Artificial neural network prediction of self-diffusion in pure compounds over multiple phase regimes.

Artificial neural networks (ANNs) were developed to accurately predict the self-diffusion constants for pure components in liquid, gas and super critical phases. The ANNs were tested on an experimental database of 6625 self-diffusion constants for 118 different chemical compounds. The presence of multiple phases results in a heavy skew in the distribution of diffusion constants and multiple approaches were used to address this challenge. First, an ANN was developed with the raw diffusion values to assess what the main drawbacks of this direct method were. The first approach for improving the predictions involved taking the log 10 of diffusion to provide a more uniform distribution and reduce the range of target output values used to develop the ANN. The second approach involved developing individual ANNs for each phase using the raw diffusion values. Results show that the log transformation leads to a model with the best self-diffusion constant predictions and an overall average absolute deviation (AAD) of 6.56%. The resultant ANN is a generalized model that can be used to predict diffusion across all three phases and over a diverse group of compounds. The importance of each input feature was ranked using a feature addition method revealing that the density of the compound has the largest impact on the ANN prediction of self-diffusion constants in pure compounds.

Selective electrochemical oxidative coupling of methane mediated by Sr2Fe1.5Mo0.5O6-δ and its chemical stability.

Efficient conversion of methane to value-added products such as olefins and aromatics has been in pursuit for the past few decades. The demand has increased further due to the recent discoveries of shale gas reserves. Oxidative and non-oxidative coupling of methane (OCM and NOCM) have been actively researched, although catalysts with commercially viable conversion rates are not yet available. Recently, [Formula: see text] (SFMO-075Fe) has been reported to activate methane in an electrochemical OCM (EC-OCM) set up with a C2 selectivity of 82.2%1. However, alkaline earth metal-based materials are known to suffer chemical instability in carbon-rich environments. Hence, here we evaluated the chemical stability of SFMO in carbon-rich conditions with varying oxygen concentrations at temperatures relevant for EC-OCM. SFMO-075Fe showed good methane activation properties especially at low overpotentials but suffered poor chemical stability as observed via thermogravimetric, powder XRD, and XPS measurements where SrCO3 was observed to be a major decomposition product along with SrMoO3 and MoC. Nevertheless, our study demonstrates that electrochemical methods could be used to selectively activate methane towards partial oxidation products such as ethylene at low overpotentials while higher applied biases result in the complete oxidation of methane to carbon dioxide and water.

Machine learning prediction of self-diffusion in Lennard-Jones fluids.

Different machine learning (ML) methods were explored for the prediction of self-diffusion in Lennard-Jones (LJ) fluids. Using a database of diffusion constants obtained from the molecular dynamics simulation literature, multiple Random Forest (RF) and Artificial Neural Net (ANN) regression models were developed and characterized. The role and improved performance of feature engineering coupled to the RF model development was also addressed. The performance of these different ML models was evaluated by comparing the prediction error to an existing empirical relationship used to describe LJ fluid diffusion. It was found that the ANN regression models provided superior prediction of diffusion in comparison to the existing empirical relationships.

Nitrogen electroreduction and hydrogen evolution on cubic molybdenum carbide: a density functional study.

We report herein a density functional theory study of the nitrogen electroreduction and hydrogen evolution reactions on cubic molybdenum carbide (MoC) in order to investigate the viability of using this material as an electro-catalyst for ammonia synthesis. Free energy diagrams for associative and dissociative Heyrovsky mechanisms showed that nitrogen reduction on cubic MoC(111) can proceed via an associative mechanism and that small negative potentials of -0.3 V vs. standard hydrogen electrode can onset the reduction of nitrogen to ammonia. Kinetic volcano plots for hydrogen evolution showed that the MoC[110] surface is expected to have a high rate for the hydrogen evolution reaction, which could compete with the reduction of nitrogen on cubic MoC. The comparison between the adsorption energies of H-adatoms and N-adatoms also shows that at low potentials adsorption of hydrogen atoms competes with nitrogen adsorption on all the MoC surfaces except the MoC(111) surface. The hydrogen evolution and accumulation of H-adatoms can be mitigated by introducing carbon vacancies i.e. increasing the ratio of metal to carbon atoms, which will significantly increase the affinity of the catalytic surface for both nitrogen molecules and N-adatoms.

Applicability of density functional theory in reproducing accurate vibrational spectra of surface bound species.

The structural equilibrium parameters, the adsorption energies, and the vibrational frequencies of the nitrogen molecule and the hydrogen atom adsorbed on the (111) surface of rhodium have been investigated using different generalized-gradient approximation (GGA), nonlocal correlation, meta-GGA, and hybrid functionals, namely, Perdew, Burke, and Ernzerhof (PBE), Revised-RPBE, vdW-DF, Tao, Perdew, Staroverov, and Scuseria functional (TPSS), and Heyd, Scuseria, and Ernzerhof (HSE06) functional in the plane wave formalism. Among the five tested functionals, nonlocal vdW-DF and meta-GGA TPSS functionals are most successful in describing energetics of dinitrogen physisorption to the Rh(111) surface, while the PBE functional provides the correct chemisorption energy for the hydrogen atom. It was also found that TPSS functional produces the best vibrational spectra of the nitrogen molecule and the hydrogen atom on rhodium within the harmonic formalism with the error of -2.62 and -1.1% for the N-N stretching and Rh-H stretching frequency. Thus, TPSS functional was proposed as a method of choice for obtaining vibrational spectra of low weight adsorbates on metallic surfaces within the harmonic approximation. At the anharmonic level, by decoupling the Rh-H and N-N stretching modes from the bulk phonons and by solving one- and two-dimensional Schrödinger equation associated with the Rh-H, Rh-N, and N-N potential energy we calculated the anharmonic correction for N-N and Rh-H stretching modes as -31 cm(-1) and -77 cm(-1) at PBE level. Anharmonic vibrational frequencies calculated with the use of the hybrid HSE06 function are in best agreement with available experiments.

Electro-reduction of nitrogen on molybdenum nitride: structure, energetics, and vibrational spectra from DFT.

We used density functional theory to study the electrochemical conversion of nitrogen to ammonia on the (001), (100/010), (101), and (111) surfaces of γ-Mo2N. Based on the calculated free energy profiles for the reduction of nitrogen by the associative and dissociative mechanisms, reactivity was found to decrease in the order (111) > (101) > (100/010) ≈ (001). Namely, the cell potentials needed to drive the reduction to ammonia increase in the following order: -0.7 V on (111), -1.2 V on (101), and -1.4 V on (100/010) and (001) surfaces. The (111) surface was found to be the most reactive for nitrogen due to (i) its ability to adsorb the N2 in the side-on position which activates N-N bonding and (ii) its high affinity for N-adatoms which also prevents accumulation of H-adatoms on the catalytic surface at low cell potentials. We have also calculated vibrational frequencies of different NxHy species adsorbed on various γ-Mo2N surfaces. The frequencies are found to depend strongly on the type of the binding sites available on the crystal facets. A comparison of the calculated frequencies with the frequencies of the corresponding species in transition metal complexes and other metal surfaces shows that the frequencies of several signature modes fall in a similar region and might be used to assign the spectra of hydrogen and nitrogen containing surface species on different metal surfaces.

Trace detection and discrimination of explosives using electrochemical potentiometric gas sensors.

In this article, selective and sensitive detection of trace amounts of pentaerythritol tetranitrate (PETN), 2,4,6-trinitrotoluene (TNT) and cyclotrimethylenetrinitramine (RDX) is demonstrated. The screening system is based on a sampling/concentrator front end and electrochemical potentiometric gas sensors as the detector. Preferential hydrocarbon and nitrogen oxide(s) mixed potential sensors based on lanthanum strontium chromite and Pt electrodes with yttria stabilized zirconia (YSZ) solid electrolyte were used to capture the signature of the explosives. Quantitative measurements based on hydrocarbon and nitrogen oxide sensor responses indicated that the detector sensitivity scaled proportionally with the mass of the explosives (1-3 µg). Moreover, the results showed that PETN, TNT, and RDX samples could be discriminated from each other by calculating the ratio of nitrogen oxides to hydrocarbon integrated area under the peak. Further, the use of front-end technology to collect and concentrate the high explosive (HE) vapors make intrinsically low vapor pressure of the HE less of an obstacle for detection while ensuring higher sensitivity levels. In addition, the ability to use multiple sensors each tuned to basic chemical structures (e.g., nitro, amino, peroxide, and hydrocarbon groups) in HE materials will permit the construction of low-cost detector systems for screening a wide spectrum of explosives with lower false positives than present-day technologies.

Titanium dioxide-supported non-precious metal oxygen reduction electrocatalyst.

A new non-precious metal oxygen reduction catalyst was developed via heat treatment of in situ polymerized polyaniline onto TiO(2) particles in the presence of Fe species. The TiO(2) provides for improved performance relative to a carbon black-based catalyst and, at a high catalyst loading, allows for reducing the performance gap between non-precious-metal catalyst and Pt/C to ca. 20 mV in RDE testing.