Search NASA⌕ Search

Engineering topics

Li, Yi-Pei

Publications and source records attributed to Li, Yi-Pei.

9 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2019

Self-Evolving Machine: A Continuously Improving Model for Molecular Thermochemistry

Because collecting precise and accurate chemistry data is often challenging, chemistry data sets usually only span a small region of chemical space, which limits the performance and the scope of applicability of data-driven models. To address this issue, we integrated an active learning machine with automatic ab initio calculations to form a self-evolving model that can continuously adapt to new species appointed by the users. In the present work, we demonstrate the self-evolving concept by modeling the formation enthalpies of stable closed-shell polycyclic species calculated at the B3LYP/6-31G(2df,p) level of theory. By combining a molecular graph convolutional neural network with a dropout training strategy, the model we developed can predict density functional theory (DFT) enthalpies for a broad range of polycyclic species and assess the quality of each predicted value. For the species which the current model is uncertain about, the automatic ab initio calculations provide additional training data to improve the performance of the model. For a test set composed of 2858 cyclic and polycyclic hydrocarbons and oxygenates, the enthalpies predicted by the model agree with the reference DFT values with a root-mean-square error of 2.62 kcal/mol. Finally, we found that a model originally trained on hydrocarbons and oxygenates can broaden its prediction coverage to nitrogen-containing species via an active learning process, suggesting that the continuous learning strategy is not only able to improve the model accuracy but is also capable of expanding the predictive capacity of a model to unseen species domains.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2019

Automated Computational Thermochemistry for Butane Oxidation: A Prelude to Predictive Automated Combustion Kinetics

arge-scale implementation of high level computational theoretical chemical kinetics offers the prospect for dramatically improving the fidelity of combustion chemical modeling. As a first step toward this goal, we developed a procedure for automatically generating the thermochemical data for combustion of an arbitrary fuel. The procedure begins by producing a list of combustion relevant species from a specification of the fuel and combustion conditions of interest. Then, for each element in the list of species, the procedure determines an internal coordinate z-matrix description of its structure, the optimal torsional configuration via Monte Carlo sampling, key rovibrational properties for that optimal geometry (including anharmonic corrections from torsional mappings and/or vibrational perturbation theory), and high level estimates of the electronic and zero-point energies via arbitrarily defined composite methods. This dataset is then converted first to partition functions, then to thermodynamic properties, and finally to NASA polynomial representations of the data. The end product is an automatically generated database of electronic structure results and thermochemical data including representations in a format appropriate for combustion simulations. The utility and functioning of this predictive automated computational thermochemistry (PACT) software package is illustrated through application to the automated generation of thermochemical data for the combustion of n-butane. Butane is chosen for this demonstration as its species list is of reasonably manageable size for debugging level computations, while still presenting most of the key challenges that need to be surmounted in the consideration of larger fuels. Furthermore, its low temperature chemistry is representative of that oc-curring with larger alkanes.

Keceli, Murat↗
DOE OSTI2016

Propene Metathesis over Supported Tungsten Oxide Catalysts: A Study of Active Site Formation

A detailed investigation was conducted on the factors influencing the properties of silica-supported tungsten oxide catalysts for propene metathesis. A principal goal of this work was to identify the processes involved in the formation of catalytically active sites. To probe the influence of dispersion, samples were prepared across a range of W loadings using two methods of catalyst preparation: incipient wetness impregnation of amorphous silica and ion exchange of mesoporous SBA-15. The samples were characterized by nitrogen adsorption, UV-vis, Raman, and X-ray absorption spectroscopy (XAS). Catalytic activity was observed to increase with W surface concentration up to the point where WO 3 nanoparticles formed. The catalytic performance of all samples was enhanced 2-fold by pretreatment in He, in comparison to pretreatment in air. In situ characterization of samples pretreated in He by Raman and XAS shows an increase in the relative concentration of isolated dioxo W(6+) species relative to mono-oxo W(6+) species, and in situ XAS data collected during propene metathesis indicated that a similar conversion occurs for air-pretreated samples in the presence of propene. For both air- and He-pretreated catalysts an activation period was observed, during which the activity increased and attained steady-state activity. This period was significantly longer for air-pretreated catalysts and was accompanied by the transient formation of acetone. While acetone was not observed during the much shorter transient of He-pretreated samples, in situ XAS provided evidence of reduction occurring in these samples upon contact with propene. In conclusion, it is also notable that, independent of the manner of catalyst preparation or pretreatment, the rate of propene metathesis is first order in propene and exhibits an activation energy of 200 kJ/mol. A model is proposed to explain why only a fraction of the isolated tungstate species is active for propene metathesis (~5%) and why this fraction increases with increasing concentration of W dispersed on silica.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2016

Theoretical Study of 4-(Hydroxymethyl)benzoic Acid Synthesis from Ethylene and 5-(Hydroxymethyl)furoic Acid Catalyzed by Sn-BEA

A sustainable route has been confirmed for the production of terephthalic acid (PTA) from 5-(hydroxymethyl)furoic acid (HMFA) and ethylene, both of which can be derived from biomass. This process starts with the production of 4-(hydroxymethyl)benzoic acid (HMBA) from HMFA and ethylene catalyzed by Sn-BEA. The subsequent oxidation of HMBA leads to PTA. Our research introduces the results of a detailed computational investigation of the mechanism of HMBA synthesis from ethylene and HMFA mediated by Sn-BEA. Density functional theory calculations show that the formation of HMBA proceeds via Diels-Alder cycloaddition of HMFA and ethylene, which is rate-limiting, followed by Lewis acid-catalyzed dehydration. The solution-phase reaction and six different pathways in Sn-BEA, including one pathway on the Si site and five different pathways on the Sn site, are investigated for the Diels-Alder cycloaddition of HMFA and ethylene. Energy decomposition analysis (EDA) reflects that the Sn site stabilizes the transition state of the Diels-Alder reaction electrostatically instead of facilitating charge transfer between HMFA and ethylene. Therefore, the preferred pathway for the Diels-Alder reaction starts with binding HMFA to the Sn site by the carbonyl oxygen, which is the configuration that maximizes electrostatic interactions between substrates and the catalyst in the transition state. The effect of substituting Sn in the active site by Zr and Ti was examined and the highest reaction barriers were for the Ti sites. Using EDA, we found that though the barriers of the Sn and Zr site are comparable, the individual contributing effects are different: lower energy penalty associated with distortion of the geometry of the Zr site overcomes less favorable electrostatic and charge transfer effects compared to the Sn site.

catalysis↗
DOE OSTI2016

Thermodynamics of Anharmonic Systems: Uncoupled Mode Approximations for Molecules

The partition functions, heat capacities, entropies, and enthalpies of selected molecules were calculated using uncoupled mode (UM) approximations, where the full-dimensional potential energy surface for internal motions was modeled as a sum of independent one-dimensional potentials for each mode. The computational cost of such approaches scales the same with molecular size as standard harmonic oscillator vibrational analysis using harmonic frequencies (HO hf ). To compute thermodynamic properties, a computational protocol for obtaining the energy levels of each mode was established. The accuracy of the UM approximation depends strongly on how the one-dimensional potentials of each modes are defined. If the potentials are determined by the energy as a function of displacement along each normal mode (UM-N), the accuracies of the calculated thermodynamic properties are not significantly improved versus the HO hf model. Significant improvements can be achieved by constructing potentials for internal rotations and vibrations using the energy surfaces along the torsional coordinates and the remaining vibrational normal modes, respectively (UM-VT). For hydrogen peroxide and its isotopologs at 300 K, UM-VT captures more than 70% of the partition functions on average. By con trast, the HO hf model and UM-N can capture no more than 50%. For a selected test set of C2 to C8 linear and branched alkanes and species with different moieties, the enthalpies calculated using the HO hf model, UM-N, and UM-VT are all quite accurate comparing with reference values though the RMS errors of the HO model and UM-N are slightly higher than UM-VT. However, the accuracies in entropy calculations differ significantly between these three models. For the same test set, the RMS error of the standard entropies calculated by UM-VT is 2.18 cal mol -1 K -1 at 1000 K. By contrast, the RMS error obtained using the HO model and UM-N are 6.42 and 5.73 cal mol -1 K -1 , respectively. For a test set composed of nine alkanes ranging from C5 to C8, the heat capacities calculated with the UM-VT model agree with the experimental values to within a RMS error of 0.78 cal mol -1 K -1 , which is less than one-third of the RMS error of the HO hf (2.69 cal mol -1 K -1) and UM-N (2.41 cal mol -1 K -1) models.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2014

Improved Force-Field Parameters for QM/MM Simulations of the Energies of Adsorption for Molecules in Zeolites and a Free Rotor Correction to the Rigid Rotor Harmonic Oscillator Model for Adsorption Enthalpies

Quantum mechanics/molecular mechanics (QM/MM) simulations provide an efficient avenue for studying reactions catalyzed in zeolite systems; however, the accuracy of such calculations is highly dependent on the zeolite MM parameters used. Previously reported parameters (P1), which were chosen to minimize the root mean square (RMS) deviations of adsorption energies compared with full QM ωB97X-D/6-31+G** adsorption energies, are shown to overestimate binding energies compared with experimental values, particularly for larger substrates. To address this issue, a new parameter set (P2) is derived by rescaling the previously reported characteristic energies of the Lennard-Jones potential in P1. The accuracy of the thermal correction for adsorption enthalpies determined by the rigid rotor-harmonic oscillator approximation (RRHO) is examined and shown to be improved by treating low-lying vibrational modes as free translational and rotational modes via a quasi-RRHO model. Here, with P2 and quasi-RRHO, adsorption energies calculated with QM/MM agree with experimental values with an RMS error of 1.8 kcal/mol for both nonpolar and polar molecules adsorbed in MFI, H-MFI, and H-BEA. By contrast, the RMS error for the same test sets obtained using parameter set P1 is 8.3 kcal/mol. Glucose-fructose isomerization catalyzed by Sn-BEA is taken as an example to demonstrate that improved values for apparent activation energies can be obtained using the methodology reported here. With parameter set P2, the apparent activation energy calculated with QM/MM reproduces the experimental value to within 1 kcal/mol. By contrast, using parameter set P1, the error is -12.9 kcal/mol. (Graph Presented).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2014

Computational Study of p -Xylene Synthesis from Ethylene and 2,5-Dimethylfuran Catalyzed by H-BEA

Detailed mechanisms for the synthesis of p -xylene as well as the primary byproducts observed experimentally, 2,5-hexadione and 2,5-dimethyl-3-[(4-methyl-1,3-cyclohexadien-1-yl)methyl]furan, from ethylene and 2,5-dimethylfuran (DMF) mediated by H-BEA are obtained using an extended QM/MM model containing 208 tetrahedral atoms. The formation of p -xylene proceeds via Diels-Alder cycloaddition of ethylene and DMF, which is rate-limiting, followed by Brønsted acid-catalyzed dehydration. Secondary addition of DMF to the substrate following the Diels-Alder reaction leads to 2,5-dimethyl-3-[(4-methyl-1,3-cyclohexadien-1-yl)methyl]furan. The analysis of the free energies associated with the mechanisms suggests that the secondary addition can be eliminated by introducing n -heptane as an inert solvent to decrease the loading of DMF in the zeolite or by using a weak Brønsted acid site to facilitate the dehydration of the Diels-Alder product, for which the rate is determined by the deprotonation via the conjugate base of the active site. Water formed in the dehydration process can react directly with DMF to form 2,5-hexadione, thereby decreasing the yield of p -xylene. However, the free-energy barriers for the formation of 2,5-heaxdione compared to the Diels-Alder reaction indicate that DMF and 2,5-hexadione will be equilibrated. Finally, the 2,5-hexadione yield can be minimized by operating at a high conversion of DMF.

Li, Yi-Pei↗
DOE OSTI2014

Analysis of the Reaction Mechanism and Catalytic Activity of Metal-Substituted Beta Zeolite for the Isomerization of Glucose to Fructose

Glucose-fructose isomerization mediated by Sn-BEA is investigated using an extended QM/MM model containing 208 tetrahedral atoms. The isomerization mechanism consists of a sequence of ring-opening, isomerization, and ring-closing processes, consistent with the previously reported experimental observations. In agreement with the experimentally observed kinetic isotope effect, the rate-determining step is found to involve a hydride shift from the C 2 carbon to the C 1 carbon. The apparent activation energy for the rate-limiting step is 22.3 kcal/mol at 343 K. The difference in the reaction barriers for the partially hydrolyzed and the fully coordinated Sn sites was investigated using energy decomposition analysis. It is found that the higher activity of the partially hydrolyzed site comes from the extra flexibility provided by the defect in the lattice. The effect of substituting Sn in the active site by Ti, Zr, V, Nb, Si, and Ge was examined, and it was found that Sn and Zr are metals that result in the lowest reaction barrier for glucose isomerization. By using energy decomposition analysis, two physical properties are shown to contribute to the magnitude of the reaction barrier: the polarizability of the metal atom in the active site and the Brønsted basicity of the oxygen atom bound to the metal atom.

anions↗