Quantum chemical calculations for reaction prediction in the development of synthetic methodologies

Chemical Science, Journal Year: 2023, Volume and Issue: 14(42), P. 11601 - 11616

Published: Jan. 1, 2023

This perspective showcases how quantum chemical calculations drive predictive strategies to explore unknown reactions, catalysts, and synthetic routes toward complex molecules in methodology development.

Language: Английский

---

Pathfinder─Navigating and Analyzing Chemical Reaction Networks with an Efficient Graph-Based Approach

Journal of Chemical Information and Modeling, Journal Year: 2022, Volume and Issue: 63(1), P. 147 - 160

Published: Dec. 14, 2022

While the field of first-principles explorations into chemical reaction space has been continuously growing, development strategies for analyzing resulting networks (CRNs) is lagging behind. A CRN consists compounds linked by reactions. Analyzing how these are transformed one another based on kinetic modeling a nontrivial task. Here, we present graph-optimization-driven algorithm and program Pathfinder to allow such an analysis CRN. The this work obtained with our open-source Chemoton network exploration software. probes reactive combinations elementary steps sorts them By encoding reactions as graph consisting compound vertices adding information about activation barriers well required reagents edges yields complete graph-theoretical representation Since probabilities formation depend starting conditions, consumption any during must be accounted reflect availability reagents. To account this, introduce costs availability. Simultaneously, determined rank in terms their probability formed. This ranking then allows us probe easily accessible first further yet unexplored terrain. We illustrate working principle abstract small Afterward, demonstrated example disproportionation iodine water comproportionation iodic acid hydrogen iodide. Both processes analyzed within same CRN, which construct autonomous software [Unsleber, J. P.;

Language: Английский

---

Multi-Time-Scale Simulation of Complex Reactive Mixtures: How Do Polyoxometalates Form?

Journal of the American Chemical Society, Journal Year: 2023, Volume and Issue: 145(34), P. 18920 - 18930

Published: July 27, 2023

Understanding the dynamics of reactive mixtures still challenges both experiments and theory. A relevant example can be found in chemistry molecular metal-oxide nanoclusters, also known as polyoxometalates. The high number species potentially involved, interconnectivity reaction network, precise control pH concentrations needed synthesis such make theoretical/computational treatment processes cumbersome. This work addresses this issue relying on a unique combination recently developed computational methods that tackle construction, kinetic simulation, analysis complex chemical networks. By using Bell-Evans-Polanyi approximation for estimating activation energies, an accurate robust linear scaling correcting computed pKa values, we report herein multi-time-scale simulations self-assembly polyoxotungstates comprise 22 orders magnitude, from tens femtoseconds to months time. very large time span was required reproduce fast acid/base equilibria (at 10-12 s), relatively slow reactions formation key clusters metatungstate 103 assembly decatungstate 106 s). Analysis data network topology shed light onto details main mechanisms, which explains origin thermodynamic followed by reaction. Simulations at alkaline fully experimental evidence since do not form under those conditions.

Language: Английский

---

Deep reaction network exploration of glucose pyrolysis

Proceedings of the National Academy of Sciences, Journal Year: 2023, Volume and Issue: 120(34)

Published: Aug. 14, 2023

Resolving the reaction networks associated with biomass pyrolysis is central to understanding product selectivity and aiding catalyst design produce more valuable products. However, even network of relatively simple [Formula: see text]-D-glucose remains unresolved due its significant complexity in terms depth number major Here, a transition-state-guided exploration has been performed that provides complete pathways most experimental products text]-D-glucose. The resulting involves over 31,000 reactions transition states computed at semiempirical quantum chemistry level approximately 7,000 kinetically relevant characterized density function theory, comprising largest reported for pyrolysis. was conducted using graph-based rules explore reactivities intermediates an adaption Dijkstra algorithm identify intermediates. This policy surprisingly (re)identified products, many proposed by previous computational studies, also identified new low-barrier mechanisms resolve outstanding discrepancies between yields isotope labeling experiments. explanatory high yield hydroxymethylfurfural pathway contributes formation hydroxyacetaldehyde during glucose Due limited domain knowledge required generate this network, approach should be transferable other complex prediction problems

Language: Английский

---

Uncertainty-Aware First-Principles Exploration of Chemical Reaction Networks

The Journal of Physical Chemistry A, Journal Year: 2024, Volume and Issue: 128(22), P. 4532 - 4547

Published: May 24, 2024

Exploring large chemical reaction networks with automated exploration approaches and accurate quantum methods can require prohibitively computational resources. Here, we present an approach that focuses on the kinetically relevant part of network by interweaving (i) large-scale reactions, (ii) identification parts through microkinetic modeling, (iii) quantification propagation uncertainties, (iv) refinement. Such uncertainty-aware a accuracy improvement has not been demonstrated before in fully mechanical approach. Uncertainties are identified local or global sensitivity analysis. The is refined rolling fashion during exploration. Moreover, uncertainties considered steering We demonstrate our for Eschenmoser–Claisen rearrangement reactions. analysis identifies only small number reactions compounds essential describing kinetics reliably, resulting efficient explorations without sacrificing requiring prior knowledge about chemistry unfolding.

Language: Английский

---

Nanoscale chemical reaction exploration with a quantum magnifying glass

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

Published: June 22, 2024

Abstract Nanoscopic systems exhibit diverse molecular substructures by which they facilitate specific functions. Theoretical models of them, aim at describing, understanding, and predicting these capabilities, are difficult to build. Viable quantum-classical hybrid come with challenges regarding atomistic structure construction quantum region selection. Moreover, if their dynamics mapped onto a state-to-state mechanism such as chemical reaction network, its exhaustive exploration will be impossible due the combinatorial explosion space. Here, we introduce “quantum magnifying glass” that allows one interactively manipulate nanoscale structures level. The glass seamlessly combines autonomous model parametrization, ultra-fast mechanical calculations, automated exploration. It represents an approach investigate complex sequences in physically consistent manner unprecedented effortlessness real time. We demonstrate features for reactions bio-macromolecules metal-organic frameworks, highlight general applicability.

Language: Английский

---

Least popular vibrational entropy model provides the best accuracy and robustness

The Journal of Chemical Physics, Journal Year: 2025, Volume and Issue: 162(12)

Published: March 28, 2025

Vibrational contributions into free energies usually amount to several kcal/mol and can significantly affect computational predictions. However, they are generally estimated incorrectly for chemical systems in solutions because the usual (employed ∼99% of cases) models vibrational entropies extremely sensitive errors low-lying frequencies (below 300 cm−1), these involve solvent molecules that neglected (computed implicitly) quantum calculations. We find only one entropy approximation—the proposed by Truhlar 2011—which is used ∼2% cases, stable frequency region does not exhibit this problem. Accordingly, approximation shows best accuracy robustness on a diverse set experimental complexation be somewhat improved even further.

Language: Английский

---

Learning Inducing Points and Uncertainty on Molecular Data by Scalable Variational Gaussian Processes

SIAM/ASA Journal on Uncertainty Quantification, Journal Year: 2025, Volume and Issue: 13(2), P. 543 - 562

Published: April 23, 2025

Language: Английский

---

Scientific Deep Machine Learning Concepts for the Prediction of Concentration Profiles and Chemical Reaction Kinetics: Consideration of Reaction Conditions

The Journal of Physical Chemistry A, Journal Year: 2024, Volume and Issue: 128(5), P. 929 - 944

Published: Jan. 25, 2024

Emerging concepts from scientific deep machine learning such as physics-informed neural networks (PINNs) enable a data-driven approach for the study of complex kinetic problems. We present an extended framework that combines advantages PINNs with detailed consideration experimental parameter variations simulation and prediction chemical reaction kinetics. The is based on truncated Taylor series expansions underlying fundamental equations, whereby external can be interpreted perturbations parameters. Accordingly, our method allows efficient settings their influence concentration profiles A particular advantage approach, in addition to univariate multivariate variations, robust model-based exploration space determine optimal conditions combination advanced insights. benefits this concept are demonstrated higher-order reactions including catalytic oscillatory systems small amounts training data. All predicted values show high level accuracy, demonstrating broad applicability flexibility approach.

Language: Английский

---

Exploring Chemical Space Using Ab Initio Hyperreactor Dynamics

ACS Central Science, Journal Year: 2024, Volume and Issue: 10(2), P. 302 - 314

Published: Jan. 31, 2024

In recent years, first-principles exploration of chemical reaction space has provided valuable insights into intricate networks. Here, we introduce ab initio hyperreactor dynamics, which enables rapid screening the accessible from a given set initial molecular species, predicting new synthetic routes that can potentially guide subsequent experimental studies. For this purpose, different hyperdynamics derived bias potentials are applied along with pressure-inducing spherical confinement system in dynamics simulations to efficiently enhance reactivity under mild conditions. To showcase advantages and flexibility approach, present systematic study method's parameters on HCN toy model apply it recently introduced for prebiotic formation glycinal acetamide interstellar ices, yields results line findings. addition, show how developed framework complicated transitions like first step nonenzymatic DNA nucleoside synthesis an aqueous environment, where fragmentation problem earlier nanoreactor approaches is avoided.

Language: Английский

---