<?xml version='1.0' encoding='UTF-8'?><?xml-stylesheet href='static/style.xsl' type='text/xsl'?><OAI-PMH xmlns="http://www.openarchives.org/OAI/2.0/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd"><responseDate>2026-11-22T09:51:46Z</responseDate><request verb="GetRecord" identifier="oai:ecommons.cornell.edu:1813/120809" metadataPrefix="dim">https://ecommons.cornell.edu/server/oai/request</request><GetRecord><record><header><identifier>oai:ecommons.cornell.edu:1813/120809</identifier><datestamp>2026-05-15T17:54:15Z</datestamp><setSpec>com_1813_35</setSpec><setSpec>col_1813_47</setSpec></header><metadata><dim:dim xmlns:dim="http://www.dspace.org/xmlns/dspace/dim" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:doc="http://www.lyncode.com/xoai" xsi:schemaLocation="http://www.dspace.org/xmlns/dspace/dim http://www.dspace.org/schema/dim.xsd">
   <dim:field mdschema="dc" element="contributor" qualifier="author">Chang, Ming-Chiang</dim:field>
   <dim:field mdschema="dc" element="contributor" qualifier="chair" lang="en_US">Thompson, Michael</dim:field>
   <dim:field mdschema="dc" element="contributor" qualifier="committeeMember" lang="en_US">Van Dover, Robert</dim:field>
   <dim:field mdschema="dc" element="contributor" qualifier="committeeMember" lang="en_US">Gomes, Carla</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="accessioned">2026-04-02T18:50:13Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="available">2026-04-02T18:50:13Z</dim:field>
   <dim:field mdschema="dc" element="date" qualifier="issued">2025-08</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="other">ProQuest Submission ID: 15047</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="other">ProQuest Publication ID: 32117293</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="uri">https://hdl.handle.net/1813/120809</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="doi">https://doi.org/10.7298/0848-d730</dim:field>
   <dim:field mdschema="dc" element="identifier" qualifier="bibid">17422613</dim:field>
   <dim:field mdschema="dc" element="description" lang="en_US">309 pages</dim:field>
   <dim:field mdschema="dc" element="description" qualifier="abstract" lang="en_US">Autonomous experimentation (AE) has the potential to accelerate materials discovery by integrating machine learning with automated synthesis and characterization, enabling rapid, closed-loop exploration of vast design spaces.A current challenge for AE is developing algorithms that enable the decision making agent to understand multimodal characterization data. Among common characterization methods, X-ray diffraction (XRD) provides dense crystallographic information which is critical for materials understanding and would be of immense value if it can be successfully integrated into AE workflows. Nevertheless, methods for the rapid and reliable analysis of XRD data are not yet fully established. In this thesis, two methods were developed to facilitate the integration of XRD data into autonomous experimentation. CrystalShift, designed for active learning use cases performs probabilistic phase labeling of single XRD pattern containing multiple phases by combining tree search and Bayesian probabilistic modeling.While CrystalShift focuses on labeling phases in a single XRD pattern, SPG-DRNet was developed to reason over larger phase spaces to construct physically viable phase processing diagrams. To process a full set of XRD patterns, SPG-DRNet utilizes a symmetry-constrained decoder to prevent hallucinating unphysical results, and uses the graph structure of the data set to reason with physical laws (e.g. Gibbs phase rule) and encourage relationships between neighboring samples. To provide AE agents with access to accurate thermal information, we also developed a temperature calibration workflow for lgLSA based on thermoreflectance measurements. This calibration allows the annealing temperature near the laser center to be quantified with an uncertainty of less than 3\%, providing the precision needed for reliable data-driven learning. Integrations of high-throughput XRD characterization with the automated phase labeling algorithms into an experimental workflow were demonstrated by investigating the Ta-Sn-Co-O system with lateral gradient laser spike annealing (lgLSA), synchrotron XRD mapping, and automated phase analysis using SPG-DRNet and CrystalShift.An extended phase-pure rutile stability region was discovered, featuring an unusually high substitutional alloy content. An unexpected cubic phase was also identified near the 1:1:1 cation composition. First-principles simulations supported the identification of this cubic phase as a pyrite structure and hinted at a possible stabilization pathway via pressure. Building on these capabilities, we extended the SARA framework, an AE method based on active learning with Bayesian optimization, to demonstrate a fully autonomous, targeted materials synthesis workflow. Applied to the Bi-Ti-O system, this approach successfully mapped the phase diagram in under two hours.Although the workflow is shown to be capable of operating independently, further performance improvements are shown to be possible through human-in-the-loop collaboration between AI and researchers. This thesis demonstrates how the combination of automated analysis, machine learning, and advanced synthesis can enable closed-loop experiments that accelerate material discovery and deepen understanding of metastable phase formation.</dim:field>
   <dim:field mdschema="dc" element="language" qualifier="iso">en</dim:field>
   <dim:field mdschema="dc" element="rights">Attribution 4.0 International</dim:field>
   <dim:field mdschema="dc" element="rights" qualifier="uri">https://creativecommons.org/licenses/by/4.0/</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Active Learning</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Autonomous Experimentation</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Deep Learning</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">Thin Films</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">X-ray Diffraction</dim:field>
   <dim:field mdschema="dc" element="subject" lang="en_US">XRD Phase Labeling</dim:field>
   <dim:field mdschema="dc" element="title" lang="en_US">Enabling Autonomous high-throughput XRD-based Phase Discovery in Thin film systems</dim:field>
   <dim:field mdschema="dc" element="type" lang="en_US">dissertation or thesis</dim:field>
   <dim:field mdschema="dc" element="format" qualifier="mimetype">application/pdf</dim:field>
   <dim:field mdschema="thesis" element="degree" qualifier="discipline">Materials Science and Engineering</dim:field>
   <dim:field mdschema="thesis" element="degree" qualifier="grantor">Cornell University</dim:field>
   <dim:field mdschema="thesis" element="degree" qualifier="level">Doctor of Philosophy</dim:field>
   <dim:field mdschema="thesis" element="degree" qualifier="name">Ph. D., Materials Science and Engineering</dim:field>
   <dim:field mdschema="dcterms" element="license">https://hdl.handle.net/1813/59810.2</dim:field>
   <dim:field mdschema="dspace" element="entity" qualifier="type">Publication</dim:field>
   <dim:field mdschema="cris" element="virtual" qualifier="collection" authority="https://ecommons.cornell.edu/handle/1813/47" confidence="600">Cornell Theses and Dissertations</dim:field>
   <dim:field mdschema="cris" element="virtual" qualifier="author">Chang, Ming-Chiang</dim:field>
   <dim:field mdschema="cris" element="virtualsource" qualifier="collection">5893a6ea-7af3-41d7-abc6-04bcd26ab5df</dim:field>
   <dim:field mdschema="others" element="access-status">open.access</dim:field>
   <dim:field mdschema="others" element="access-status">open.access</dim:field>
   <dim:field mdschema="cerif" element="openaire" authority="" confidence="-1">&lt;Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="3b2cbc96-e7ff-4688-9462-bb603ae17661">
	&lt;Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843&lt;/Type>
	&lt;Language>en&lt;/Language>
   	&lt;Title>Enabling Autonomous high-throughput XRD-based Phase Discovery in Thin film systems&lt;/Title>
   	&lt;PublishedIn>
    	&lt;Publication>
      	&lt;/Publication>
   	&lt;/PublishedIn>
   	&lt;PublicationDate>2025-08&lt;/PublicationDate>
   	&lt;DOI>https://doi.org/10.7298/0848-d730&lt;/DOI>
   	&lt;Authors>
      	&lt;Author>
        	&lt;DisplayName>Chang, Ming-Chiang&lt;/DisplayName>
         	&lt;Affiliation>
         		&lt;OrgUnit>
         		&lt;/OrgUnit>
         	&lt;/Affiliation>
      	&lt;/Author>
	&lt;/Authors>
   	&lt;Editors>
	&lt;/Editors>
    &lt;Publishers>
        &lt;Publisher>
            &lt;OrgUnit />
        &lt;/Publisher>
    &lt;/Publishers>
    &lt;License>https://creativecommons.org/licenses/by/4.0/&lt;/License>
    &lt;Keyword>Active Learning&lt;/Keyword>
    &lt;Keyword>Autonomous Experimentation&lt;/Keyword>
    &lt;Keyword>Deep Learning&lt;/Keyword>
    &lt;Keyword>Thin Films&lt;/Keyword>
    &lt;Keyword>X-ray Diffraction&lt;/Keyword>
    &lt;Keyword>XRD Phase Labeling&lt;/Keyword>
   	&lt;Abstract>Autonomous experimentation (AE) has the potential to accelerate materials discovery by integrating machine learning with automated synthesis and characterization, enabling rapid, closed-loop exploration of vast design spaces.A current challenge for AE is developing algorithms that enable the decision making agent to understand multimodal characterization data. Among common characterization methods, X-ray diffraction (XRD) provides dense crystallographic information which is critical for materials understanding and would be of immense value if it can be successfully integrated into AE workflows. Nevertheless, methods for the rapid and reliable analysis of XRD data are not yet fully established. In this thesis, two methods were developed to facilitate the integration of XRD data into autonomous experimentation. CrystalShift, designed for active learning use cases performs probabilistic phase labeling of single XRD pattern containing multiple phases by combining tree search and Bayesian probabilistic modeling.While CrystalShift focuses on labeling phases in a single XRD pattern, SPG-DRNet was developed to reason over larger phase spaces to construct physically viable phase processing diagrams. To process a full set of XRD patterns, SPG-DRNet utilizes a symmetry-constrained decoder to prevent hallucinating unphysical results, and uses the graph structure of the data set to reason with physical laws (e.g. Gibbs phase rule) and encourage relationships between neighboring samples. To provide AE agents with access to accurate thermal information, we also developed a temperature calibration workflow for lgLSA based on thermoreflectance measurements. This calibration allows the annealing temperature near the laser center to be quantified with an uncertainty of less than 3\%, providing the precision needed for reliable data-driven learning. Integrations of high-throughput XRD characterization with the automated phase labeling algorithms into an experimental workflow were demonstrated by investigating the Ta-Sn-Co-O system with lateral gradient laser spike annealing (lgLSA), synchrotron XRD mapping, and automated phase analysis using SPG-DRNet and CrystalShift.An extended phase-pure rutile stability region was discovered, featuring an unusually high substitutional alloy content. An unexpected cubic phase was also identified near the 1:1:1 cation composition. First-principles simulations supported the identification of this cubic phase as a pyrite structure and hinted at a possible stabilization pathway via pressure. Building on these capabilities, we extended the SARA framework, an AE method based on active learning with Bayesian optimization, to demonstrate a fully autonomous, targeted materials synthesis workflow. Applied to the Bi-Ti-O system, this approach successfully mapped the phase diagram in under two hours.Although the workflow is shown to be capable of operating independently, further performance improvements are shown to be possible through human-in-the-loop collaboration between AI and researchers. This thesis demonstrates how the combination of automated analysis, machine learning, and advanced synthesis can enable closed-loop experiments that accelerate material discovery and deepen understanding of metastable phase formation.&lt;/Abstract>
	&lt;Access xmlns="http://purl.org/coar/access_right" 
    >
    &lt;/Access>
&lt;/Publication>
</dim:field>
</dim:dim>
</metadata></record></GetRecord></OAI-PMH>