NPNathaniel H. Park

Nathaniel H. Park

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Nathaniel H. Park has a total of 77 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Rapid Functionalization of Polycarbonates and Polyurethanes via Organocatalyzed SNAr

    by Eddy W. P. Tan, Yi Yan Yang, James L. Hedrick, Nathaniel H. Park

    Robust methodologies for polymer postsynthesis functionalization enable preparation of new high-performance material scaffolds with uniquely tailored properties. Here we demonstrate that polycarbonates containing pendant fluoroarene groups provide a convenient handle for incorporation of different functional thiols into the polymer backbone via organocatalytic nucleophilic aromatic substitution using silyl-protected thiols. This transformation proceeds under mild conditions and can be performed in a single pot immediately following polymerization or on purified substrates in a second step. In both cases, the transformation is typically complete within 5 min. Overall, the methodology presented herein enables a more convergent approach for development of functional high-performance materials.

  • AccFG: Accurate Functional Group Extraction and Molecular Structure Comparison

    by Xuan Liu, Sarathkrishna Swaminathan, Dmitry Zubarev, Brandi Ransom, Nathaniel Park, Kristin Schmidt, Huimin Zhao

    Functional groups are specific molecular patterns that influence a molecule's chemical and physical properties. Identifying them is crucial for understanding the structure-activity relationships. Currently, accurately extracting functional groups from complex molecular structures remains a challenge. In this work, we present AccFG, a tool designed for precise functional group extraction and molecular structure comparison. AccFG generates refined and structured functional group descriptors for molecules based on a predefined library, while lowering the barrier to encoding and extending the library with new functional group definitions. Additionally, AccFG's output includes identified functional groups and heterocycles with their corresponding mapped atom numbers, which enables molecular structure comparisons at the functional group level with alkane differences (differences in carbon backbone). We evaluated its performance on various data sets and examples with manual examination. The results demonstrate that AccFG can accurately extract functional groups from individual molecules and identify functional group-level differences between molecular pairs, effectively bridging the molecular modality and textual descriptors. We anticipate that AccFG will serve as a useful tool for processing functional groups, and its comprehensive, structured output enables fine-grained insights between molecular structures and properties. The code is available at https://github.com/xuanliugit/AccFG.


  • Open Access

    Agentic Mixture-of-Workflows for Multi-Modal Chemical Search

    by Tiffany J. Callahan, Nathaniel H. Park, Sara Capponi

    The vast and complex materials design space demands innovative strategies to integrate multidisciplinary scientific knowledge and optimize materials discovery. While large language models (LLMs) have demonstrated promising reasoning and automation capabilities across various domains, their application in materials science remains limited due to a lack of benchmarking standards and practical implementation frameworks. To address these challenges, we introduce Mixture-of-Workflows for Self-Corrective Retrieval-Augmented Generation (CRAG-MoW) - a novel paradigm that orchestrates multiple agentic workflows employing distinct CRAG strategies using open-source LLMs. Unlike prior approaches, CRAG-MoW synthesizes diverse outputs through an orchestration agent, enabling direct evaluation of multiple LLMs across the same problem domain. We benchmark CRAG-MoWs across small molecules, polymers, and chemical reactions, as well as multi-modal nuclear magnetic resonance (NMR) spectral retrieval. Our results demonstrate that CRAG-MoWs achieve performance comparable to GPT-4o while being preferred more frequently in comparative evaluations, highlighting the advantage of structured retrieval and multi-agent synthesis. By revealing performance variations across data types, CRAG-MoW provides a scalable, interpretable, and benchmark-driven approach to optimizing AI architectures for materials discovery. These insights are pivotal in addressing fundamental gaps in benchmarking LLMs and autonomous AI agents for scientific applications.


  • Open Access

    High-throughput computational analysis of kinetic barriers to ring-closing depolymerization for aliphatic polycarbonates

    by Brandi Ransom, Riccardo Bosio, Dmitry Zubarev, James Hedrick, Nathaniel Park

    The chemical reversion of polymers via ring-closing depolymerization (RCD) to their monomeric constituents is a highly promising avenue to enable end of life-cycle recycling and reuse. However, most reported systems using RCD revolve around bespoke monomer designs to facilitate facile depolymerization and there exists relatively few investigations into influence of functional groups on the ability of a particular monomer to cleanly undergo depolymerization. Here, we perform computational investigations into the energy barriers for RCD of 6-membered aliphatic carbonates in different solvents. The results demonstrate highlight clear trends observed in prior experimental studies, validating the utility of computational investigations towards understanding RCD. Experimental evaluation of thermal depolymerization of two monomers confirmed the ability of the computationally investigated monomers to cleanly undergo RCD. Overall, this work highlights the utility of high-throughput energy barrier computations to provide meaningful insight into broad reactivity trends that would be highly laborious to access experimentally.

  • Open Access

    Rapid Functionalization of Polycarbonates and Polyurethanes via Organocatalyzed SNAr

    by Eddy Wei Ping Tan, Yi Yan Yang, James Hedrick, Nathaniel Park

    Robust methodologies for polymer post-synthesis functionalization enable preparation of new high-performance material scaffolds with uniquely tailored properties. Here, we demonstrate that polycarbonates containing pendant fluoroarene groups provide a convenient handle for incorporation of different functional thiols into the polymer backbone via organocatalytic nucleophilic aromatic substitution using silyl-protected thiols. This transformation proceeds under mild conditions can be performed in a single-pot immediately following polymerization or on purified substrates in a second step. In both cases, the transformation is typically complete within 5 minutes. Overall, the methodology presented herein enables a more convergent approach for development of functional high-performance materials.

  • Cyclopropenimine‐Mediated CO2 Activation for the Synthesis of Polyurethanes and Small‐Molecule Carbonates and Carbamates

    by Dino Wu, Robert T. Martin, Jeanette Piña, Junho Kwon, Michael P. Crockett, Andy A. Thomas, Osvaldo Gutierrez, Nathaniel H. Park, James L. Hedrick, Luis M. Campos

    Carbon dioxide (CO2) is an abundant C1 feedstock with tremendous potential to produce versatile building blocks in synthetic applications. Given the adverse impact of CO2 on the atmosphere, it is of paramount importance to devise strategies for upcycling it into useful materials, such as polymers and fine chemicals. To activate such stable molecule, superbases offer viable modes of binding to CO2. In this study, a superbase cyclopropenimine derivative was found to exhibit exceptional proficiency in activating CO2 and mediating its polymerization at ambient temperature and pressure for the synthesis of polyurethanes. The versatility of this reaction can be extended to monofunctional amines and alcohols, yielding a variety of functional carbonates and carbamates.


  • Cyclopropenimine‐Mediated CO2 Activation for the Synthesis of Polyurethanes and Small‐Molecule Carbonates and Carbamates

    by Dino Wu, Robert T. Martin, Jeanette Piña, Junho Kwon, Michael P. Crockett, Andy A. Thomas, Osvaldo Gutierrez, Nathaniel H. Park, James L. Hedrick, Luis M. Campos

    Carbon dioxide (CO2) is an abundant C1 feedstock with tremendous potential to produce versatile building blocks in synthetic applications. Given the adverse impact of CO2 on the atmosphere, it is of paramount importance to devise strategies for upcycling it into useful materials, such as polymers and fine chemicals. To activate such stable molecule, superbases offer viable modes of binding to CO2. In this study, a superbase cyclopropenimine derivative was found to exhibit exceptional proficiency in activating CO2 and mediating its polymerization at ambient temperature and pressure for the synthesis of polyurethanes. The versatility of this reaction can be extended to monofunctional amines and alcohols, yielding a variety of functional carbonates and carbamates.


  • Ring-Opening Polymerization of Cyclic Esters and Carbonates with (Thio)urea/Cyclopropenimine Organocatalytic Systems

    by Romain Morodo, David M. Dumas, Jia Zhang, Kai H. Lui, Paul J. Hurst, Riccardo Bosio, Luis M. Campos, Nathaniel H. Park, Robert M. Waymouth, James L. Hedrick

    Organocatalyzed ring-opening polymerization is a powerful tool for the synthesis of a variety of functional, readily degradable polyesters and polycarbonates. We report the use of (thio)ureas in combination with cyclopropenimine bases as a unique catalyst for the polymerization of cyclic esters and carbonates with a large span of reactivities. Methodologies of exceptionally effective and selective cocatalyst combinations were devised to produce polyesters and polycarbonates with narrow dispersities (Đ = 1.01-1.10). Correlations of the pKa of the various ureas and cyclopropenimine bases revealed the critical importance of matching the pKa of the two cocatalysts to achieve the most efficient polymerization conditions. It was found that promoting strong H-bonding interactions with a noncompetitive organic solvent, such as CH2Cl2, enabled greatly increased polymerization rates. The stereoselective polymerization of rac-lactide afforded stereoblock poly(lactides) that crystallize as stereocomplexes, as confirmed by wide-angle X-ray scattering.


  • Open Access

    Author Correction: Artificial intelligence driven design of catalysts and materials for ring opening polymerization using a domain-specific language

    by Nathaniel H. Park, Matteo Manica, Jannis Born, James L. Hedrick, Tim Erdmann, Dmitry Yu. Zubarev, Nil Adell-Mill, Pedro L. Arrechea


  • Open Access

    Artificial intelligence driven design of catalysts and materials for ring opening polymerization using a domain-specific language

    by Nathaniel H. Park, Matteo Manica, Jannis Born, James L. Hedrick, Tim Erdmann, Dmitry Yu. Zubarev, Nil Adell-Mill, Pedro L. Arrechea

    Advances in machine learning (ML) and automated experimentation are poised to vastly accelerate research in polymer science. Data representation is a critical aspect for enabling ML integration in research workflows, yet many data models impose significant rigidity making it difficult to accommodate a broad array of experiment and data types found in polymer science. This inflexibility presents a significant barrier for researchers to leverage their historical data in ML development. Here we show that a domain specific language, termed Chemical Markdown Language (CMDL), provides flexible, extensible, and consistent representation of disparate experiment types and polymer structures. CMDL enables seamless use of historical experimental data to fine-tune regression transformer (RT) models for generative molecular design tasks. We demonstrate the utility of this approach through the generation and the experimental validation of catalysts and polymers in the context of ring-opening polymerization—although we provide examples of how CMDL can be more broadly applied to other polymer classes. Critically, we show how the CMDL tuned model preserves key functional groups within the polymer structure, allowing for experimental validation. These results reveal the versatility of CMDL and how it facilitates translation of historical data into meaningful predictive and generative models to produce experimentally actionable output.