Synergistic Effects of Transition Metals and Coordination Environments on Nitrate Reduction in Tetraethynylporphyrin Analyzed by Machine Learning and Verified Using First-Principles Calculations

Chemical heuristics have been fundamental to the advancement of chemistry and materials science. These heuristics are typically established by scientists using knowledge and creativity to extract patterns from limited datasets. Machine learning offers opportunities to perfect this approach using computers and larger datasets. Here, we discuss the relationships between traditional heuristics and machine learning approaches. We show how traditional rules can be challenged by large-scale statistical assessment and how traditional concepts commonly used as features are feeding the machine learning techniques. We stress the waste involved in relearning chemical rules and the challenges in terms of data size requirements for purely data-driven approaches. Our view is that heuristic and machine learning approaches are at their best when they work together.

Learning What Makes Catalysts Good

Whither Second-Sphere Coordination?

Electrocatalytic sulfur reduction reaction (SRR) is emerging as an effective strategy to combat the polysulfide shuttling effect, which remains a critical factor impeding the practical application of the Li-S battery. Single-atom catalyst (SAC), one of the most studied catalytic materials, has shown considerable potential in addressing the polysulfide shuttling effect in a Li-S battery. However, the role played by transition metal vs coordination mode in electrocatalytic SRR is trial-and-error, and the general understanding that guides the synthesis of the specific SAC with desired property remains elusive. Herein, we use first-principles calculations and machine learning to screen a comprehensive data set of graphene-based SACs with different transition metals, heteroatom doping, and coordination modes. The results reveal that the type of transition metal plays the decisive role in SAC for electrocatalytic SRR, rather than the coordination mode. Specifically, the 3d transition metals exhibit admirable electrocatalytic SRR activity for all of the coordination modes. Compared with the reported N3C1 and N4 coordinated graphene-based SACs covering 3d, 4d, and 5d transition metals, the proposed para-MnO2C2 and para-FeN2C2 possess significant advantages on the electrocatalytic SRR, including a considerably low overpotential down to 1 mV and reduced Li2S decomposition energy barrier, both suggesting an accelerated conversion process among the polysulfides. This study may clarify some understanding of the role played by transition metal vs coordination mode for SAC materials with specific structure and desired catalytic properties toward electrocatalytic SRR and beyond.

As promising next-generation candidates for applications in aero-engines, L12-strengthened cobalt (Co)-based superalloys have attracted extensive attention. However, the L12 strengthening phase in first-generation Co-Al-W-based superalloys is metastable and both its solvus temperature and mechanical properties still need to be improved. Therefore, it is necessary to discover new L12-strengthened Co-based superalloy systems with a stable L12 phase by exploring the effect of alloying elements on its stability. Traditional first-principles calculations are capable of providing the crystal structure and mechanical properties of the L12 phase doped by transition metals but suffer from low efficiency and relatively high computational costs. The present study combines machine learning (ML) with first-principles calculations to accelerate crystal structure and mechanical property predictions, with the latter providing both the training and validation datasets. Three ML models are established and trained for predicting the occupancy of alloying elements in the supercell and the stability and the mechanical properties of the L12 phase. The ML predictions are evaluated using first-principles calculations and the accompanying data are used to further refine the ML models. Our ML-accelerated first-principles calculation approach offers more efficient predictions of the crystal structure and mechanical properties for Co-V-Ta- and Co-Al-V-based systems than the traditional counterpart. This approach is applicable to expediting crystal structure and mechanical property calculations and thus the design and discovery of other advanced materials beyond Co-based superalloys.

Machine learning for predicting thermal transport properties of solids

Batch active learning for accelerating the development of interatomic potentials

The ultraviolet luminescence of β-galliumsesquioxide

Perspective on machine learning in energy material discovery

This study presents a novel approach which combines first-principles calculations with machine learning techniques to predict the 4f-5d transition energy of Ce3+ ions in garnet-type oxides. This is important in various technological applications, e.g., light-emitting materials, and solid-state lighting devices. However, it is difficult and time-consuming to accurately determine this energy using conventional methods. A linear regression model for the 4f-5d transition energy of Ce3+ in garnet-type oxides was created by machine learning, using electronic and structural parameters as attributes. The electronic parameters were obtained through first-principles calculations. This model is highly effective in estimating transition energy values for various Ce3+-doped garnet-type oxides. The machine learning model improved the accuracy of transition energy calculations. Furthermore, we have developed predictive models that estimate first-principles calculation results based on structural data. Systematic first-principles calculations of fictitious garnet-type oxides with gradually changed structural parameters were performed and used as the training data. The predictive models of the electronic parameters such as the bond order, the crystal field splitting (εcfs) and the net charge were used to create a two-step predictive model based solely on structural parameters.

Ammonia (NH3) is pivotal in modern industry and represents a promising next‐generation carbon‐free energy carrier. Electrocatalytic nitrate reduction reaction (eNO3RR) presents viable solutions for NH3 production and removal of ambient nitrate pollutants. However, the development of eNO3RR is hindered by lacking the efficient electrocatalysts. To address this challenge, we synthesized a series of macrocyclic molecular catalysts for the heterogeneous eNO3RR. These materials possess different coordination environments around metal centers by surrounding subunits. Consequently, electronic structures of the active centers can be altered, enabling tunable activity towards eNO3RR. Our investigation reveals that metal center with an N2(pyrrole)‐N2(pyridine) configuration demonstrates superior activity over the others and achieves a high NH3 Faradaic efficiency (FE) of over 90 % within the tested range, where the highest FE of approximately 94 % is obtained. Furthermore, it achieves a production rate of 11.28 mg mgcat−1 h−1, and a turnover frequency of up to 3.28 s−1. Further tests disclose that these molecular catalysts with diverse coordination environments showed different magnetic moments. Theoretical calculation results indicate that variated coordination environments can result in a d‐band center variation which eventually affects rate‐determining step energy and calculated magnetic moments, thus establishing a correlation between electronic structure, experimental activity, and computational parameters.

Ammonia (NH3) is pivotal in modern industry and represents a promising next-generation carbon-free energy carrier. Electrocatalytic nitrate reduction reaction (eNO3RR) presents viable solutions for NH3 production and removal of ambient nitrate pollutants. However, the development of eNO3RR is hindered by lacking the efficient electrocatalysts. To address this challenge, we synthesized a series of macrocyclic molecular catalysts for the heterogeneous eNO3RR. These materials possess different coordination environments around metal centers by surrounding subunits. Consequently, electronic structures of the active centers can be altered, enabling tunable activity towards eNO3RR. Our investigation reveals that metal center with an N2(pyrrole)-N2(pyridine) configuration demonstrates superior activity over the others and achieves a high NH3 Faradaic efficiency (FE) of over 90 % within the tested range, where the highest FE of approximately 94 % is obtained. Furthermore, it achieves a production rate of 11.28 mg mgcat -1 h-1, and a turnover frequency of up to 3.28 s-1. Further tests disclose that these molecular catalysts with diverse coordination environments showed different magnetic moments. Theoretical calculation results indicate that variated coordination environments can result in a d-band center variation which eventually affects rate-determining step energy and calculated magnetic moments, thus establishing a correlation between electronic structure, experimental activity, and computational parameters.

The ethylene-forming enzyme (EFE) is a Fe(II)/2-oxoglutarate (2OG) and l-arginine (l-Arg)-dependent oxygenase that primarily decomposes 2OG into ethylene while also catalyzing l-Arg hydroxylation. While the hydroxylation mechanism in EFE is similar to other Fe(II)/2OG-dependent oxygenases, the formation of ethylene is unique. Various redesign strategies have aimed to increase ethylene production in EFE, but success has been limited, highlighting the need for alternate approaches. It is crucial to incorporate an accurate and comprehensive description of the integrative and multidimensional effects of the protein environment to enhance the redesign strategy in metalloenzymes, particularly in EFE. This involves understanding the role of the second coordination sphere (SCS) and long-range (LR) interacting residues, correlated motions, electronic structure, intrinsic electric field (IntEF), as well as the stabilization of transition states and reaction intermediates. In this study, we employ a molecular dynamics-based quantum mechanics/molecular mechanics approach to examine the integrative effects of the protein environment on reactions catalyzed by EFE variants from the first coordination sphere (FCS, D191E), SCS (A198V and R171A) and LR (E215A). The study uncovers how substitutions at different positions in EFE similarly impact the ethylene-forming reaction while posing distinct effects on the hydroxylation reaction. Results predict the effect of the variants in controlling the 2OG coordination mode in the Fe(II) center. Specifically, the study suggests that D191E uniquely prefers transitioning from an off-line to an in-line 2OG coordination mode before dioxygen binding. However, studies on the 2OG flip in the presence of off-line approaching dioxygen and dioxygen binding in the D191E variant indicate that the 2OG flip might not be feasible in the 5C Fe(II) state. Calculations show the possibility of a hydrogen atom transfer (HAT)-assisted oxygen flip in EFE and its variants (other than D191E). MD simulations elucidate the characteristic dynamic change in the α7 region in the D191E variant that might contribute to its increased hydroxylation reaction. Results indicate the possibility of forming an in-line ferryl from the IM2 (Fe(III)-partial bond intermediate) in the D191E variant. This alternative pathway from IM2 may also exist in WT EFE and other variants, which are yet to be explored. The study also delineates the impact of substitutions on the electronic structure and IntEF. Overall, the calculations support the idea that understanding the integrative and multidimensional effects of the protein environment on the reactions catalyzed by EFE variants provides the basics for improved enzyme redesign protocols of EFE to increase ethylene production. The results of this study will also contribute to the development of alternate redesign strategies for other metalloenzymes.

For theoretical search of novel rare earth (RE)-doped or transition metal (TM)-doped phosphor materials, it is important to optimize the combination of luminescent ions and host crystals. For this purpose, information on the structure-property relationships such as the structure-energy relationship is indispensable. Accordingly, prediction of the absorption and emission energies for hypothetical materials is essentially important.In our group, we conduct research to obtain such information by three approaches. (1) Analysis of detailed electronic structures based on first-principles calculations, (2) Prediction of emission and absorption energies based on first-principles calculations and machine learning, (3) Creation of structure-property maps based on systematic first-principles calculations.In this talk, these approaches will be explained in detail and some of our results will be also presented.

Obtaining a suitable chemical composition for high-entropy alloys (HEAs) with superior mechanical properties and good biocompatibility is still a formidable challenge through conventional trial-and-error methods. Here, based on a large amount of experimental data, a machine learning technique may be used to establish the relationship between the composition and the mechanical properties of the biocompatible HEAs. Subsequently, first-principles calculations are performed to verify the accuracy of the prediction results from the machine learning model. The predicted Young’s modulus and yield strength of HEAs performed very well in the previous experiments. In addition, the effect on the mechanical properties of alloying an element is investigated in the selected Ti-Zr-Hf-Nb-Ta HEA with the high crystal symmetry. Finally, the Ti8-Zr20-Hf16-Nb35-Ta21 HEA predicted by the machine learning model exhibits a good combination of biocompatibility and mechanical performance, attributed to a significant electron flow and charge recombination. This work reveals the importance of these strategies, combined with machine learning and first-principles calculations, on the development of advanced biocompatible HEAs.