Identifying transition states-saddle points on the potential energy surface connecting reactant and product minima-is central to predicting kinetic barriers and understanding chemical reaction mechanisms. In this work, we train a fully differentiable equivariant neural network potential, NewtonNet, on thousands of organic reactions and derive the analytical Hessians. By reducing the computational cost by several orders of magnitude relative to the density functional theory (DFT) ab initio source, we can afford to use the learned Hessians at every step for the saddle point optimizations. We show that the full machine learned (ML) Hessian robustly finds the transition states of 240 unseen organic reactions, even when the quality of the initial guess structures are degraded, while reducing the number of optimization steps to convergence by 2-3× compared to the quasi-Newton DFT and ML methods. All data generation, NewtonNet model, and ML transition state finding methods are available in an automated workflow.

We investigated the pyrolysis of propyne and propene using aerosol mass spectrometry coupled with tunable synchrotron vacuum ultraviolet photoionization. The results show that propyne forms particles at pyrolysis temperatures approximately 100 K below the lowest temperature at which propene produces particles. The aerosol mass spectra show that the carbon-to-hydrogen (C/H) ratios of the pyrolysis products from both reactants increase with temperature between the particle onset temperature and 1275 K. In this temperature range, the mass peaks corresponding to species with 18 or more carbon atoms are more saturated (i.e., have lower C/H ratios) than species predicted by the stabilomer grid. The observation of relatively high saturation (low C/H ratios) at temperatures near particle onset is consistent with the presence of aliphatically linked hydrocarbons during the preliminary stages of soot formation. The masses of the species presumed to be aliphatically linked suggest that these linked species are adducts of species smaller than chrysene (C18H12). We performed a comparative analysis of the initial PAH-growth pathways consistent with aerosol mass spectra and associated mass-specific photoionization-efficiency curves. The results from the comparative analysis indicate that the development of PAHs up to the size of acenaphthylene (C12H8) follow similar formation pathways for propyne and propene, but the isomeric composition of species larger than acenaphthylene diverges for the two reactants.

Methoxymethanol (CH3OCH2OH) is a reactive C2 ether-alcohol that is formed by coupling events in both heterogeneous and homogeneous systems. It is found in complex reactive environments-for example those associated with catalytic reactors, combustion systems, and liquid-phase mixtures of oxygenates. Using tunable synchrotron-generated vacuum-ultraviolet photons between 10.0 and 11.5 eV, we report on the photoionization spectroscopy of methoxymethanol. We determine that the lowest-energy photoionization process is the dissociative ionization of methoxymethanol via H-atom loss to produce [C2H5O2]+, a fragment cation with a mass-to-charge ratio (m/z) = 61.029. We measure the appearance energy of this fragment ion to be 10.24 ± 0.05 eV. The parent cation is not detected in the energy range examined. To elucidate the origin of the m/z = 61.029 (C2H5O2) fragment, we used automated electronic structure calculations to identify key stationary points on the cation potential energy surface and compute conformer-specific microcanonical rate coefficients for the important unimolecular processes. The calculated H-atom dissociation pathway results in a [C2H5O2]+ fragment appearance at 10.21 eV, in excellent agreement with experimental results.

A crucial chain-branching step in autoignition is the decomposition of ketohydroperoxides (KHP) to form an oxy radical and OH. Other pathways compete with chain-branching, such as "Korcek" dissociation of γ-KHP to a carbonyl and an acid. Here we characterize the formation of a γ-KHP and its decomposition to formic acid+acetone products from observations of n-butane oxidation in two complementary experiments. In jet-stirred reactor measurements, KHP is observed above 590 K. The KHP concentration decreases with increasing temperature, whereas formic acid and acetone products increase. Observation of characteristic isotopologs acetone-d₃ and formic acid-d₀ in the oxidation of CH₃ CD₂ CD₂ CH₃ is consistent with a Korcek mechanism. In laser-initiated oxidation experiments of n-butane, formic acid and acetone are produced on the timescale of KHP removal. Modelling the time-resolved production of formic acid provides an estimated upper limit of 2 s^-1 for the rate coefficient of KHP decomposition to formic acid+acetone.