Moving from postdictive to predictive kinetics in reaction engineering

Uncertainty Qualification for Deep Learning-Based Elementary Reaction Property Prediction

The prediction of the thermodynamic and kinetic properties of elementary reactions has shown rapid improvement due to the implementation of deep learning (DL) methods. While various studies have reported the success in predicting reaction properties, the quantification of prediction uncertainty has seldom been investigated, thus compromising the confidence in using these predicted properties in practical applications. Here, we integrated graph convolutional neural networks (GCNN) with three uncertainty prediction techniques, including deep ensemble, Monte Carlo (MC)-dropout, and evidential learning, to provide insights into the uncertainty quantification and utility. The deep ensemble model outperforms others in accuracy and shows the highest reliability in estimating prediction uncertainty across all elementary reaction property data sets. We also verified that the deep ensemble model showed a satisfactory capability in recognizing epistemic and aleatoric uncertainties. Additionally, we adopted a Monte Carlo Tree Search method for extracting the explainable reaction substructures, providing a chemical explanation for DL predicted properties and corresponding uncertainties. Finally, to demonstrate the utility of uncertainty qualification in practical applications, we performed an uncertainty-guided calibration of the DL-constructed kinetic model, which achieved a 25% higher hit ratio in identifying dominant reaction pathways compared to that of the calibration without uncertainty guidance.

Accelerated design of nickel-cobalt based catalysts for CO2 hydrogenation with human-in-the-loop active machine learning

Thermo-catalytic conversion of CO2 into more valuable compounds, such as methane, is an attractive strategy for energy storage in chemical bonds and creating a carbon-based circular economy. However, designing heterogeneous catalysts remains a challenging, time- and resource-consuming task. Herein, we present an interpretable, human-in-the-loop active machine learning framework to efficiently plan catalytic experiments, execute them in an automated set-up, and estimate the effect of experimental variables on the catalytic activity. A dataset with 48 catalytic activity tests was compiled from a design space of Ni-Co/Al2O3 catalysts with over 50 million potential combinations in only eight iterations. This small dataset was found sufficient to predict CO2 conversion, methane selectivity, and methane space-time yield with remarkable accuracy (R 2 > 0.9) for untested catalysts and reaction conditions. New experiments and catalysts were selected with this methodology, leading to experimental conditions that improved the methane space-time yield by nearly 50% in comparison to the previously obtained maximum in the dataset. Interpretation of the model predictions unveiled the effect of each catalyst descriptor and reaction condition on the outcome. Particularly, the strong predicted inverse trend between the calcination temperature and the catalytic activity was validated experimentally, and characterization implied an underlying structure-performance relationship. Finally, it is demonstrated that the deployed active learning model is excellently suited to predict and fit kinetic trends with a minimal amount of data. This data-driven framework is a first step to faster, model-based, and interpretable design of catalysts and holds promise for broader applications across catalytic processes.

Accurately Predicting Barrier Heights for Radical Reactions in Solution Using Deep Graph Networks

Quantitative estimates of reaction barriers and solvent effects are essential for developing kinetic mechanisms and predicting reaction outcomes. Here, we create a new data set of 5,600 unique elementary radical reactions calculated using the M06-2X/def2-QZVP//B3LYP-D3(BJ)/def2-TZVP level of theory. A conformer search is done for each species using TPSS/def2-TZVP. Gibbs free energies of activation and of reaction for these radical reactions in 40 common solvents are obtained using COSMO-RS for solvation effects. These balanced reactions involve the elements H, C, N, O, and S, contain up to 19 heavy atoms, and have atom-mapped SMILES. All transition states are verified by an intrinsic reaction coordinate calculation. We next train a deep graph network to directly estimate the Gibbs free energy of activation and of reaction in both gas and solution phases using only the atom-mapped SMILES of the reactant and product and the SMILES of the solvent. This simple input representation avoids computationally expensive optimizations for the reactant, transition state, and product structures during inference, making our model well-suited for high-throughput predictive chemistry and quickly providing information for (retro-)synthesis planning tools. To properly measure model performance, we report results on both interpolative and extrapolative data splits and also compare to several baseline models. During training and testing, the data set is augmented by including the reverse direction of each reaction and variants with different resonance structures. After data augmentation, we have around 2 million entries to train the model, which achieves a testing set mean absolute error of 1.16 kcal mol-1 for the Gibbs free energy of activation in solution. We anticipate this model will accelerate predictions for high-throughput screening to quickly identify relevant reactions in solution, and our data set will serve as a benchmark for future studies.

Benchmark of general-purpose machine learning-based quantum mechanical method AIQM1 on reaction barrier heights

Artificial intelligence-enhanced quantum mechanical method 1 (AIQM1) is a general-purpose method that was shown to achieve high accuracy for many applications with a speed close to its baseline semiempirical quantum mechanical (SQM) method ODM2*. Here, we evaluate the hitherto unknown performance of out-of-the-box AIQM1 without any refitting for reaction barrier heights on eight datasets, including a total of ∼24 thousand reactions. This evaluation shows that AIQM1's accuracy strongly depends on the type of transition state and ranges from excellent for rotation barriers to poor for, e.g., pericyclic reactions. AIQM1 clearly outperforms its baseline ODM2* method and, even more so, a popular universal potential, ANI-1ccx. Overall, however, AIQM1 accuracy largely remains similar to SQM methods (and B3LYP/6-31G* for most reaction types) suggesting that it is desirable to focus on improving AIQM1 performance for barrier heights in the future. We also show that the built-in uncertainty quantification helps in identifying confident predictions. The accuracy of confident AIQM1 predictions is approaching the level of popular density functional theory methods for most reaction types. Encouragingly, AIQM1 is rather robust for transition state optimizations, even for the type of reactions it struggles with the most. Single-point calculations with high-level methods on AIQM1-optimized geometries can be used to significantly improve barrier heights, which cannot be said for its baseline ODM2* method.

Non-physical Species in Chemical Kinetic Models: A Case Study of Diazenyl Hydroxy and Diazenyl Peroxide

Predictive chemical kinetic models often consider hundreds to thousands of intermediate species. An even greater number of species are required to generate pressure-dependent reaction networks for gas-phase systems. As this immense chemical search space is being explored using automated tools by applying reaction templates, it is probable that non-physical species will infiltrate the model without being recognized by the compute or a human as such. These non-physical species might obey chemical intuition as well as requirements coded in the software, e. g., obeying element electron valence constraints, and may consequently remain unnoticed. Non-physical species become an acute problem when their presence affects a model observable. Correcting a pressure-dependent network containing a non-physical species may significantly affect the computed rate coefficient. The present work discusses and analyzes two specific cases of such species, diazenyl hydroxy (⋅N=NOH) and diazenyl peroxide (⋅N=NOOH), both previously suggested as intermediates in nitrogen combustion systems. A comprehensive conformational search did not identify any non-fragmented energy well, and energy scans performed for diazenyl peroxide (⋅N=NOOH), at DFT and CCSD(T) show that it barrierlessly decomposes. This work highlights a broad implication for future automated chemical kinetic model generation, and provides a significant motivation to standardize non-physical species identification in chemical kinetic models.

High accuracy barrier heights, enthalpies, and rate coefficients for chemical reactions

Quantitative chemical reaction data, including activation energies and reaction rates, are crucial for developing detailed kinetic mechanisms and accurately predicting reaction outcomes. However, such data are often difficult to find, and high-quality datasets are especially rare. Here, we use CCSD(T)-F12a/cc-pVDZ-F12//ωB97X-D3/def2-TZVP to obtain high-quality single point calculations for nearly 22,000 unique stable species and transition states. We report the results from these quantum chemistry calculations and extract the barrier heights and reaction enthalpies to create a kinetics dataset of nearly 12,000 gas-phase reactions. These reactions involve H, C, N, and O, contain up to seven heavy atoms, and have cleaned atom-mapped SMILES. Our higher-accuracy coupled-cluster barrier heights differ significantly (RMSE of ∼5 kcal mol-1) relative to those calculated at ωB97X-D3/def2-TZVP. We also report accurate transition state theory rate coefficients [Formula: see text] between 300 K and 2000 K and the corresponding Arrhenius parameters for a subset of rigid reactions. We believe this data will accelerate development of automated and reliable methods for quantitative reaction prediction.

Fast Predictions of Reaction Barrier Heights: Toward Coupled-Cluster Accuracy

Quantitative estimates of reaction barriers are essential for developing kinetic mechanisms and predicting reaction outcomes. However, the lack of experimental data and the steep scaling of accurate quantum calculations often hinder the ability to obtain reliable kinetic values. Here, we train a directed message passing neural network on nearly 24,000 diverse gas-phase reactions calculated at CCSD(T)-F12a/cc-pVDZ-F12//ωB97X-D3/def2-TZVP. Our model uses 75% fewer parameters than previous studies, an improved reaction representation, and proper data splits to accurately estimate performance on unseen reactions. Using information from only the reactant and product, our model quickly predicts barrier heights with a testing MAE of 2.6 kcal mol^-1 relative to the coupled-cluster data, making it more accurate than a good density functional theory calculation. Furthermore, our results show that future modeling efforts to estimate reaction properties would significantly benefit from fine-tuning calibration using a transfer learning technique. We anticipate this model will accelerate and improve kinetic predictions for small molecule chemistry.

Concluding Remarks on Faraday Discussion on Unimolecular Reactions

This Faraday Discussion, marking the centenary of Lindemann’s explanation of the pressure-dependence of unimolecular reactions, presented recent advances in measuring and computing collisional energy transfer efficiencies, microcanonical rate coefficients, pressure-dependent...

Kinetic Modeling of API Oxidation: (1) The AIBN/H2O/CH3OH Radical "Soup"

Stress testing of active pharmaceutical ingredients (API) is an important tool used to gauge chemical stability and identify potential degradation products. While different flavors of API stress testing systems have been used in experimental investigations for decades, the detailed kinetics of such systems as well as the chemical composition of prominent reactive species, specifically reactive oxygen species, are unknown. As a first step toward understanding and modeling API oxidation in stress testing, we investigated a typical radical "soup" solution an API is subject to during stress testing. Here we applied ab initio electronic structure calculations to automatically generate and refine a detailed chemical kinetics model, taking a fresh look at API oxidation. We generated a detailed kinetic model for a representative azobis(isobutyronitrile) (AIBN)/H₂O/CH₃OH stress-testing system with a varied cosolvent ratio (50%/50%-99.5%/0.5% vol water/methanol) for 5.0 mM AIBN and representative pH values of 4-10 at 40 °C that was stirred and open to the atmosphere. At acidic conditions, hydroxymethyl alkoxyl is the dominant alkoxyl radical, and at basic conditions, for most studied initial methanol concentrations, cyanoisopropyl alkoxyl becomes the dominant alkoxyl radical, albeit at an overall lower concentration. At acidic conditions, the levels of cyanoisopropyl peroxyl, hydroxymethyl peroxyl, and hydroperoxyl radicals are relatively high and comparable, while, at both neutral and basic pH conditions, superoxide becomes the prominent radical in the system. The present work reveals the prominent species in a common model API stress testing system at various cosolvent and pH conditions, sets the stage for an in-depth quantitative API kinetic study, and demonstrates the usage of novel software tools for automated chemical kinetic model generation and ab initio refinement.

Reaction Mechanism Generator v3.0: Advances in Automatic Mechanism Generation

In chemical kinetics research, kinetic models containing hundreds of species and tens of thousands of elementary reactions are commonly used to understand and predict the behavior of reactive chemical systems. Reaction Mechanism Generator (RMG) is a software suite developed to automatically generate such models by incorporating and extrapolating from a database of known thermochemical and kinetic parameters. Here, we present the recent version 3 release of RMG and highlight improvements since the previously published description of RMG v1.0. Most notably, RMG can now generate heterogeneous catalysis models in addition to the previously available gas- and liquid-phase capabilities. For model analysis, new methods for local and global uncertainty analysis have been implemented to supplement first-order sensitivity analysis. The RMG database of thermochemical and kinetic parameters has been significantly expanded to cover more types of chemistry. The present release includes parallelization for faster model generation and a new molecule isomorphism approach to improve computational performance. RMG has also been updated to use Python 3, ensuring compatibility with the latest cheminformatics and machine learning packages. Overall, RMG v3.0 includes many changes which improve the accuracy of the generated chemical mechanisms and allow for exploration of a wider range of chemical systems.

The impact of NOx addition on the ignition behaviour of n-pentane

Modern engine concepts present several opportunities for nitrogen combustion chemistry, particularly the interaction of NOx (NO + NO2) with fuel fragments and products of partial combustion.