Synthesis of Heterogeneous Catalysts in Catalyst Informatics to Bridge Experiment and High-Throughput Calculation.

The coupling of high-throughput calculations with catalyst informatics is proposed as an alternative way to design heterogeneous catalysts. High-throughput first-principles calculations for the oxidative coupling of methane (OCM) reaction are designed and performed where 1972 catalyst surface planes for the CH4 to CH3 reaction are calculated. Several catalysts for the OCM reaction are designed based on key elements that are unveiled via data visualization and network analysis. Among the designed catalysts, several active catalysts such as CoAg/TiO2, Mg/BaO, and Ti/BaO are found to result in high C2 yield. Results illustrate that designing catalysts using high-throughput calculations is achievable in principle if appropriate trends and patterns within the data generated via high-throughput calculations are identified. Thus, high-throughput calculations in combination with catalyst informatics offer a potential alternative method for catalyst design.

Unveiling the reaction pathways of hydrocarbons via experiments, computations and data science.

Reaction networks of hydrocarbons are explored using first principles calculations, data science, and experiments. Transforming hydrocarbon data into networks reveals the prevalence of the formation and reaction of various molecules. Graph theory is implemented to extract knowledge from the reaction network. In particular, centralities analysis reveals that H+, CCC, CH3+, CC, and [CH2+]C have high degrees and are thus very likely to form or react with other molecules. Additionally, H+, CH3+, C2H5+, C8H15+, C8H17+, and C6H11+ are found to have high control throughout the network and lead towards a series of additional reactions. The constructed network is also validated in experiments while the shortest path analysis is implemented for further comparison between experiment and the network. Thus, combining network analysis with first principles calculations uncovers key points in the development of various hydrocarbons that can be used to improve catalyst design and targeted synthesis of desired hydrocarbons.

Unveiling gas-phase oxidative coupling of methane via data analysis.

Unveiling the details of the mechanisms of a chemical reaction is a difficult task as reaction mechanisms are strongly coupled with reaction conditions. Here, catalysts informatics combined with high-throughput experimental data is implemented to understand the oxidative coupling of methane (OCM) reaction. In particular, pairwise correlation and data visualization are performed to reveal the relation between reaction conditions and selectivity/conversion. In addition, machine learning is used to fill the gap between experimental data points; thus, a more detailed understanding of the OCM reaction against reaction conditions can be achieved. Therefore, catalysts informatics is proposed for understanding the details of the reaction mechanism, thereby aiding reaction design.

Redesigning the Materials and Catalysts Database Construction Process Using Ontologies.

Materials and catalyst informatics are emerging fields that are a result of shifts in terms of how materials and catalysts are discovered in the fields of materials science and catalysis. However, these fields are not reaching their full potential due to issues related to database creation and curation. Issues such as lack of uniformity, data selectivity, and the presence of bias affect the quality and usefuless of materials databases, especially when attempting to search for materials descriptors. Without uniform rules and frameworks, databases are limited in use outside of the intent of the creators of the database. Ontologies are therefore investigated as a means of redesigning the way materials and catalysts databases are designed and created. In particular, an ontology consisting of information found within the periodic table as well as commonly used related data is constructed and applied toward the search for materials descriptors. Additional ontologies are also developed for two databases-a database consisting of computational data related to perovskites and a database consisting of experimental data related to oxidative coupling of methane (OCM) catalysts-in order to investigate the impact of merging ontologies.

Descriptors for predicting the lattice constant of body centered cubic crystal.

The prediction of the lattice constant of binary body centered cubic crystals is performed in terms of first principle calculations and machine learning. In particular, 1541 binary body centered cubic crystals are calculated using density functional theory. Results from first principle calculations, corresponding information from periodic table, and mathematically tailored data are stored as a dataset. Data mining reveals seven descriptors which are key to determining the lattice constant where the contribution of descriptors is also discussed and visualized. Support vector regression (SVR) technique is implemented to train the data where the predicted lattice constants have the mean score of 83.6% accuracy via cross-validation and maximum error of 4% when compared to experimentally determined lattice constants. In addition, trained SVR is successful in predicting material combinations from a desired lattice constant. Thus, a set of descriptors for determining the lattice constant is identified and can be used as a base descriptor for lattice constants of further complex crystals. This would allow for the acceleration of the search for lattice constants of desired atomic compositions as well as the prediction of new materials based on a specified lattice constant.

Reactivity of Two-Dimensional Au9, Pt9, and Au18Pt18 against Common Molecules.

Adsorption of common molecules over two-dimensional Au9, Pt9, and Au18Pt18 is investigated with implementation of first-principles calculations. In general, it is found that Pt9 and Au18Pt18 exhibit low adsorption energies where Au18Pt18 preserves the structural integrity of the molecule and surface. In particular, adsorption of molecules onto Au18Pt18 frequently results in low adsorption energies and high reactivity with minor surface reconstruction of Au18Pt18 and average bond lengths of molecules. The decrease in adsorption energy can be attributed to the presence of platinum, while gold can be considered responsible for structural stability. In addition, molecule dissociation is observed in the cases of H2, HCl, CH4, SO, and SO2 when Pt atoms are involved. Thus, two-dimensional Au9, Pt9, and Au18Pt18 show low adsorption energies against common molecules, reflecting adsorption energies observed in small Au and Pt clusters. These results demonstrate that Au18Pt18 can successfully utilize the low adsorption energies associated with platinum while preserving the integrity of the surface structure using gold atoms, making it possible to adsorb desired molecules using select areas of the Au18Pt18 surface.

Low temperature pollutant trapping and dissociation over two-dimensional tin.

Newly discovered two-dimensional tin, named stanene, has been theoretically predicted and found to have unique electronic properties. Stanene is found to have a buckled structure which could be a key structure against chemical reactivity. Hence, the reactivity of stanene against key air pollutants NO, NO2, SO, SO2, CO, and CO2 is investigated within first principles calculations. The results showed that stanene is reactive against those air pollutants. Furthermore, the dissociation activation energies of those pollutants over stanene are lower than previously reported catalysts. The physical origin of low dissociation barriers rests in the charge transfer from stanene to those pollutants, resulting in bond weakening. Hence, one can predict that unique reactivities of stanene offer low temperature trapping and dissociation of air pollutants.

Toward the Golden Age of Materials Informatics: Perspective and Opportunities.

Materials informatics is reaching the transition point and is evolving from its early stages of adoption and development and moving toward its golden age. Here, the transformation of the early stage of materials informatics toward the next level of materials informatics is explored. In particular, it has become crucial to be able to manipulate materials synthesis data, materials properties data, and materials characterization data. Through the use of ontology, material design and understanding can be carried out simultaneously in a whitebox manner. Here, a perspective on the ultimate goal of materials informatics along with potential key components is discussed.

Catalysts informatics: paradigm shift towards data-driven catalyst design.

Designing catalysts is a challenging matter as catalysts are involved with various factors that impact synthesis, catalysts, reactor and reaction. In order to overcome these difficulties, catalysts informatics is proposed as an alternative way to design and understand catalysts. The underlying concept of catalysts informatics is to design the catalysts from trends and patterns found in catalysts data. Here, three key concepts are introduced: experimental catalysts database, knowledge extraction from catalyst data via data science, and a catalysts informatics platform. Methane oxidation is chosen as a prototype reaction for demonstrating various aspects of catalysts informatics. This work summarizes how catalysts informatics plays a role in catalyst design. The work covers big data generation via high throughput experiments, machine learning, catalysts network method, catalyst design from small data, catalysts informatics platform, and the future of catalysts informatics via ontology. Thus, the proposed catalysts informatics would help innovate how catalysts can be designed and understood.

Designing transformer oil immersion cooling servers for machine learning and first principle calculations.

A transfomer oil immersion cooling server is designed and constructed for machine learning applications and first principle calculations that are carried out for materials-related research. CPU, motherboard, random access memory, hard disk drive, solid state drive, graphic card, and the power supply unit are submerged into the transformer oil in order to cool the entire system. Benchmark tests reveal that overall performance is improved while performance times for multicore calculations are dramatically improved. Furthermore, calculation times for machine learning with large data sets and density functional theory calculations are shortened during single core calculations. Thus, a transformer oil immersion cooling server is proposed to be an alternative cooling system used for improving the performance of first principle calculations and machine learning.

Representing the Methane Oxidation Reaction via Linking First-Principles Calculations and Experiment with Graph Theory.

Representing the chemical reaction is a challenging matter faced in chemistry due to the complex molecular interactions and difficulties faced when determining intermediate reactions that may occur throughout the reaction. Graph theory and network analysis are used with first-principles calculations and experiments to investigate possible intermediate reactions that may occur during a reaction; in this case, catalyst-free methane oxidation is chosen as the prototype reaction. Network analysis is used to help illuminate several key intermediate compounds that potentially appear throughout the course of the prototype reaction and the detailed mechanisms of methane oxidation while showing good agreement with experimental data. Presenting the chemical reaction as a network, therefore, makes it possible to link experimental and computational data in a space that accounts for the impact of intermediate reactions upon the outcome of the overall reaction, thereby making network analysis an alternative method for representing chemical reactions.

Representing Catalytic and Processing Space in Methane Oxidation Reaction via Multioutput Machine Learning.

Multioutput support vector regression (SVR) is implemented to simultaneously predict the selectivities and the CH4 conversion against experimental conditions in methane oxidation catalysts. The predictions unveil the details of how each selectivity and CH4 conversion behaves in each catalyst. In particular, the selectivity and the CH4 conversion of Mn-Na2WO4/SiO2, Ti-Na2WO4/SiO2, Pd-Na2WO4/SiO2, and Na2WO 4/SiO2 are predicted, and the effects of Mn, Ti, and Pd are unveiled. In addition, the trade-off points of CO and C2H6 are identified for each catalyst, leading to maximization of the C2H6 yield. Thus the simultaneous prediction of the reaction trend with catalysts not only will help with the understanding of the catalyst activities but also will provide guidance for designing the experimental conditions.

Constructing catalyst knowledge networks from catalyst big data in oxidative coupling of methane for designing catalysts.

Designing high performance catalysts for the oxidative coupling of methane (OCM) reaction is often hindered by inconsistent catalyst data, which often leads to difficulties in extracting information such as combinatorial effects of elements upon catalyst performance as well as difficulties in reaching yields beyond a particular threshold. In order to investigate C2 yields more systematically, high throughput experiments are conducted in an effort to mass-produce catalyst-related data in a way that provides more consistency and structure. Graph theory is applied in order to visualize underlying trends in the transformation of high-throughput data into networks, which are then used to design new catalysts that potentially result in high C2 yields during the OCM reaction. Transforming high-throughput data in this manner has resulted in a representation of catalyst data that is more intuitive to use and also has resulted in the successful design of a myriad of catalysts that elicit high C2 yields, several of which resulted in yields greater than those originally reported in the high-throughput data. Thus, transforming high-throughput catalytic data into catalyst design-friendly maps provides a new method of catalyst design that is more efficient and has a higher likelihood of resulting in high performance catalysts.

Catalysis Gene Expression Profiling: Sequencing and Designing Catalysts.

Identification of catalysts is a difficult matter as catalytic activities involve a vast number of complex features that each catalyst possesses. Here, catalysis gene expression profiling is proposed from unique features discovered in catalyst data collected by high-throughput experiments as an alternative way of representing the catalysts. Combining constructed catalyst gene sequences with hierarchical clustering results in catalyst gene expression profiling where natural language processing is used to identify similar catalysts based on edit distance. In addition, catalysts with similar properties are designed by modifying catalyst genes where the designed catalysts are experimentally confirmed to have catalytic activities that are associated with their catalyst gene sequences. Thus, the proposed method of catalyst gene expressions allows for a novel way of describing catalysts that allows for similarities in catalysts and catalytic activity to be easily recognized while enabling the ability to design new catalysts based on manipulating chemical elements of catalysts with similar catalyst gene sequences.

Data-Driven Identification of the Reaction Network in Oxidative Coupling of the Methane Reaction via Experimental Data.

Identifying details of chemical reactions is a challenging matter for both experiments and computations. Here, the reaction pathway in oxidative coupling of methane (OCM) is investigated using a series of experimental data and data science techniques in which data are analyzed using a variety of visualization techniques. Data visualization, pairwise correlation, and machine learning unveil the relationships between experimental conditions and the selectivities of CO, CO2, C2H4, C2H6, and H2 in the OCM reaction. More importantly, the reaction network for the OCM reaction is constructed on the basis of the scores provided by machine learning and experimental data. In particular, the proposed reaction map not only contains the chemical compound but also contains experimental conditions. Thus, data-driven identification of chemical reactions can be achieved in principle via a series of experimental data, leading to more efficient experimental design and catalyst development.

Multidimensional Classification of Catalysts in Oxidative Coupling of Methane through Machine Learning and High-Throughput Data.

Understanding the unique features of catalysts is a complex matter as it requires quantitative analysis with a relatively large selection of catalyst data. Here, unique features of each catalyst within the oxidative methane of coupling (OCM) reaction are investigated by combining data science and high throughput experimental data. Visualization of high-throughput OCM data reveals that there are several groups of catalysts based on their response against experimental conditions. Unsupervised machine learning, in particular, the Gaussian mixture model, classifies the OCM catalysts into six groups based on similarity in catalytic activities. Data visualization and parallel coordinates unveil the unique catalytic features of each classified group. Each classified group is statistically analyzed where unique features of each group are defined in term of C2 selectivity, CH4 conversion, and their composition in each calssified group. Thus, systematic design of catalysts can be achieved in principle on the basis of the unique features of catalysts uncovered via data science.

Visualizing Scientists' Cognitive Representation of Materials Data through the Application of Ontology.

The introduction of data science as a viable new approach to research has led toward the establishment of materials informatics. However, issues relating to the infrastructure of data collection and organization in materials science have hindered the development of materials informatics. Issues related to data quality, conflicting terminologies between subfields, and inconsistent recording practices make it difficult to share data and implement data science. Furthermore, one can consider that scientific discoveries have occurred via the rules that are unconsciously defined by the scientist's mind, which has made scientific discovery an unintentional process. Here, ontology is proposed as a new way to structure databases as well as model scientific understandings of data. By implementing ontology during the database creation process, it not only becomes possible to define and visualize the experiences and knowledge held by researchers but also provides a way of creating a field-wide standard of defining data, the ability to incorporate data semantics, a method to increase the solid choice of descriptors for determining the materials' properties, and the space to merge databases in a more interactive and coherent manner. Ontology can also help improve database management by providing a way to incorporate new scientific discoveries into existing databases, which can have a positive effect on the search for new materials and material design.

Data Driven Determination in Growth of Silver from Clusters to Nanoparticles and Bulk.

The thresholds among atomic clusters, nanoparticles, and the bulk state have been ambiguous. A potential solution is to determine cluster growth toward bulk, but this is challenging to determine with experiments and computation. Data science is proposed to predict atomic cluster growth and determine the cluster-nanoparticle-bulk thresholds using Ag clusters as a prototype element. Supervised machine learning reveals that Ag cluster growth has nonlinear models where nonlinear machine learning is found to accurately predict binding energy. Unsupervised machine learning discovers three groups (cluster, semiclusters, and nanoparticles) where linear regression is used to predict the binding energy in each group. Furthermore, machine learning reveals the linear relationship between binding energy and the surface-to-volume ratio of Ag nanoparticles. This allows for a binding energy estimation of large Ag nanoparticles and a revelation of how Ag nanoparticles grow toward the bulk. Thus, data science is proposed as a powerful tool for determining cluster growth and thresholds for clusters, nanoparticles, and bulk states.

Creating Machine Learning-Driven Material Recipes Based on Crystal Structure.

Determining the manner in which crystal structures are formed is considered a great mystery within materials science. Potential solutions have the possibility to be uncovered by revealing hidden patterns within the material data. Data science is therefore implemented in order to link the material data to the crystal structure. In particular, unsupervised and supervised machine learning techniques are used where the Gaussian mixture model is employed in order to understand the data structure of the materials database while random forest classification is used to predict the crystal structure. As a result, the unsupervised and supervised machine learning techniques reveal descriptors for determining the crystal structure via the materials database. In addition, predicting atomic combinations from the crystal structure is also achieved using a trained machine where the first-principles calculations confirm the stability of predicted materials. Thus, one can consider that the estimation of the crystal structure can be achieved in principle via the combination of material data and machine learning, thereby leading to the advancement of crystal structure prediction.

Structural stability and electronic properties of an octagonal allotrope of two dimensional boron nitride.

An octagonal allotrope of two dimensional boron nitride is explored through first principles calculations. Calculations show that two dimensional octagonal boron nitride can be formed with a binding energy comparable to two dimensional hexagonal boron nitride. In addition, two dimensional octagonal boron nitride is found to have a band gap smaller than two dimensional hexagonal boron nitride, suggesting the possibility of semiconductive attributes. Two dimensional octagonal boron nitride also has the ability to layer through physisorption. Defects present within two dimensional octagonal boron nitride also lead toward the introduction of a magnetic moment through the absence of boron atoms. The presence of defects is also found to render both hexagonal and octagonal boron nitrides reactive against hydrogen, where greater reactivity is seen in the presence of nitrogen. Thus, two dimensional octagonal boron nitride is confirmed with potential to tailor properties and reactivity through lattice shape and purposeful introduction of defects.