Workflows and automated platforms for nanocrystal synthesis, purification, and characterization

Nanocrystals (NCs) are a class of materials with promising potential across a wide range of applications. Most NCs are synthesized using solution-phase methods that enable synthesis of nanostructures across a wide range of compositions at a large scale. These solution-phase methods all rely on specific processes of precursor reaction, formation of nuclei, and growth of nuclei to form NCs with desired shapes and sizes. Unfortunately, such mechanisms of NC formation are at present poorly understood, that in turn limits our ability to synthesize desired NCs at scale. Understanding the synthesis mechanisms for NC synthesis remains a major barrier to the widespread application of NCs. Conventional manual batch-reactor solution-based NC synthesis suffers from both low throughput and significant synthesis reproducibility that in turn leads to poor understanding of their formation mechanisms. This dissertation thus focuses on developing automated synthesis, purification, and characterizations approach to enable the rapid generation of reproducible datasets for understanding the mechanisms of NC nucleation and growth. In Chapter 2, an automated batch reactor that enables the reproducible synthesis of NCs is reported. This reactor platform was capable of reproducible hot injection synthesis of CdSe (< 0.2% variation across runs) and was used to collect a large dataset for the hot-injection synthesis of CdSe NCs. Using the dataset collected from this synthesis platform, machine learning models are used to predict the synthesis outcomes of this synthesis chemistry in Chapter 3. ML explanability tools were then used to understand the impact of individual synthesis parameters on synthesis outcomes and thus uncover their influence on CdSe nucleation and growth. The model explanability tool SHAP has demonstrated to not only show the relative importance of different process parameters such as temperatures and concentrations but also allows us to select features that can lead to more accurate models. In Chapter 4, an automated NC purification platform using size-exclusion chromatography (SEC) was used to enable one-step purification of NCs in nonpolar inorganic phases to produce high-purity NCs suitable for imaging and other structural characterization applications. This platform allows for rapid purification of crude QD synthesis mixtures (< 100s per sample) to generate QDs in solvents free of excess nonvolatile solvents and ligands (as seen from NMR analysis) and can be seamlessly integrated into existing synthesis and characterization platforms. In Chapter 5, building upon the platforms and workflows described in previous chapters, plans for future research directions are provided. These proposals seek to further develop capabilities for automated synthesis and characterization platforms by integrating existing (Chapter 2-4) and planned platforms with powerful structural and chemical characterization tools. Data from these integrated platforms will then be used by physics-based modeling approaches to uncover mechanisms of NC nucleation and growth. Some preliminary results on designing and testing these planned platforms will also be discussed here. Finally, in Chapter 6, a summary of insights gained in this dissertation as well as some perspectives about the outlook on capabilities in self-driving laboratories and NC synthesis will be discussed, with emphasis on improving characterization capabilities and modeling efforts. Overall, this dissertation focuses on developing automated capabilities in NC synthesis, purification, and characterization with a focus on synthesis reproducibility, structural characterization, and interpretable models to understand NC nucleation and growth.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Chemical Engineering

Grantor

University of Illinois Urbana-Champaign

Year dc:date

2025

Author dc:creator

• Xu, Rui Hua Jeff

Contributors dc:contributor

• Kenis, Paul J A
• Shim, Moonsub
• Peters, Baron G
• Schroeder, Charles M

• Automation
• Nanoparticles
• Machine Learning

Statement dc:rights

• Copyright 2025 Rui Hua Jeff Xu

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/129550

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F129550