate structures. This work seeks to establish the theoretical foundations needed for critical bottom-up assembly of particulate materials. A transport-directed 3D self-assembly will enable the design and synthesis of multi-component materials incorporating hierarchical constructs. These investigations will transform nano-particle self-assembly technologies.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 08, 2022
- Source ID
- W911NF2210172
Entities
People
- Safa Jamali
Organizations
- Army Contracting Command
- Northeastern University
- United States Army