MonteCat: A Basin-Hopping-Inspired Catalyst Descriptor Search Algorithm for Machine Learning Models

Proposing relevant catalyst descriptors that can relate the information on a catalyst's composition to its actual performance is an ongoing area in catalyst informatics, as it is a necessary step to improve our understanding on the target reactions. Herein, a small descriptor-engineered data set containing 3289 descriptor variables and the performance of 200 catalysts for the oxidative coupling of methane (OCM) is analyzed, and a descriptor search algorithm based on the workflow of the Basin-hopping optimization methodology is proposed to select the descriptors that better fit a predictive model. The algorithm, which can be considered wrapper in nature, consists of the successive generation of random-based modifications to the descriptor subset used in a regression model and adopting them depending on their effect on the model's score. The results are presented after being tested on linear and Support Vector Regression models with average cross-validation r2 scores of 0.8268 and 0.6875, respectively.

High-throughput experimentation for photocatalytic water purification in practical environments

High-throughput screening instrument was developed for photocatalytic water purification, enabling the simultaneous testing of 132 photocatalytic reactions under uniform visible light irradiation, temperature control, and stirring. The instrument was used to investigate the effects of different catalysts (TiO2, ZnO, α-Fe2O3) and environmental waters (seawater, urban wastewater, and industrial wastewater) on dye degradation. It was observed environmental ions, particularly carbonate and phosphate ions, significantly reduced catalyst activity by inhibiting the adsorption of dye molecules. To develop effective catalysts for dye degradation in industrial wastewater, 15 types of noble metal nanoparticles (NPs) were supported on photocatalysts. The study found that noble metal NPs with high work functions and oxidation resistance, such as Au and Pt, exhibited higher activity even in the industrial wastewater, likely converting environmental ions into active species. These findings, based on 432 test results, demonstrate the effectiveness of the developed high-throughput screening instrument for optimizing photocatalytic water purification.

Automatic feature engineering for catalyst design using small data without prior knowledge of target catalysis

The empirical aspect of descriptor design in catalyst informatics, particularly when confronted with limited data, necessitates adequate prior knowledge for delving into unknown territories, thus presenting a logical contradiction. This study introduces a technique for automatic feature engineering (AFE) that works on small catalyst datasets, without reliance on specific assumptions or pre-existing knowledge about the target catalysis when designing descriptors and building machine-learning models. This technique generates numerous features through mathematical operations on general physicochemical features of catalytic components and extracts relevant features for the desired catalysis, essentially screening numerous hypotheses on a machine. AFE yields reasonable regression results for three types of heterogeneous catalysis: oxidative coupling of methane (OCM), conversion of ethanol to butadiene, and three-way catalysis, where only the training set is swapped. Moreover, through the application of active learning that combines AFE and high-throughput experimentation for OCM, we successfully visualize the machine's process of acquiring precise recognition of the catalyst design. Thus, AFE is a versatile technique for data-driven catalysis research and a key step towards fully automated catalyst discoveries.

Parallel Catalyst Synthesis Protocol for Accelerating Heterogeneous Olefin Polymerization Research

The data scientific approach has become an indispensable tool for capturing structure-performance relationships in complex systems, where the quantity and quality of data play a crucial role. In heterogeneous olefin polymerization research, the exhaustive and multi-step nature of Ziegler-Natta catalyst synthesis has long posed a bottleneck in synthetic throughput and data generation. In this contribution, a custom-designed 12-parallel reactor system and a catalyst synthesis protocol were developed to achieve the parallel synthesis of a magnesium ethoxide-based Ziegler-Natta catalyst. The established system, featuring a miniature reaction vessel with magnetically suspended stirring, allows for over a tenfold reduction in synthetic scale while ensuring the consistency and reliability of the synthesis. We demonstrate that the established protocol is highly efficient for the generation of a catalyst library with diverse compositions and physical features, holding promise as a foundation for the data-driven establishment of the structure-performance relationship in heterogeneous olefin polymerization catalysis.

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.

Preventing Premature Convergence in Evolutionary Structure Determination of Complex Molecular Systems: Demonstration in Few-Nanometer-Sized TiCl4-Capped MgCl2 Nanoplates

The combination of genetic algorithm-based global search and local geometry optimization enables nonempirical structure determination for complex materials such as practical solid catalysts. However, premature convergence in the genetic algorithm hinders the determination of the global minimum for complicated molecular systems. Here, we implemented a distributed genetic algorithm based on the migration from a structure database for avoiding the premature convergence, and thus we realized the structure determination for TiCl₄-capped MgCl₂ nanoplates with experimentally consistent sizes. The obtained molecular models are featured with a realistic size and nonideal surfaces, representing actual primary particles of catalysts.

Enhancing Mechanical Properties of Graft-Type Nanocomposites Using Organically Modified SiO2 and Polypropylene Containing Reactive Methoxy Groups

In situ grafting of a reactive matrix and nanofillers is a promising strategy to fabricate graft-type polypropylene (PP)-based nanocomposites, where the grafting efficiency is affected by the initial dispersion of nanofillers in the matrix. In this work, influences of surface organic modification of nanofillers were investigated on properties of PP/SiO₂ nanocomposites using poly(propylene-co-octenyltrimethoxysilane) as a reactive matrix. The surface modification of SiO₂, especially with longer alkyl chains, led to improved dispersion of nanoparticles, thus promoting the grafting reaction and mechanical properties. The combination of in situ grafting and surface modification of nanofillers provided several benefits, most notably in balancing the strength and the toughness, which could not be achieved by the grafting alone.

High-Throughput Screening and Literature Data Driven Machine Learning Assisting Investigation of Multi-component La2O3-based Catalysts for Oxidative Coupling of Methane

Multi-component La2O3-based catalysts for oxidative coupling of methane (OCM) were designed based on high-throughput screening (HTS) and literature datasets with multi-output machine learning (ML) approaches including random forest regression (RFR),...

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.

Catalyst data created through high-throughput experimentation is transformed into catalyst knowledge networks, leading to a new method of catalyst design where successfully designed catalysts result in high C₂ yields during the OCM reaction.

Preparation and Desalination Performance of PA/UiO-66/PES Composite Membranes

UiO-66 nanoparticles are considered highly potential fillers for the application in desalination membranes. In this study, UiO-66 nanoparticles were anchored to PES membrane substrates, which were subsequently subjected to the interfacial polymerization reaction to coat a layer of polyamide (PA) on their surface. For comparison, a blank membrane incorporating no UiO-66 and a reference membrane incorporating ZrO2 (instead of UiO-66) were prepared. All prepared membranes were tested for their desalination performance. The membranes containing UiO-66 were found to outperform the blank and the reference counterparts. The reason for this outperformance is possibly attributed to the hydrophilicity of UiO-66 nanoparticles and the presence of nanochannels in their structure.

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.

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.

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.

Dataset of energetically accessible structures of MgCl2/TiCl4 clusters for Ziegler-Natta catalysts

This data article provides a dataset of the energetically accessible structures including the most stable structures of xMgCl₂/yTiCl₄ nanoplates (x = 6-19, y = 0-4). TiCl₄-capped MgCl₂ nanoplates are regarded as the building block of the Ziegler-Natta catalyst. The most stable structures were determined for MgCl₂/TiCl₄ nanoplates of different sizes and chemical compositions using a combination of the genetic algorithm and the DFT geometry optimization. The evolution in the genetic algorithm produced a number of meta-stable structures. A set of isomeric structures having similar energy to the most stable structure (termed energetically accessible structures) are provided as realistic models of MgCl₂/TiCl₄ nanoplates. These structures are useful for further investigation on the structural distribution of Ti species on MgCl₂ regarding the Ziegler-Natta catalyst.

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.

Modulator-free approach towards missing-cluster defect formation in Zr-based UiO-66

By rigorous control of water, missing-cluster defects in Zr-based UiO-66 were generated to a remarkable extent without the need of acidic modulators. The presence of missing-cluster defects created hierarchical pore structures, which had a profound effect on the catalytic performance.

Development of Large-Scale Stopped-Flow Technique and its Application in Elucidation of Initial Ziegler-Natta Olefin Polymerization Kinetics

The stopped-flow (SF) technique has been extensively applied to study Ziegler-Natta (ZN) olefin polymerization kinetics within an extremely short period (typically <0.2 s) for understanding the nature of the active sites as well as the polymerization mechanisms through microstructure analyses of obtained polymers. In spite of its great applicability, a small amount of polymer that is yielded in a short-time polymerization has been a major bottleneck for polymer characterizations. In order to overcome this limitation, a large-scale SF (LSF) system has been developed, which offers stable and scalable polymerization over an expanded time range from a few tens milliseconds to several seconds. The scalability of the LSF technique has been further improved by introducing a new quenching protocol. With these advantages, the LSF technique has been effectively applied to address several unknown issues in ZN catalysis, such as the role of physical and chemical transformations of a catalyst on the initial polymerization kinetics, and regiochemistry of ZN propylene polymerization. Here, we review the development of the LSF technique and recent efforts for understanding heterogeneous ZN olefin polymerization catalysis with this new system.

Nano-Dispersed Ziegler-Natta Catalysts for 1 μm-Sized Ultra-High Molecular Weight Polyethylene Particles

A catalytic approach to synthesize microfine ultra-high molecular weight polyethylene (UHMWPE) particles was proposed based on the exploitation of nano-sized catalysts. By utilizing MgO nanoparticles as a core material, a Ziegler-Natta-type MgO/MgCl2/TiCl4 core-shell catalyst with the particle size in a nano-range scale was prepared in a simple preparation step. The organic modification of MgO surfaces prior to catalyzation prevented agglomeration and facilitated the full dispersion of catalyst particles at a primary particle level for the first time. The nano-dispersed catalysts successfully afforded a direct access to UHMWPE having the particle size in the range of 1-2 μm at a reasonable activity. Extremely fine polymer particles yielded several advantages, especially at a significantly lower fusion temperature in compression molding.

High-Throughput Synthesis of Support Materials for Olefin Polymerization Catalyst

Rational catalyst design necessitates fundamental knowledge on the structure-performance relationship, while the synthetic throughput for heterogeneous Ziegler-Natta olefin polymerization catalysts has long prevented the acquisition of a statistical database. In this contribution, an in-house reactor system was developed to realize the parallel synthesis of support materials for Ziegler-Natta catalysts for the first time. The developed system enabled parallel synthesis of 24 magnesium ethoxide samples with excellent reproducibility and morphological control comparable to a conventional experiment. Our demonstration revealed that the generation of diverse particle characteristics could be achieved through the addition of a third component as a structural modulator, in which the in-house parallel reactor system combined with the first principle component analysis enabled fast screening of effective modulators.