Expansive Quantum Mechanical Exploration of Chemical Reaction Paths

Baiardi, Alberto; Grimmel, Stephanie A.; Steiner, Miguel; Türtscher, Paul L.; Unsleber, Jan P.; Weymuth, Thomas; Reiher, Markus

doi:10.1021/acs.accounts.1c00472

Cited by 25 publications

(17 citation statements)

References 75 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Combined with active learning − and global optimization algorithms, − catalysts with optimal catalytic properties can be quickly identified for a given mechanism. Because it is often difficult to experimentally characterize all mechanistically relevant intermediates due to their transient nature, computational approaches that explore reaction mechanisms are also desired. , In this case, ML combined with automated VHTS workflows can accelerate the exploration of potential reactive intermediates and reaction pathways. − …”

Section: Introductionmentioning

confidence: 99%

Machine Learning Models Predict Calculation Outcomes with the Transferability Necessary for Computational Catalysis

Duan

Nandy

Adamji

et al. 2022

J. Chem. Theory Comput.

View full text Add to dashboard Cite

Virtual high-throughput screening (VHTS) and machine learning (ML) have greatly accelerated the design of single-site transition-metal catalysts. VHTS of catalysts, however, is often accompanied with a high calculation failure rate and wasted computational resources due to the difficulty of simultaneously converging all mechanistically relevant reactive intermediates to expected geometries and electronic states. We demonstrate a dynamic classifier approach, i.e., a convolutional neural network that monitors geometry optimizations on the fly, and exploit its good performance and transferability in identifying geometry optimization failures for catalyst design. We show that the dynamic classifier performs well on all reactive intermediates in the representative catalytic cycle of the radical rebound mechanism for the conversion of methane to methanol despite being trained on only one reactive intermediate. The dynamic classifier also generalizes to chemically distinct intermediates and metal centers absent from the training data without loss of accuracy or model confidence. We rationalize this superior model transferability as arising from the use of electronic structure and geometric information generated on-the-fly from density functional theory calculations and the convolutional layer in the dynamic classifier. When used in combination with uncertainty quantification, the dynamic classifier saves more than half of the computational resources that would have been wasted on unsuccessful calculations for all reactive intermediates being considered.

show abstract

Section: Introductionmentioning

confidence: 99%

Machine Learning Models Predict Calculation Outcomes with the Transferability Necessary for Computational Catalysis

Duan

Nandy

Adamji

et al. 2022

J. Chem. Theory Comput.

View full text Add to dashboard Cite

show abstract

“…Compared with a series of static explorations on the potential energy surface (PES), including the latest development of first-principles heuristics, 9 enhanced chemical reactions 10 and interactive quantum mechanics, 11 these methods have better results for local or specific chemical reaction processes. 12 Whereas an important concept of AMK is the use of molecular dynamics (MD) or other sampling methods to obtain enough reasonable initial guesses of TS configurations, and then screening these configurations from low-level to high-level calculations, to ultimately achieve the full automation of TS searches. Although the AMK method has been extensively applied in many fields, such as van der Waals diversity calculation and catalysis, 2,13,14 it remains characterized by the following shortcomings: i.…”

Section: Introductionmentioning

confidence: 99%

“…AMK adopts a large number of shell scripts for data assimilation and the connection of different calculation levels. Compared with a series of static explorations on the potential energy surface (PES), including the latest development of first‐principles heuristics, 9 enhanced chemical reactions 10 and interactive quantum mechanics, 11 these methods have better results for local or specific chemical reaction processes 12 . Whereas an important concept of AMK is the use of molecular dynamics (MD) or other sampling methods to obtain enough reasonable initial guesses of TS configurations, and then screening these configurations from low‐level to high‐level calculations, to ultimately achieve the full automation of TS searches.…”

Section: Introductionmentioning

confidence: 99%

Automated reaction mechanisms and kinetics with the nudged elastic band method‐based AMK_Mountain and its description of the preliminary alkaline hydrolysis of nitrocellulose monomer

Zhang

Liang

et al. 2022

J Comput Chem

View full text Add to dashboard Cite

To improve the transition state (TS) search capability in complex chemical environments, AMK_Mountain is constructed based on the automated reaction mechanisms and kinetics (AutoMeKin) package. AMK_Mountain does not distinguish the reaction type of the TSs, which is beneficial to obtaining a more comprehensive reaction mechanism. In this study, the first step of the alkaline hydrolysis process of nitrocellulose monomer was adopted as the research object, and 730 possible initial configurations are constructed and 22 TSs pass high‐level calculations. Energy difference and interaction region indicator reveal that the first step of alkaline hydrolysis is mainly the combination of nitrogen‐containing functional groups at the positions α and β with hydroxide anions, followed by the formation of nitric acid and the further loss of protons to form nitrate. Overall, in combination with GFN2‐xTB and ORCA, the AMK_Mountain technique provides a reliable method for the location of the TSs in complex environments.

show abstract

“…Within the past decade, approaches have been developed for the automated exploration of reactive chemical systems. [29][30][31][32][33][34][35] A software for this purpose is SCINE Chemoton, 36,37 which next to fully fledged exploration protocols, provides a range of so-called double-sided transition state finding algorithms, i.e., algorithms which search for a reaction given both reactant and product structures. Another example is autodE, 38 which requires only the SMILES strings of all reactant and product structures.…”

Section: Introductionmentioning

confidence: 99%

Quantum Chemical Data Generation as Fill-In for Reliability Enhancement of Machine-Learning Reaction and Retrosynthesis Planning

Toniato¹,

Unsleber

Vaucher³

et al. 2022

Preprint

View full text Add to dashboard Cite

Data-driven synthesis planning has seen remarkable successes in recent years by virtue of modern approaches of artificial intelligence that efficiently exploit vast databases with experimental data on chemical reactions. However, this success story is intimately connected to the availability of existing experimental data. It may well occur in retrosynthetic and synthesis design tasks that predictions in individual steps of a reaction cascade are affected by large uncertainties. In such cases, it will, in general, not be easily possible to provide missing data from autonomously conducted experiments on demand. However, first-principles calculations can, in principle, provide missing data to enhance the confidence of an individual prediction or for model retraining. Here, we demonstrate the feasibility of such an ansatz and examine resource requirements for conducting autonomous first-principles calculations on demand.

show abstract

Expansive Quantum Mechanical Exploration of Chemical Reaction Paths

Cited by 25 publications

References 75 publications

Machine Learning Models Predict Calculation Outcomes with the Transferability Necessary for Computational Catalysis

Machine Learning Models Predict Calculation Outcomes with the Transferability Necessary for Computational Catalysis

Automated reaction mechanisms and kinetics with the nudged elastic band method‐based AMK_Mountain and its description of the preliminary alkaline hydrolysis of nitrocellulose monomer

Quantum Chemical Data Generation as Fill-In for Reliability Enhancement of Machine-Learning Reaction and Retrosynthesis Planning

Contact Info

Product

Resources

About