Deep data mining in a real space: separation of intertwined electronic responses in a lightly doped BaFe2As2

Machine-Learning Predictions of Critical Temperatures from Chemical Compositions of Superconductors

In the quest for advanced superconducting materials, the accurate prediction of critical temperatures (Tc) poses a formidable challenge, largely due to the complex interdependencies between superconducting properties and the chemical and structural characteristics of a given material. To address this challenges, we have developed a machine-learning framework that aims to elucidate these complicated and hitherto poorly understood structure-property and property-property relationships. This study introduces a novel machine-learning-based workflow, termed the Gradient Boosted Feature Selection (GBFS), which has been tailored to predict Tc for superconductors by employing a distributed gradient-boosting framework. This approach integrates exploratory data analyses, statistical evaluations, and multicollinearity reduction techniques to select highly relevant features from a high-dimensional feature space, derived solely from the chemical composition of materials. Our methodology was rigorously tested on a data set comprising approximately 16,400 chemical compounds with around 12,000 unique chemical compositions. The GBFS workflow enabled the development of a classification model that distinguishes compositions likely to exhibit Tc values greater than 10 K. This model achieved a weighted average F1-score of 0.912, an AUC-ROC of 0.986, and an average precision score of 0.919. Additionally, the GBFS workflow underpinned a regression model that predicted Tc values with an R2 of 0.945, an MAE of 3.54 K, and an RMSE of 6.57 K on a test set obtained via random splitting. Further exploration was conducted through out-of-sample Tc predictions, particularly those exceeding the liquid nitrogen temperature, and out-of-distribution predictions for (Ca1-xLax)FeAs2 based on varying lanthanum content. The outcome of our study underscores the significance of systematic feature analysis and selection in enhancing predictive model performance, offering various advantages over models that rely primarily on algorithmic complexity. This research not only advances the field of superconductivity but also sets a precedent for the application of machine learning in materials science.

Automated and Autonomous Experiments in Electron and Scanning Probe Microscopy

Machine learning and artificial intelligence (ML/AI) are rapidly becoming an indispensable part of physics research, with domain applications ranging from theory and materials prediction to high-throughput data analysis. In parallel, the recent successes in applying ML/AI methods for autonomous systems from robotics to self-driving cars to organic and inorganic synthesis are generating enthusiasm for the potential of these techniques to enable automated and autonomous experiments (AE) in imaging. Here, we aim to analyze the major pathways toward AE in imaging methods with sequential image formation mechanisms, focusing on scanning probe microscopy (SPM) and (scanning) transmission electron microscopy ((S)TEM). We argue that automated experiments should necessarily be discussed in a broader context of the general domain knowledge that both informs the experiment and is increased as the result of the experiment. As such, this analysis should explore the human and ML/AI roles prior to and during the experiment and consider the latencies, biases, and prior knowledge of the decision-making process. Similarly, such discussion should include the limitations of the existing imaging systems, including intrinsic latencies, non-idealities, and drifts comprising both correctable and stochastic components. We further pose that the role of the AE in microscopy is not the exclusion of human operators (as is the case for autonomous driving), but rather automation of routine operations such as microscope tuning, etc., prior to the experiment, and conversion of low latency decision making processes on the time scale spanning from image acquisition to human-level high-order experiment planning. Overall, we argue that ML/AI can dramatically alter the (S)TEM and SPM fields; however, this process is likely to be highly nontrivial and initiated by combined human-ML workflows and will bring challenges both from the microscope and ML/AI sides. At the same time, these methods will enable opportunities and paradigms for scientific discovery and nanostructure fabrication.

Unsupervised machine learning of topological phase transitions from experimental data

Identifying phase transitions is one of the key challenges in quantum many-body physics. Recently, machine learning methods have been shown to be an alternative way of localising phase boundaries from noisy and imperfect data without the knowledge of the order parameter. Here, we apply different unsupervised machine learning techniques, including anomaly detection and influence functions, to experimental data from ultracold atoms. In this way, we obtain the topological phase diagram of the Haldane model in a completely unbiased fashion. We show that these methods can successfully be applied to experimental data at finite temperatures and to the data of Floquet systems when post-processing the data to a single micromotion phase. Our work provides a benchmark for the unsupervised detection of new exotic phases in complex many-body systems.

K-means-driven Gaussian Process data collection for angle-resolved photoemission spectroscopy

We propose the combination of k-means clustering with Gaussian Process (GP) regression in the analysis and exploration of 4D angle-resolved photoemission spectroscopy (ARPES) data. Using cluster labels as the driving metric on which the GP is trained, this method allows us to reconstruct the experimental phase diagram from as low as 12% of the original dataset size. In addition to the phase diagram, the GP is able to reconstruct spectra in energy-momentum space from this minimal set of data points. These findings suggest that this methodology can be used to improve the efficiency of ARPES data collection strategies for unknown samples. The practical feasibility of implementing this technology at a synchrotron beamline and the overall efficiency implications of this method are discussed with a view on enabling the collection of more samples or rapid identification of regions of interest.

Automated Searching and Identification of Self-Organized Nanostructures

Currently, researchers spend significant time manually searching through large volumes of data produced during scanning probe imaging to identify specific patterns and motifs formed via self-assembly and self-organization. Here, we use a combination of Monte Carlo simulations, general statistics, and machine learning to automatically distinguish several spatially correlated patterns in a mixed, highly varied data set of real AFM images of self-organized nanoparticles. We do this regardless of feature-scale and without the need for manually labeled training data. Provided that the structures of interest can be simulated, the strategy and protocols we describe can be easily adapted to other self-organized systems and data sets.

Host Cell Prediction of Exosomes Using Morphological Features on Solid Surfaces Analyzed by Machine Learning

Exosomes are extracellular nanovesicles released from any cells and found in any body fluid. Because exosomes exhibit information of their host cells (secreting cells), their analysis is expected to be a powerful tool for early diagnosis of cancers. To predict the host cells, we extracted multidimensional feature data about size, shape, and deformation of exosomes immobilized on solid surfaces by atomic force microscopy (AFM). The key idea is combination of support vector machine (SVM) learning for individual exosome particles and their interpretation by principal component analysis (PCA). We observed exosomes derived from three different cancer cells on SiO₂/Si, 3-aminopropyltriethoxysilane-modified-SiO₂/Si, and TiO₂ substrates by AFM. Then, 14-dimensional feature vectors were extracted from AFM particle data, and classifiers were trained in 14-dimensional space. The prediction accuracy for host cells of test AFM particles was examined by the cross-validation test. As a result, we obtained prediction of exosome host cells with the best accuracy of 85.2% for two-class SVM learning and 82.6% for three-class one. By PCA of the particle classifiers, we concluded that the main factors for prediction accuracy and its strong dependence on substrates are incremental decrease in the PCA-defined aspect ratio of the particles with their volume.

Deep data analysis via physically constrained linear unmixing: universal framework, domain examples, and a community-wide platform

Many spectral responses in materials science, physics, and chemistry experiments can be characterized as resulting from the superposition of a number of more basic individual spectra. In this context, unmixing is defined as the problem of determining the individual spectra, given measurements of multiple spectra that are spatially resolved across samples, as well as the determination of the corresponding abundance maps indicating the local weighting of each individual spectrum. Matrix factorization is a popular linear unmixing technique that considers that the mixture model between the individual spectra and the spatial maps is linear. Here, we present a tutorial paper targeted at domain scientists to introduce linear unmixing techniques, to facilitate greater understanding of spectroscopic imaging data. We detail a matrix factorization framework that can incorporate different domain information through various parameters of the matrix factorization method. We demonstrate many domain-specific examples to explain the expressivity of the matrix factorization framework and show how the appropriate use of domain-specific constraints such as non-negativity and sum-to-one abundance result in physically meaningful spectral decompositions that are more readily interpretable. Our aim is not only to explain the off-the-shelf available tools, but to add additional constraints when ready-made algorithms are unavailable for the task. All examples use the scalable open source implementation from https://github.com/ramkikannan/nmflibrary that can run from small laptops to supercomputers, creating a user-wide platform for rapid dissemination and adoption across scientific disciplines.

Improving superconductivity in BaFe2As2-based crystals by cobalt clustering and electronic uniformity

Quantum materials such as antiferromagnets or superconductors are complex in that chemical, electronic, and spin phenomena at atomic scales can manifest in their collective properties. Although there are some clues for designing such materials, they remain mainly unpredictable. In this work, we find that enhancement of transition temperatures in BaFe₂As₂-based crystals are caused by removing local-lattice strain and electronic-structure disorder by thermal annealing. While annealing improves Néel-ordering temperature in BaFe₂As₂ crystal (T_N = 132 K to 136 K) by improving in-plane electronic defects and reducing overall a-lattice parameter, it increases superconducting-ordering temperature in optimally cobalt-doped BaFe₂As₂ crystal (T_c = 23 to 25 K) by precipitating-out the cobalt dopants and giving larger overall a-lattice parameter. While annealing improves local chemical and electronic uniformity resulting in higher T_N in the parent, it promotes nanoscale phase separation in the superconductor resulting in lower disparity and strong superconducting band gaps in the dominant crystal regions, which lead to both higher overall T_c and critical-current-density, J_c.

Chemically induced topological zero mode at graphene armchair edges

The electronic and magnetic properties of chemically modified graphene armchair edges are studied using a combination of tight-binding calculations, first-principles modelling, and low temperature scanning tunneling microscopy (STM) experiments. The atomically resolved STM images of the hydrogen etched graphitic edges suggest the presence of localized states at the Fermi level for certain armchair edges. We demonstrate theoretically that the topological zero-energy edge mode may emerge at armchair boundaries with asymmetrical chemical termination of the two outermost atoms in the unit cell. We particularly focus our attention on armchair edges terminated by various combinations of the hydrogen (H, H₂) and methylene (CH₂) groups. The inclusion of the spin component in our calculations reveals the appearance of π-electron-based magnetism at the armchair edges under consideration.