Computational tools for materials design: applications in dynamic covalent chemistry, polymers, and electrochemical systems

Cencer, Morgan

Permalink

https://hdl.handle.net/2142/115661 Copy

Description

Title

Computational tools for materials design: applications in dynamic covalent chemistry, polymers, and electrochemical systems

Author(s)

Cencer, Morgan

Issue Date

2022-03-02

Director of Research (if dissertation) or Advisor (if thesis)

Moore, Jeffrey S

Doctoral Committee Chair(s)

Moore, Jeffrey S

Committee Member(s)

• Tajkhorshid, Emad
• Zimmerman, Steven C
• Pogorelov, Taras P

Department of Study

Chemistry

Discipline

Chemistry

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• computational chemistry
• materials design
• rule-based models
• DFT
• FROMP

Abstract

Modern materials chemistry covers a huge chemical space, from hydrogels to metal organic frameworks. The development of all this array of materials has been based on a variety of experimental, analytical, and computational techniques. Computational techniques have become increasingly vital to both material design and data analysis as technological advances in computation and automation have made scientific computation easier and data sets larger. This thesis discusses a variety of simulation, modeling, and data analysis techniques as applied to the fields of dynamic covalent chemistry, polymers, and electrochemistry. The first chapter provides a broad overview of computational techniques. The second chapter discusses the use of rule-based models to understand error correction and assembly in dynamic covalent chemistry systems. Chapter three focuses on how rule-based models can aid in understanding depolymerization mechanisms. In chapter four, electrochemical processes are modeled using density functional theory. Chapter five introduces the use of machine learning and density functional theory for monomer discovery. The final chapter provides a summary of these projects and the outlook for computational techniques in materials design. Rule-based models are ideal for dynamic covalent chemistry because rational design of large scale 3D molecular structures (covalent organic frameworks, molecular cages, etc.) is currently limited by our understanding of how the individual molecular constituents assemble into the larger structure. The most efficient routes to large scale 3D molecular structures depend on reversible interactions (either covalent or non-covalent), which can disconnect misplaced bonds (error correction), and reach the thermodynamic minimum product. However, error correction can be stymied by kinetic traps, which are persistent reaction intermediates. Kinetic traps are most common with multivalent precursors and slow exchange chemistries. With simple systems or very fast chemistries thermodynamic modeling is often sufficient to predict reaction outcomes. As complexity increases or exchange speeds slow, kinetic modeling is also necessary to understand the time course of the reaction and predict kinetic traps. Rule-based models can successfully capture the critical aspects of a dynamic reaction and predict error correction time. These models can give guidance to the planning of a dynamic covalent synthesis by predicting time to maximum yield of a desired product based on rate coefficients and valency. They can also capture details of the reaction network and provide an estimate of reaction kinetics that are challenging to directly measure. This type of modeling approach is particularly well suited to dynamic chemistries as a few simple rules can capture all the reactivity of the system, and the reaction networks are generated algorithmically. Another application for rule-based models is in understanding depolymerization mechanisms. Polymers that can degrade upon exposure to a specific trigger are of great interest as they could reduce plastic waste, make up transient electronics, act as biodegradeable medical implants, or as environmental sensors. There are a wide variety of chemistries used for triggered degradation, including those that are irreversible and those that fully depolymerize back down to monomer. As any triggered degradation has the trigger reaction and then an unzipping reaction, it can be quite challenging to experimentally measure the kinetics of each reaction. We apply rule-based models to determine the kinetic parameters that correctly model experimental data, to distinguish between mechanisms, and to probe the effect of mechanism, rate, and molecular weight on overall degradation behavior. We also develop new techniques to properly model dispersity based on experimental size-exclusion chromatography data and to directly calculate molecular weight within the model. The models here are directly tied to specific chemical systems, but they are also solid backbones for anyone desiring to build rule-based models for their own systems. Electrochemistry is vital to developing a green energy economy, whether it is through converting carbon dioxide in the atmosphere to value added chemicals, or through creating innovative new batteries. Optimizing reaction conditions or designing battery materials requires detailed understanding of the molecular and atomic level interactions between electrolytes, solvents, and reaction intermediates. Chapter four covers three approaches for using density functional theory (DFT) to gain insight into the behavior of electrochemical systems. First, we use dynamic DFT to probe interactions between carbon dioxide reduction reaction intermediates and surrounding solvent and electrolyte molecules. We show how different cations and solvents affect the stability of the [CO2]- radical anion by examining its angle and distance to cation in dynamic simulations. We identify that the strength of [CO2]- interactions can be tailored through choosing an appropriate cation and solvent combination. Second, we use static DFT to probe the relationship between cluster complexation energy and electrochemical behavior. We show that complexation energy can be easily predicted using machine learning and a database of over 300 complexes. Third, we develop a database of reduction potentials for divalent metal complexes of interest for designing beyond lithium batteries. We calculated the 1e- and 2e- reduction potentials for 78 different complexes in 22 different solvents, analyze the data trends, and identify which complexes are unstable upon reduction. These three studies use atomistic modeling to provide guidance on which systems are suitable and stable for specific applications, and worth further experimental study. Thermoset polymers and composites are widely used in airplanes, cars, and structural components due to their strength, stiffness, and low density. However, traditional thermosets require large amounts of energy to cure. Frontal ring-opening metathesis polymerization (FROMP) only requires a small initial thermal or photo stimulus which triggers a self-propagating exothermic reaction that fully cures the monomer to polymer. Currently, there are only a small number of monomers that can sustain FROMP. We use machine learning to identify the key aspects of a monomer that contribute to its FROMP behavior. We developed a large pool of over eleven million candidate monomers, and created a representative sample of those candidates. We did density functional theory calculations and scored for synthetic accessibility scores for all the representative sample. We discuss the active learning model that can be built from the known monomers and the candidate monomers to identify new FROMP monomers. The computational methods discussed here are quite varied, but they share the basic hypothesis that modeling system and applying systematic data analysis can provide vital information to understand complex chemical systems. As shown in this dissertation, the future of chemical discovery relies on the interweaving of chemical knowledge, experimental results, theory, simulation, and data science.

Graduation Semester

2022-05

Type of Resource

Thesis

Copyright and License Information

Copyright 2022 Morgan Cencer

Owning Collections