Study on the Reaction Mechanism and Kinetics of the Thermal Decomposition of Nitroethane

Synergies Between Quantum Mechanics and Machine Learning in Reaction Prediction

Machine learning (ML) and quantum mechanical (QM) methods can be used in two-way synergy to build chemical reaction expert systems. The proposed ML approach identifies electron sources and sinks among reactants and then ranks all source-sink pairs. This addresses a bottleneck of QM calculations by providing a prioritized list of mechanistic reaction steps. QM modeling can then be used to compute the transition states and activation energies of the top-ranked reactions, providing additional or improved examples of ranked source-sink pairs. Retraining the ML model closes the loop, producing more accurate predictions from a larger training set. The approach is demonstrated in detail using a small set of organic radical reactions.

Thermal decomposition pathways for 1,1-diamino-2,2-dinitroethene (FOX-7)

In this study, we computationally investigate the initial and subsequent steps in the chemical mechanism for the gas-phase thermal decomposition of 1,1-diamino-2,2-dinitroethene (FOX-7). We determine the key exothermic step in the gas-phase thermal decomposition of FOX-7 and explore the similarities and differences between FOX-7 and other geminal dinitro energetic materials. The calculations reveal a mechanism for NO loss involving a 3-member cyclic intermediate, rather than a nitro-nitrite isomerization, that occurs in the radical intermediates formed throughout the decomposition mechanism.

Comment on “Decomposition mechanisms of trinitroalkyl compounds: a theoretical study from aliphatic to aromatic nitro compounds” by G. Fayet, P. Rotureau, B. Minisini, Phys. Chem. Chem. Phys., 2014, 16, 6614

The approach proposed in the original paper yields spurious contributions to both enthalpy and entropy of activation of barrierless reactions. This renders reaction branching ratios intrinsically biased towards radical decomposition of nitro species.

Decomposition mechanisms of trinitroalkyl compounds: a theoretical study from aliphatic to aromatic nitro compounds

The chemical mechanisms involved in the decomposition of trinitroethyl compounds were studied for both aliphatic and aromatic derivatives using density functional theory calculations. At first, in the case of 1,1,1-trinitrobutane, used as a reference molecule, two primary channels were highlighted among the five investigated ones: the breaking of the C-N bond and the HONO elimination. Then, the influence of various structural parameters was studied for these two reactions by changing the length of the carbon chain, adding substituents or double bonds along the carbon chain. If some slight changes in activation energies were observed for most of these features, no modification of the competition between the two investigated reactions was highlighted and the breaking of the C-N bond remained the favoured mechanism. At last, the reactions involving the trinitroalkyl fragments were highlighted to be more competitive than reactions involving nitro groups linked to aromatic cycles in two aromatic systems (4-(1,1,1-trinitrobutyl)-nitrobenzene and 2-(1,1,1-trinitrobutyl)-nitrobenzene). This showed that aromatic nitro compounds with trinitroalkyl derivatives decompose from their alkyl part and may be considered more likely as aliphatic than as aromatic regarding the initiation of their decomposition process.

Does addition of NO2 to carbon-centered radicals yield RONO or RNO2? An investigation using distonic radical ions

Nitrogen dioxide is used as a "radical scavenger" to probe the position of carbon-centered radicals within complex radical ions in the gas phase. As with analogous neutral radical reactions, this addition results in formation of an [M + NO2](+) adduct, but the structural identity of this species remains ambiguous. Specifically, the question remains: do such adducts have a nitro- (RNO2) or nitrosoxy- (RONO) moiety, or are both isomers present in the adduct population? In order to elucidate the products of such reactions, we have prepared and isolated three distonic phenyl radical cations and observed their reactions with nitrogen dioxide in the gas phase by ion-trap mass spectrometry. In each case, stabilized [M + NO2](+) adduct ions are observed and isolated. The structure of these adducts is probed by collision-induced dissociation and ultraviolet photodissociation action spectroscopy and a comparison made to the analogous spectra of authentic nitro- and nitrosoxy-benzenes. We demonstrate unequivocally that for the phenyl radical cations studied here, all stabilized [M + NO2](+) adducts are exclusively nitrobenzenes. Electronic structure calculations support these mass spectrometric observations and suggest that, under low-pressure conditions, the nitrosoxy-isomer is unlikely to be isolated from the reaction of an alkyl or aryl radical with NO2. The combined experimental and theoretical results lead to the prediction that stabilization of the nitrosoxy-isomer will only be possible for systems wherein the energy required for dissociation of the RO-NO bond (or other low energy fragmentation channels) rises close to, or above, the energy of the separated reactants.

A Novel Mechanism for Nitric Oxide Production in Nitroalkyl Radicals that Circumvents Nitro-Nitrite Isomerization

In this study, we present a novel mechanism for NO loss from nitroalkyl radicals that circumvents the traditional higher-energy nitro-nitrite isomerization. We characterize the intrinsic reaction coordinate at the B3LYP/6-311++g(3df,2p) level of theory and calculate the transition-state energies using the G4 composite method; the subsequent dynamics en route to the highly exothermic NO + acetone product channel proceeds through a three-membered ring intermediate. Crossed laser-molecular beam scattering experiments on the 2-nitro-2-propyl radical confirm the importance of this new mechanism in determining the product branching.

Speeding up chemical searches using the inverted index: the convergence of chemoinformatics and text search methods

In ligand-based screening, retrosynthesis, and other chemoinformatics applications, one often seeks to search large databases of molecules in order to retrieve molecules that are similar to a given query. With the expanding size of molecular databases, the efficiency and scalability of data structures and algorithms for chemical searches are becoming increasingly important. Remarkably, both the chemoinformatics and information retrieval communities have converged on similar solutions whereby molecules or documents are represented by binary vectors, or fingerprints, indexing their substructures such as labeled paths for molecules and n-grams for text, with the same Jaccard-Tanimoto similarity measure. As a result, similarity search methods from one field can be adapted to the other. Here we adapt recent, state-of-the-art, inverted index methods from information retrieval to speed up similarity searches in chemoinformatics. Our results show a several-fold speed-up improvement over previous methods for both threshold searches and top-K searches. We also provide a mathematical analysis that allows one to predict the level of pruning achieved by the inverted index approach and validate the quality of these predictions through simulation experiments. All results can be replicated using data freely downloadable from http://cdb.ics.uci.edu/ .

Learning to predict chemical reactions

Being able to predict the course of arbitrary chemical reactions is essential to the theory and applications of organic chemistry. Approaches to the reaction prediction problems can be organized around three poles corresponding to: (1) physical laws; (2) rule-based expert systems; and (3) inductive machine learning. Previous approaches at these poles, respectively, are not high throughput, are not generalizable or scalable, and lack sufficient data and structure to be implemented. We propose a new approach to reaction prediction utilizing elements from each pole. Using a physically inspired conceptualization, we describe single mechanistic reactions as interactions between coarse approximations of molecular orbitals (MOs) and use topological and physicochemical attributes as descriptors. Using an existing rule-based system (Reaction Explorer), we derive a restricted chemistry data set consisting of 1630 full multistep reactions with 2358 distinct starting materials and intermediates, associated with 2989 productive mechanistic steps and 6.14 million unproductive mechanistic steps. And from machine learning, we pose identifying productive mechanistic steps as a statistical ranking, information retrieval problem: given a set of reactants and a description of conditions, learn a ranking model over potential filled-to-unfilled MO interactions such that the top-ranked mechanistic steps yield the major products. The machine learning implementation follows a two-stage approach, in which we first train atom level reactivity filters to prune 94.00% of nonproductive reactions with a 0.01% error rate. Then, we train an ensemble of ranking models on pairs of interacting MOs to learn a relative productivity function over mechanistic steps in a given system. Without the use of explicit transformation patterns, the ensemble perfectly ranks the productive mechanism at the top 89.05% of the time, rising to 99.86% of the time when the top four are considered. Furthermore, the system is generalizable, making reasonable predictions over reactants and conditions which the rule-based expert does not handle. A web interface to the machine learning based mechanistic reaction predictor is accessible through our chemoinformatics portal ( http://cdb.ics.uci.edu) under the Toolkits section.

Tree and Hashing Data Structures to Speed up Chemical Searches: Analysis and Experiments

In many large chemoinformatics database systems, molecules are represented by long binary fingerprint vectors whose components record the presence or absence of particular functional groups or combinatorial features. For a given query molecule, one is interested in retrieving all the molecules in the database with a similarity to the query above a certain threshold. Here we describe a method for speeding up chemical searches in these large databases of small molecules by combining previously developed tree and hashing data structures to prune the search space without any false negatives. More importantly, we provide a mathematical analysis that allows one to predict the level of pruning, and validate the quality of the predictions of the method through simulation experiments.