Automated algorithms for band gap analysis from optical absorption spectra

Advancing high-throughput combinatorial aging studies of hybrid perovskite thin films via precise automated characterization methods and machine learning assisted analysis

To optimize material stability, automated high-throughput workflows are of increasing interest. However, many of those workflows either employ synthesis techniques not suitable for large-area depositions or are carried out in ambient conditions, which limits the transferability of the results. While combinatorial approaches based on vapour-based depositions are inherently scalable, their potential for controlled stability assessments has yet to be exploited. Based on MAPbI3 thin films as a prototypical system, we demonstrate a combinatorial inert-gas workflow to study intrinsic materials degradation, closely resembling conditions in encapsulated devices. Specifically, we probe the stability of MAPbI3 thin films with varying residual PbI2 content. A comprehensive set of automated characterization techniques is used to investigate the structure and phase constitution of pristine and aged thin films. A custom-designed in situ UV-Vis aging setup is used for real-time photospectroscopy measurements of the material libraries under relevant aging conditions, such as heat or light-bias exposure. These measurements are used to gain insights into the degradation kinetics, which can be linked to intrinsic degradation processes such as autocatalytic decomposition. Despite scattering effects, which complicate the conventional interpretation of in situ UV-Vis results, we demonstrate how a machine learning model trained on the comprehensive characterization data before and after the aging process can link changes in the optical spectra to phase changes during aging. Consequently, this approach does not only enable semi-quantitative comparisons of material stability but also provides detailed insights into the underlying degradation processes which are otherwise mostly reported for investigations on single samples.

In Situ UV-Vis-NIR Absorption Spectroscopy and Catalysis

This review highlights in situ UV-vis-NIR range absorption spectroscopy in catalysis. A variety of experimental techniques identifying reaction mechanisms, kinetics, and structural properties are discussed. Stopped flow techniques, use of laser pulses, and use of experimental perturbations are demonstrated for in situ studies of enzymatic, homogeneous, heterogeneous, and photocatalysis. They access different time scales and are applicable to different reaction systems and catalyst types. In photocatalysis, femto- and nanosecond resolved measurements through transient absorption are discussed for tracking excited states. UV-vis-NIR absorption spectroscopies for structural characterization are demonstrated especially for Cu and Fe exchanged zeolites and metalloenzymes. This requires combining different spectroscopies. Combining magnetic circular dichroism and resonance Raman spectroscopy is especially powerful. A multitude of phenomena can be tracked on transition metal catalysts on various supports, including changes in oxidation state, adsorptions, reactions, support interactions, surface plasmon resonances, and band gaps. Measurements of oxidation states, oxygen vacancies, and band gaps are shown on heterogeneous catalysts, especially for electrocatalysis. UV-vis-NIR absorption is burdened by broad absorption bands. Advanced analysis techniques enable the tracking of coking reactions on acid zeolites despite convoluted spectra. The value of UV-vis-NIR absorption spectroscopy to catalyst characterization and mechanistic investigation is clear but could be expanded.

Machine learning predicted emission of water-stable CdTe quantum dots

Quantum dots (QDs) have attracted much attention and exhibit many attractive properties, including high absorption coefficient, adjustable bandgap, high brightness, long-term stability, and size-dependent emission. It is known that to obtain high-quality luminescent properties (i.e. emission color, color purity, quantum yield, and stability), the synthesis parameters must be precisely controlled. In this work, we have constructed a database with CdTe aqueous synthesis parameters and spectroscopic results and applied machine learning algorithms to better understand the influence of the main synthesis parameters of CdTe QDs on their final emission properties. A strong dependence of the final emission wavelength with the reaction time and surface ligands and precursors concentrations was demonstrated. These parameters adjusted synchronously were shown to be very useful for provide ideal synthesis conditions for the preparation of CdTe QDs with desirable emission wavelengths. Moreover, applying the algorithms correctly allows for obtaining information and insights into the growth kinetics of QDs under different synthetic conditions.

Prospects and challenges for autonomous catalyst discovery viewed from an experimental perspective

Autonomous catalysis research requires elaborate integration of operando experiments into automated workflows. Suitable experimental data for analysis by artificial intelligence can be measured more readily according to standard operating procedures.

Research data infrastructure for high-throughput experimental materials science

The High-Throughput Experimental Materials Database (HTEM-DB, htem.nrel.gov) is a repository of inorganic thin-film materials data collected during combinatorial experiments at the National Renewable Energy Laboratory (NREL). This data asset is enabled by NREL's Research Data Infrastructure (RDI), a set of custom data tools that collect, process, and store experimental data and metadata. Here, we describe the experimental data flow from the RDI to the HTEM-DB to illustrate the strategies and best practices currently used for materials data at NREL. Integration of the data tools with experimental instruments establishes a data communication pipeline between experimental researchers and data scientists. This work motivates the creation of similar workflows at other institutions to aggregate valuable data and increase their usefulness for future machine learning studies. In turn, such data-driven studies can greatly accelerate the pace of discovery and design in the materials science domain.

•

Automated curation of experimental materials data

•

Integration of data tools into the experimental laboratory

•

Simple, effective, and flexible data archival system

•

Collection of metadata for enhanced total data value

For machine learning to make significant contributions to a scientific domain, algorithms must ingest and learn from high-quality, large-volume datasets. The Research Data Infrastructure (RDI) that feeds the High-Throughput Experimental Materials Database (HTEM-DB, htem.nrel.gov) provides such a dataset from existing experimental data streams at the National Renewable Energy Laboratory (NREL). The described methods for curating experimental data can be applied to other materials research laboratory settings, paving the way for increased application of machine learning to materials science. In turn, the resulting new materials and new knowledge will benefit the society by advancing new technologies in energy, fuels, computing, security, and other important areas.

A Generalized Semiempirical Approach to the Modeling of the Optical Band Gap of Ternary Al-(Ga, Nb, Ta, W) Oxides Containing Different Alumina Polymorphs

A generalization of the modeling equation of optical band gap values for ternary oxides, as a function of cationic ratio composition, is carried out based on the semiempirical correlation between the differences in the electronegativity of oxygen and the average cationic electronegativity proposed some years ago. In this work, a novel approach is suggested to account for the differences in the band gap values of the different polymorphs of binary oxides as well as for ternary oxides existing in different crystalline structures. A preliminary test on the validity of the proposed modeling equations has been carried out by using the numerous experimental data pertaining to alumina and gallia polymorphs as well as the crystalline ternary Ga(1-x)AlxO3 polymorphs (α-Ga(1-x)AlxO3 and β-Ga(1-x)AlxO3) covering a large range of optical band gap values (4.50-8.50 eV). To make a more rigorous test of the modeling equation, we extended our investigation to amorphous ternary oxides anodically formed on Al-d-metal alloys (Al-Nb, Al-Ta, and Al-W) covering a large range of d-metal composition (xd-metal ≥ 0.2). In the last case, the novel approach allows one to overcome some difficulties experienced in fitting the optical band gap dependence from the Al-d-metal mixed anodic oxide composition as well as to provide a rationale for the departure, at the lowest d-metal content (xd-metal < 0.2), from the behavior observed for anodic films containing higher d-metal content.

Progress and prospects for accelerating materials science with automated and autonomous workflows

Accelerating materials research by integrating automation with artificial intelligence is increasingly recognized as a grand scientific challenge to discover and develop materials for emerging and future technologies. While the solid state materials science community has demonstrated a broad range of high throughput methods and effectively leveraged computational techniques to accelerate individual research tasks, revolutionary acceleration of materials discovery has yet to be fully realized. This perspective review presents a framework and ontology to outline a materials experiment lifecycle and visualize materials discovery workflows, providing a context for mapping the realized levels of automation and the next generation of autonomous loops in terms of scientific and automation complexity. Expanding autonomous loops to encompass larger portions of complex workflows will require integration of a range of experimental techniques as well as automation of expert decisions, including subtle reasoning about data quality, responses to unexpected data, and model design. Recent demonstrations of workflows that integrate multiple techniques and include autonomous loops, combined with emerging advancements in artificial intelligence and high throughput experimentation, signal the imminence of a revolution in materials discovery.

Synthesis, optical imaging, and absorption spectroscopy data for 179072 metal oxides

Optical absorption spectroscopy is an important materials characterization for applications such as solar energy generation. This data descriptor describes the to date (Dec 2018) largest publicly available curated materials science dataset for near infrared to near UV (UV-Vis) light absorbance, composition and processing properties of metal oxides. By supplying the complete synthesis and processing history of each of the 179072 samples from 99965 unique compositions we believe the dataset will enable the community to develop predictive models for materials, such as prediction of optical properties based on composition and processing, and ultimately serve as a benchmark dataset for continued integration of machine learning in materials science. The dataset is also a resource for identifying materials composition and synthesis to attain specific optical properties.

Design Type(s)	classification objective • observation design
Measurement Type(s)	absorption spectrum
Technology Type(s)	ultraviolet-visible spectrophotometry
Factor Type(s)	metal oxide
Sample Characteristic(s)

Machine-accessible metadata file describing the reported data (ISA-Tab format)

Machine learning of optical properties of materials - predicting spectra from images and images from spectra

As the materials science community seeks to capitalize on recent advancements in computer science, the sparsity of well-labelled experimental data and limited throughput by which it can be generated have inhibited deployment of machine learning algorithms to date. Several successful examples in computational chemistry have inspired further adoption of machine learning algorithms, and in the present work we present autoencoding algorithms for measured optical properties of metal oxides, which can serve as an exemplar for the breadth and depth of data required for modern algorithms to learn the underlying structure of experimental materials science data. Our set of 178 994 distinct materials samples spans 78 distinct composition spaces, includes 45 elements, and contains more than 80 000 unique quinary oxide and 67 000 unique quaternary oxide compositions, making it the largest and most diverse experimental materials set utilized in machine learning studies. The extensive dataset enabled training and validation of 3 distinct models for mapping between sample images and absorption spectra, including a conditional variational autoencoder that generates images of hypothetical materials with tailored absorption properties. The absorption patterns auto-generated from sample images capture the salient features of ground truth spectra, and band gap energies extracted from these auto-generated patterns are quite accurate with a mean absolute error of 180 meV, which is the approximate uncertainty from traditional extraction of the band gap energy from measurements of the full transmission and reflection spectra. Optical properties of materials are not only ubiquitous in materials applications but also emblematic of the confluence of underlying physical phenomena yielding the type of complex data relationships that merit and benefit from neural network-type modelling.