MACHINE LEARNING ASSISTED STRATEGIC SYNTHESIS OF TISSUE MIMETIC ELASTOMERS

Over the course of evolution, biological creatures in nature have developed various elegant mechanisms to defend themselves. Particularly, soft biological tissues not only serve as cushions but at the same time, also prevent tearing. Meanwhile, some tissues, such as the skin of chameleons, can also display adaptive coloration which protects them from predators and helps them attract spouses. Inspired by the multifunctionality of biological tissues, this study focused on developing materials that possess a combination of these unique properties. To characterize the nonlinear elasticity of tissues and synthetic materials that mimic this property, we used firmness β and Young’s modulus E_0. To unravel the origin of mechanical properties of tissues, we studied the stress-strain curves of previously measured tissues from literatures. We demonstrated that the mechanical properties of tissues were tied to their functions and structural organization of collagens. To target the nonlinear elasticity synthetically, we used linear-bottlebrush-linear (LBL) triblock copolymers that micro-phase separate into physical networks, which we named plastomers. The triblock was produced by a two-step atomic transfer radical polymerization (ATRP) synthesis: the bottlebrush macroinitiator was synthesized by grafting-through polymerization followed by linear chain extension from both ends of the macroinitiator. The synthetic challenges and synthetic outcomes on the effect of mechanical properties of plastomers were investigated. Rigorous kinetic studies were performed to optimize the synthetic conditions for producing bottlebrush macroinitiator with high chain end fidelity. Next, we investigated in the control of mechanical properties by varying architectural parameters as well as mixing experiments. We showed that there is still a gap between synthetic plastomers and biological tissues. In particular, we lacked synthetic materials that possessed high firmness (β > 0.8) and high modulus (E_0 > 105 Pa). To bridge this gap, we needed to target plastomers with specific firmness and modulus. Therefore, we developed statistical and machine learning models that predicted the mechanical properties of triblocks based on chemical and architectural parameters. Finally, we investigated in incorporating structural coloration into plastomers. We studied factors, such as architectural parameters of the plastomers and swelling that controlled the reflected color of the plastomers. Specifically, we utilized ultraviolet-visible (UV-VS) spectroscopy and small angle X-ray scattering (SAXS) to demonstrate the effect of these factors on reflected wavelength and periodicity of the plastomers.Doctor of Philosoph