Statistical learning of materials properties or functions so far starts with a largely silent, nonchallenged step: the choice of the set of descriptive parameters (termed descriptor). However, when the scientific connection between the descriptor and the actuating mechanisms is unclear, the causality of the learned descriptor-property relation is uncertain. Thus, a trustful prediction of new promising materials, identification of anomalies, and scientific advancement are doubtful. We analyze this issue and define requirements for a suitable descriptor. For a classic example, the energy difference of zinc blende or wurtzite and rocksalt semiconductors, we demonstrate how a meaningful descriptor can be found systematically.

Big Data of Materials Science -- Critical Role of the Descriptor

Recent literature suggests that the fields of machine learning (ML) and high-throughput experimentation (HTE) have separately received considerable attention from chemists and engineers, leading to the development of powerful reactivity models and platforms capable of rapidly performing thousands of reactions. The merger of ML with HTE presents a wealth of opportunities for the exploration of chemical space, but the integration of the two has yet to be fully realized. We highlight examples of recent developments in ML and HTE that collectively suggest the utility of their integration. Our analysis highlights the complementarity of the two fields, while exposing a number of obstacles that can and should be overcome to take full advantage of this merger and thereby accelerate chemical research.

Toward Machine Learning-Enhanced High-Throughput Experimentation

The prosperity and lifestyle of our society are very much governed by achievements in condensed matter physics, chemistry and materials science, because new products for sectors such as energy, the environment, health, mobility and information technology (IT) rely largely on improved or even new materials. Examples include solid-state lighting, touchscreens, batteries, implants, drug delivery and many more. The enormous amount of research data produced every day in these fields represents a gold mine of the twenty-first century. This gold mine is, however, of little value if these data are not comprehensively characterized and made available. How can we refine this feedstock; that is, turn data into knowledge and value? For this, a FAIR (findable, accessible, interoperable and reusable) data infrastructure is a must. Only then can data be readily shared and explored using data analytics and artificial intelligence (AI) methods. Making data 'findable and AI ready' (a forward-looking interpretation of the acronym) will change the way in which science is carried out today. In this Perspective, we discuss how we can prepare to make this happen for the field of materials science. 

/pdf/fair-data-enabling-new-horizons-for-materials-research-2wgf2ras.pdf

FAIR data enabling new horizons for materials research

The surface oxygen vacancy formation energy (EOᵥₐc) is an important parameter in determining the catalytic activity of metal oxides. Estimating these energies can therefore lead to data-driven design of promising catalyst candidates. In the present study, we determine EOᵥₐc for various insulating and semiconducting oxides. Statistical investigations indicate that the band gap, bulk formation energy, and electron affinity are factors that strongly influence EOᵥₐc. Electrons enter defect states after O desorption, and these states can be in the valence band, mid-gap, or in the conduction band. Subsequent adsorption of O₂, NO, CO, CO₂, and H₂ molecules on an O-deficient surface is also investigated. These molecules become preferentially adsorbed at the defect sites, and EOᵥₐc is identified as the dominant factor that determines the adsorption mode as well as a descriptor that shows good correlation with the adsorption energy.

Density Functional Theory Calculations of Oxygen Vacancy Formation and Subsequent Molecular Adsorption on Oxide Surfaces

Modern research methods produce large amounts of scientifically valuable data. Tools to process and analyze such data have advanced rapidly. Yet, access to large amounts of high‐quality data remains limited in many fields, including catalysis research. Implementing the concept of FAIR data (Findable, Accessible, Interoperable, Reusable) in the catalysis community would improve this situation dramatically. The German NFDI initiative (National Research Data Infrastructure) aims to create a unique research data infrastructure covering all scientific disciplines. One of the consortia, NFDI4Cat, proposes a concept that serves all aspects and fields of catalysis research. We present a perspective on the challenging path ahead. Starting out from the current state, research needs are identified. A vision for a integrating all research data along the catalysis value chain, from molecule to chemical process, is developed. Respective core development topics are discussed, including ontologies, metadata, required infrastructure, IP, and the embedding into research community. This Concept paper aims to inspire not only researchers in the catalysis field, but to spark similar efforts also in other disciplines and on an international level.

/pdf/a-unified-research-data-infrastructure-for-catalysis-1idtjuikie.pdf

A Unified Research Data Infrastructure for Catalysis Research – Challenges and Concepts

The “Seven Pillars” of oxidation catalysis proposed by Robert K. Grasselli represent an early example of phenomenological descriptors in the field of heterogeneous catalysis. Major advances in the theoretical description of catalytic reactions have been achieved in recent years and new catalysts are predicted today by using computational methods. To tackle the immense complexity of high-performance systems in reactions where selectivity is a major issue, analysis of scientific data by artificial intelligence and data science provides new opportunities for achieving improved understanding. Modern data analytics require data of highest quality and sufficient diversity. Existing data, however, frequently do not comply with these constraints. Therefore, new concepts of data generation and management are needed. Herein we present a basic approach in defining best practice procedures of measuring consistent data sets in heterogeneous catalysis using “handbooks”. Selective oxidation of short-chain alkanes over mixed metal oxide catalysts was selected as an example.

/pdf/towards-experimental-handbooks-in-catalysis-r03vwlq4yy.pdf

Giulia Bellini

Papers

Towards Experimental Handbooks in Catalysis