田纳西大学Vogiatzis封面丨探索可用于优化催化剂的分子拓扑

01

Background

Heme and non-heme enzymes have the power to activate and dissociate strong C–H bonds for the formation of a plethora of products with biological relevance. The catalytically active site of these enzymes is an Fe(IV)–oxo. Since the trapping and characterization of the first non-heme Fe(IV)–oxo intermediate in the catalytic cycle of TauD (intermediate J) in 2003, many inorganic chemistry groups have synthesized non-heme Fe(IV)–oxo model complexes. These models have contributed on the elucidation of the electronic structure and spectroscopic properties of these novel intermediates, which ultimate target to provide mechanistic insights into their function in biology.

02

Research Summary

Computational studies provide a fundamental understanding of the electronic effects that control the reactivity of the Fe(IV)–oxo species, but also provide directions for the synthesis of the next generation of catalytic complexes and materials. Recently, graduate students Grier Jones, Dr. Brett Smith, and Dr. Justin Kirkland from the Vogiatzis lab present a novel computational procedure that combines quantum chemistry with machine learning for the accurate and reliable examination of thousands of coordination environments of Fe(IV)–oxo sites.

The new methodology utilizes a novel molecular fingerprinting method (Persistent Images) based on persistent homology, an applied branch of topology, that can encode the geometric and electronic structure together with molecular topology. Persistence Images can help the model to “learn” the topological features of a various ligand environments around the Fe(IV)–oxo site and how they affect the C–H activation step.

In this study, we have designed a database of 50 Fe(IV)–oxo species with varying coordination environments which are further functionalized for a total of approximately 181k structures. DFT calculations are then performed on a subset of the molecular database to determine spin states and C–H bond activation energies. The collected data are then curated based on a series of chemically informed criteria. To avoid performing 181k DFT calculations on the total chemical compound space, we developed machine learning models that utilize our novel molecular representation (persistence images). In particular, we have developed a novel similarity search algorithm, followed by training a regression model to predict C–H activation energies and a classification model to predict the spin states. The priority is to provide high-fidelity predictions for C–H activation barriers. For that purpose, we divided the full database into low- and high-fidelity structures, and we introduced a metric (δΔG^‡) which evaluates the effect of a specific ligand modification with respect to the parent, unsubstituted structure.

These new insights aid to the construction of a theoretical framework for the design of novel catalysts for energetically less demanding industrial processes, such as the oxyfunctionalization of natural gas.

03

Article Information

Data-driven ligand field exploration of Fe(IV)–oxo sites for C–H activation

Grier M. Jones, Brett A. Smith, Justin K. Kirkland and Konstantinos D. Vogiatzis

Inorg. Chem. Front., 2023, 10, 1062-1075

https://doi.org/10.1039/D2QI01961B

*文中图片皆来源上述文章

点击“阅读原文”直达上述文章

04

Corresponding Author

Prof. Konstantinos D. Vogiatzis

University of Tennessee

Konstantinos "Kostas" Vogiatzis is from Athens, Greece, and he is an Associate Professor at the University of Tennessee. His research focuses on the development of new computational methods based on electronic structure theory and machine learning for the theoretical examination of reactivity, catalysis, and separation processes.

Email: kvogiatz@utk.edu

Twitter: @VogLab

推荐阅读

微信改版，公众号文章不再以时间轴排列啦！

将Frontiers Journals设为星标⭐ 不错过更多精彩内容！

喜欢今天的内容？

👇    就来分享、点赞、在看三连吧    👇