A Full-Dimensional Potential Energy Surface and Dynamics of the Multichannel Reaction between H and HO2

Fundamental Invariant Neural Network (FI-NN) Potential Energy Surface for the OH + CH3OH Reaction with Analytical Forces

The hydrogen abstraction reaction of OH + CH3OH plays a great role in combustion and atmospheric and interstellar chemistry and has been extensively studied theoretically and experimentally. Theoretically, the numerical gradients with respect to the Cartesian coordinates of atoms in molecular simulations on our recent potential energy surface (PES) for the title reaction trained using the permutationally invariant polynomial neural network (PIP-NN) approach hinder the extensive calculation because of the unaffordable computation cost. To address this issue, we in this work report a new full-dimensional accurate analytical PES for the title reaction using the fundamental invariant neural network (FI-NN) approach based on 140,192 points of the quality UCCSD(T)-F12a/AVTZ. Besides, the spin-orbit (SO) corrections of OH in the entrance channel were determined at the level of complete active space self-consistent field with the AVTZ basis set. As a compromise between computational cost and efficiency, the Δ-machine learning approach was employed to construct the SO-corrected PES. Based on this new FI-NN PES with analytical forces, thermal rate coefficients and various dynamic properties, including the integral cross sections, the differential cross sections, and the product energy partitioning, were determined by running a total of 5.5 million trajectories. The use of analytical gradients of the FI-NN PES accelerated simulations and about 99% of computation cost was saved, compared to that for the PIP-NN PES with numerical gradients. Such a significant acceleration is achieved mainly by replacing PIPs with FIs.

Importance of Spin Channels from Radical-Radical Reactions in Hydrogen-Oxygen Combustion Mechanisms at High Temperatures

Radical-radical reactions can generate two channels with high and low spins. In this work, ten radical-radical reactions with different spin channels and four radical-molecule reactions in hydrogen-oxygen combustion were systematically investigated from a theoretical perspective. The potential energy surface (PES) of radical-radical reactions reveals that the high- and low-spin states of the reactant are energetically degenerate and the two channels are energetically feasible. The difference in rate constants between the high- and low-spin channels gradually decreases as the temperature increases. Then, the kinetic parameters of the 14 bimolecular reactions in the hydrogen-oxygen mechanism of the University of California, San Diego (UCSD), were replaced to simulate the ignition delay time and laminar flame speed. The simulation results agree well with the available experimental findings, indicating the necessity of considering both high- and low-spin channels for kinetic simulation.

Quasi-Classical Trajectory Calculation of Rate Constants Using an Ab Initio Trained Machine Learning Model (aML-MD) with Multifidelity Data

Machine learning (ML) provides a great opportunity for the construction of models with improved accuracy in classical molecular dynamics (MD). However, the accuracy of a ML trained model is limited by the quality and quantity of the training data. Generating large sets of accurate ab initio training data can require significant computational resources. Furthermore, inconsistent or incompatible data with different accuracies obtained using different methods may lead to biased or unreliable ML models that do not accurately represent the underlying physics. Recently, transfer learning showed its potential for avoiding these problems as well as for improving the accuracy, efficiency, and generalization of ML models using multifidelity data. In this work, ab initio trained ML-based MD (aML-MD) models are developed through transfer learning using DFT and multireference data from multiple sources with varying accuracy within the Deep Potential MD framework. The accuracy of the force field is demonstrated by calculating rate constants for the H + HO2 → H2 + 3O2 reaction using quasi-classical trajectories. We show that the aML-MD model with transfer learning can accurately predict the rate constants while reducing the computational cost by more than five times compared to the use of more expensive quantum chemistry training data sets. Hence, the aML-MD model with transfer learning shows great potential in using multifidelity data to reduce the computational cost involved in generating the training set for these potentials.

Role of curvature in acridone for 1 O2 oxidation of a natural product homoallylic alcohol: A novel iso-hydroperoxide intermediate

A density functional theoretical (DFT) study is presented, implicating a 1 O2 oxidation process to reach a dihydrobenzofuran from the reaction of the natural homoallylic alcohol, glycocitrine. Our results predict an interconversion between glycocitrine and an iso-hydroperoxide intermediate [R(H)O+ -O- ] that provides a key path in the chemistry which then follows. Formations of allylic hydroperoxides are unlikely from a 1 O2 'ene' reaction. Instead, the dihydrobenzofuran arises by 1 O2 oxidation facilitated by a 16° curvature of the glycocitrine ring imposed by a pyramidal N-methyl group. This curvature facilitates the formation of the iso-hydroperoxide, which is analogous to the iso species CH2 I+ -I- and CHI2 + -I- formed by UV photolysis of CH2 I2 and CHI3 . The iso-hydroperoxide is also structurally reminiscent of carbonyl oxides (R2 C=O+ -O- ) formed in the reaction of carbenes and oxygen. Our DFT results point to intermolecular process, in which the iso-hydroperoxide's fate relates to O-transfer and H2 O dehydration reactions for new insight into the biosynthesis of dihydrobenzofuran natural products.

Deciphering Dynamics of the Cl + SiH4 → H + SiH3Cl Reaction on a Machine Learning Made Globally Accurate Full-Dimensional Potential Energy Surface

Chemical reaction dynamics needs the joint effort from both experiment and theory, and theory is useful to rationalize the experimental results by offering intimate details of chemical reaction dynamics and to explore new reaction pathways. With the aid of machine learning, we develop here an accurate full-dimensional potential energy surface (PES) for the reaction between Cl + SiH₄. This PES can describe well the hydrogen abstraction channel to HCl + SiH₃. It can also give a good description for the hydrogen substitution channel to H + SiH₃Cl, which is the focus of the current study and has never been reported by theory. The dynamics of this substitution channel is revealed in detail by calculating ample quasi-classical trajectories (QCTs) on the new PES. The computed product angular distributions are in good agreement with the only crossed molecular beam experiment. Both theory and experiment suggest that this channel takes place mainly via the typical S_N2 inversion mechanism. Theory reveals that there also exists a novel torsion mechanism for the substitution channel. Two dynamic mechanisms are analyzed in detail. The present detailed theoretical dynamics study sheds insightful and novel understanding for this fundamentally important chemical reaction.

Formation of Hydrogen Peroxide from O-(H2O)n Clusters

Hydrogen peroxide (H₂O₂) has recently received much attention as a safe and clean energy carrier for hydrogen molecules. In this study, based on direct ab initio molecular dynamics (AIMD) calculations, we demonstrated that H₂O₂ is directly formed via the photoelectron detachment of O^-(H₂O)_n (n = 1-6) (water clusters of an oxygen radical anion). Three electronic states of oxygen atoms were examined in the calculations: O(X)(H₂O)_n (X = ³P, ¹D, and ¹S states). After the photoelectron detachment of O^-(H₂O)_n (n = 1) to the ¹S state, a complex comprising O(¹S) and H₂O, O(¹S)-OH₂, was formed. A hydrogen atom of H₂O immediately transferred to O(¹S) during an intracluster reaction to form H₂O₂ as the final product. Simulations were run to obtain a total of 33 trajectories for n = 1 that all led to the formation of H₂O₂. The average reaction time of H₂O₂ formation was calculated to be 57.7 fs in the case of n = 1, indicating that the reaction was completed within 100 fs of electron detachment. All the reaction systems O(¹S)(H₂O)_n (n = 1-6) indicated the formation of H₂O₂ by the same mechanism. The reaction times for n = 2-6 were calculated to range between 80 and 180 fs, indicating that the reaction for n = 1 is faster than that of the larger clusters, that is, the larger the cluster size, the slower the reaction is. The reaction dynamics of the triplet O(³P) and singlet O(¹D) potential energy surfaces were calculated for comparison. All calculations yielded the dissociation product O(X)(H₂O)_n → O(X) + (H₂O)_n (X = ³P and ¹D), indicating that the O(¹S) state contributes to the formation of H₂O₂. The reaction mechanism was discussed based on the theoretical results.