UniversalQM/MMapproaches for general nanoscale applications

Wiley Interdisciplinary Reviews Computational Molecular Science, Journal Year: 2023, Volume and Issue: 13(4)

Published: Feb. 1, 2023

Abstract Quantum mechanics/molecular mechanics (QM/MM) hybrid models allow one to address chemical phenomena in complex molecular environments. Whereas this modeling approach can cope with a large system size at moderate computational costs, the are often tedious construct and require manual preprocessing expertise. As result, transferability new application areas be limited many parameters not easy adjust reference data that typically scarce. Therefore, it is desirable devise automated procedures of controllable accuracy, which enables such standardized black‐box‐type manner. Although diverse best‐practice protocols have been set up for construction individual components QM/MM model (e.g., MM potential, type embedding, choice QM region), reconcile all steps still rare. Here, we review state art focus on automation. We elaborate parametrization, atom‐economical physically‐motivated region selection, embedding schemes incorporate mutual polarization as critical model. In view broad scope field, mostly restrict discussion methodologies build de novo based first‐principles data, uncertainty quantification, error mitigation high potential Ultimately, able reliable fast efficient way without being constrained by specific or technical limitations. This article categorized under: Electronic Structure Theory > Combined Methods

Language: Английский

---

First principles reaction discovery: from the Schrodinger equation to experimental prediction for methane pyrolysis

Chemical Science, Journal Year: 2023, Volume and Issue: 14(27), P. 7447 - 7464

Published: Jan. 1, 2023

Our recent success in exploiting graphical processing units (GPUs) to accelerate quantum chemistry computations led the development of ab initio nanoreactor, a computational framework for automatic reaction discovery and kinetic model construction. In this work, we apply nanoreactor methane pyrolysis, from path refinement modeling. Elementary reactions occurring during pyrolysis are revealed using GPU-accelerated molecular dynamics simulations. Subsequently, these paths refined at higher level theory with optimized reactant, product, transition state geometries. Reaction rate coefficients calculated by based on paths. The discovered lead 53 species 134 reactions, which is validated against experimental data simulations literature models. We highlight advantage leveraging local brute force Monte Carlo sensitivity analysis approaches efficient identification important reactions. Both can further improve accuracy model. results work demonstrate power computationally affordable systematic accurate

Language: Английский

---

Graph-Driven Reaction Discovery: Progress, Challenges, and Future Opportunities

The Journal of Physical Chemistry A, Journal Year: 2022, Volume and Issue: 126(40), P. 7051 - 7069

Published: Oct. 3, 2022

Graph-based descriptors, such as bond-order matrices and adjacency matrices, offer a simple compact way of categorizing molecular structures; furthermore, descriptors can be readily used to catalog chemical reactions (i.e., bond-making -breaking). As such, number graph-based methodologies have been developed with the goal automating process generating reaction network models describing possible mechanistic chemistry in given set reactant species. Here, we outline evolution these discovery schemes, particular emphasis on more recent methods incorporating semiempirical ab initio electronic structure calculations, minimum-energy path refinements, transition state searches. Using representative examples from homogeneous catalysis interstellar chemistry, highlight how schemes increasingly act "virtual vessels" for interrogating questions. Finally, where challenges remain, including issues accuracy calculation speeds, well inherent challenge dealing vast size accessible space.

Language: Английский

---

An atom-in-molecule adaptive polarized valence single-ζ atomic orbital basis for electronic structure calculations

The Journal of Chemical Physics, Journal Year: 2023, Volume and Issue: 159(16)

Published: Oct. 25, 2023

Many low-cost or semiempirical quantum mechanical-based electronic structure methods suffer from the use of unpolarized minimal atomic orbital (AO) basis sets. In this work, we overcome limitation by a fully DFT variationally optimized, adaptive set consistently available for elements up to radon (Z = 86). The new key feature is make linear coefficients primitive Gaussians in contracted AO dependent on effective charge atom molecule, i.e., each symmetry-unique obtains its "own" specifically adapted functions. way, physically important "breathing" AOs molecule with (a) (expansion/contraction anionic/cationic states) and (b) number close-lying bonded neighbor atoms accounted for. required charges are obtained specially developed extended Hückel type Hamiltonian coordination numbers geometry. Proper analytical derivatives resulting functions can easily be derived. Moreover, electric field-dependent, thus improving description of, e.g., dipole moments polarizabilities. termed q-vSZP (charge valence single-ζ, polarized) thoroughly benchmarked atomic/molecular thermochemical properties compared standard double-ζ sets at level accurate ωB97X-D4 functional. It shown that clearly superior existing sets, often reaching quality even better results. We expect it optimal choice future mechanical methods.

Language: Английский

---

High-throughput ab initio reaction mechanism exploration in the cloud with automated multi-reference validation

The Journal of Chemical Physics, Journal Year: 2023, Volume and Issue: 158(8)

Published: Feb. 7, 2023

Quantum chemical calculations on atomistic systems have evolved into a standard approach to studying molecular matter. These often involve significant amount of manual input and expertise, although most this effort could be automated, which would alleviate the need for expertise in software hardware accessibility. Here, we present AutoRXN workflow, an automated workflow exploratory high-throughput electronic structure systems, (i) density functional theory methods are exploited deliver minimum transition-state structures corresponding energies properties, (ii) coupled cluster then launched optimized provide more accurate energy property estimates, (iii) multi-reference diagnostics evaluated back check results subject them multi-configurational potential cases. All carried out cloud environment support massive computational campaigns. Key features all components autonomy, stability, operator interference. We highlight with example autonomous reaction mechanism exploration mode action homogeneous catalyst asymmetric reduction ketones.

Language: Английский

---

A human-machine interface for automatic exploration of chemical reaction networks

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: May 1, 2024

Abstract Autonomous reaction network exploration algorithms offer a systematic approach to explore mechanisms of complex chemical processes. However, the resulting networks are so vast that an all potentially accessible intermediates is computationally too demanding. This renders brute-force explorations unfeasible, while with completely pre-defined or hard-wired constraints, such as element-specific coordination numbers, not flexible enough for systems. Here, we introduce STEERING WHEEL guide otherwise unbiased automated exploration. The algorithm intuitive, generally applicable, and enables one focus on specific regions emerging network. It also allows guiding data generation in context mechanism exploration, catalyst design, other optimization challenges. demonstrated elucidation transition metal catalysts. We highlight how catalytic cycles reproducible way. objectives fully adjustable, allowing harness both structure-specific (accurate) calculations well broad high-throughput screening possible intermediates.

Language: Английский

---

SCINE—Software for chemical interaction networks

The Journal of Chemical Physics, Journal Year: 2024, Volume and Issue: 160(22)

Published: June 10, 2024

The software for chemical interaction networks (SCINE) project aims at pushing the frontier of quantum calculations on molecular structures to a new level. While individual as well simple relations between them have become routine in chemistry, developments pushed field high-throughput calculations. Chemical may be created by search specific properties design attempt, or they can defined set elementary reaction steps that form network. modules SCINE been designed facilitate such studies. features are (i) general applicability applied methodologies ranging from electronic structure (no restriction elements periodic table) microkinetic modeling (with little restrictions molecularity), full modularity so also stand-alone programs exchanged external packages fulfill similar purpose (to increase options computational campaigns and provide alternatives case tasks hard impossible accomplish with certain programs), (ii) high stability autonomous operations control steering an operator easy possible, (iii) embedding into complex heterogeneous environments taken individually context A graphical user interface unites all ensures interoperability. All components made available open source free charge.

Language: Английский

---

ReNeGate: A Reaction Network Graph-Theoretical Tool for Automated Mechanistic Studies in Computational Homogeneous Catalysis

Journal of Chemical Theory and Computation, Journal Year: 2022, Volume and Issue: 18(12), P. 7470 - 7482

Published: Nov. 2, 2022

Exploration of the chemical reaction space transformations in multicomponent mixtures is one main challenges contemporary computational chemistry. To remove expert bias from mechanistic studies and to discover new chemistries, an automated graph-theoretical methodology proposed, which puts forward a network formalism homogeneous catalysis reactions utilizes analysis tool for studies. The method can be used analyzing trajectories with single multiple catalytic species provide unique conformers catalysts including multinuclear catalyst clusters along other mixture components. presented three-step approach has integrated ability handle systems arbitrary complexity (mixtures reactants, precursors, ligands, additives, solvents). It not limited predefined rules, does require prealignment components consistent coordinate, agnostic nature transformations. Conformer exploration, reactive event identification, are steps taken identifying pathways given starting precatalytic as input. Such allows us efficiently explore realistic conditions either previously observed or completely unknown events context representing different intermediates. Our workflow exploration exclusively focuses on identification thermodynamically feasible conversion channels, representative (secondary) deactivation inhibition paths, usually most difficult anticipate based solely knowledge. Thus, sought removed at all steps, intuition choice thermodynamic constraint imposed by applicable experimental terms threshold energy values allowed capabilities proposed have been tested exploring reactivity Mn complexes relevant hydrogenation chemistry verify postulated activation mechanisms unravel unexpected channels rare events.

Language: Английский

---

A non-self-consistent tight-binding electronic structure potential in a polarized double-ζ basis set for all spd-block elements up to Z = 86

The Journal of Chemical Physics, Journal Year: 2023, Volume and Issue: 158(12)

Published: Feb. 16, 2023

Existing semiempirical molecular orbital methods suffer from the usually minimal atomic-orbital (AO) basis set used to simplify calculations. Here, a completely new and consistently parameterized tight-binding electronic structure Hamiltonian evaluated in deeply contracted, properly polarized valence double-zeta (vDZP) is described. The inner-shell electrons are accounted for by standard, large-core effective potentials approximations them. primary target of this so-called density matrix method reproduce one-particle P ωB97X-V range-separated hybrid functional theory (DFT) calculation exactly same set. Additional properties considered energies, dipole polarizabilities moments, polarizability derivatives. key features as follows: (a) it non-self-consistent with an overall fixed number only three required diagonalizations; (b) AO overlap integrals needed construct matrix; (c) P-dependent terms emulating non-local exchange included; (d) element-specific empirical parameters (about 50 per element) need be determined. globally achieves high accuracy at speedup compared ωB97X-V/vDZP reference about 3-4 orders magnitude. It performs robustly difficult transition metal complexes, highly charged or zwitterionic systems, chemically unusual bonding situations, indicating generally robust approximation (self-consistent) Kohn-Sham potential. As example application, vibrational Raman spectrum entire protein 327 atoms respect DFT shown. This may out-of-the-box generate molecular/atomic machine learning applications accurate high-speed methods.

Language: Английский

---

Quantum chemical data generation as fill-in for reliability enhancement of machine-learning reaction and retrosynthesis planning

Digital Discovery, Journal Year: 2023, Volume and Issue: 2(3), P. 663 - 673

Published: Jan. 1, 2023

We demonstrate and discuss the feasibility of autonomous first-principles mechanistic explorations for providing quantum chemical data to enhance confidence data-driven retrosynthetic synthesis design based on molecular transformers.

Language: Английский

---