arXiv subjects

Sergei Kalinin

Publications and source records attributed to Sergei Kalinin.

At least 19 recordsLinked to original sources

arXiv2026

Multitask Scanning Probe Microscopy

Scanning probe microscopy provides nanoscale access to structural, electrical, electromechanical, magnetic, and mechanical properties of materials. Its increasing use for wafer-scale characterization and combinatorial materials exploration creates a need to distribute measurements efficiently across large spatial domains. This is particularly important when available modalities differ in acquisition time and potential for tip and sample damage, making exhaustive multimodal mapping over spatial grids impractical. Here, we demonstrate multitask scanning probe microscopy, a live, closed-loop workflow in which a multitask Gaussian process learns spatial and cross-modal relationships and autonomously selects both the next measurement location and the next experimental protocol. The approach is implemented on an automated large-sample atomic force microscope and demonstrated on a composition-spread AlScN wafer using tapping-mode and Dual AC Resonance Tracking (DART) measurements. Paired initial measurements establish the relation between the tasks, after which noncoincident measurements are used to update both response landscapes. The resulting workflow extends active learning in scanning probe microscopy from spatial sampling to autonomous allocation of measurement modalities and provides a basis for combining rapid, weakly perturbative imaging with slower contact, electrical, electromechanical, magnetic, or spectroscopic measurements.

cond-mat.mtrl-sci↗

arXiv2026

LLM-Guided Open Hypothesis Learning from Autonomous Scanning Probe Microscopy Experiments

Autonomous experimentation has transformed microscopy and materials discovery by enabling closed-loop optimization including imaging and spectroscopy tuning, strucutre property relationship discovery, and exploration of combinatorial libraries. However, most current workflows remain limited to selecting measurements within fixed objective or hypothesis spaces, rather than generating new physical models from experimental data. Here, we introduce an open hypothesis-learning framework that combines symbolic regression with large-language-model-based physical evaluation and implement it for autonomous scanning probe microscopy. Symbolic regression generates candidate analytical relationships directly from sparse measurements, while the language-model evaluator ranks these candidates according to physical plausibility, scaling behavior, and consistency with known mechanisms. We demonstrate the approach on autonomous piezoresponse force microscopy measurements of ferroelectric domain switching in a PZT thin film. Starting from five seed measurements, the workflow evolves from physically incomplete candidate expressions toward interpretable voltage-time growth laws consistent with kinetic domain-wall motion. This work extends autonomous microscopy from closed-loop optimization toward open hypothesis discovery, where candidate physical laws emerge from the experiment itself rather than being specified in advance. More broadly, the framework establishes a route for integrating symbolic regression, physical reasoning, and adaptive experimentation into hierarchical autonomous scientific workflows.

cond-mat.mtrl-sci↗

arXiv2026

From Photons to Electrons: Accelerated Materials Discovery via Random Libraries and Automated Scanning Transmission Electron Microscopy

The real-world implementation of materials prediction algorithms remains limited by persistent characterization bottlenecks in materials discovery, where photon-based probe techniques (e.g., XRD or Raman) impose long acquisition times and access latencies, restricting exploration to quasi-ternary composition spaces typically realized as compositional libraries. Here, we argue that a paradigm shift from photon- to electron-based characterization can realign materials characterization with modern high-throughput synthesis. We formulate cost functions and exploration strategies for STEM-based chemical and structural characterization and use Monte Carlo simulations to show that random chemical libraries, where compositionally distinct regions are co-located within a single specimen and interrogated in situ by electron spectroscopies, can sample high-dimensional composition and phase spaces with orders-of-magnitude greater effective coverage than conventional spread-library/X-ray approaches. We further demonstrate autonomous discovery on a laboratory STEM platform, where ML-based autotuning and scripted control enable iterative region selection and characterization without human intervention. Finally, we outline extensions to labeled or position-encoded libraries that preserve compositional and processing metadata, enabling joint exploration of composition and process spaces. Together, these results establish electron-based, ML-enabled STEM as a scalable pathway toward combinatorially rich materials discovery.

cond-mat.mtrl-sci↗

arXiv2026

Sequential versus Manifold Bayesian Optimization under Realistic Experimental Time Constraints

Bayesian optimization (BO) is widely used for autonomous materials discovery, yet its classical sequential formulation is insufficient for design of experimental workflows that often combine parallel or batch synthesis with inherently serial characterization. Methods such as combinatorial spread libraries and printed libraries sample a defined low-D manifold in the chemical space of the system. Here, we introduce a time-aware framework for comparing sequential and manifold BO under experimentally realistic constraints. By explicitly modeling synthesis and characterization times, we define an effective experimental time metric that enables fair, time-normalized benchmarking of optimization strategies. Using numerical experiments in ternary and quaternary compositional spaces, we show that sequential BO remains optimal for short-term experiments or when batching provides no effective time advantage, whereas manifold BO becomes favorable once multiplexed synthesis enables faster accumulation of measurements. We identify a small set of physically interpretable parameters that govern the transition between these regimes. These results establish a general, experimentally grounded framework for selecting optimization strategies in self-driving laboratories and autonomous materials discovery workflows. The accompanying analysis code is publicly available at https://github.com/Slautin/2025_GP_BO_Manifolds.

cond-mat.mtrl-sci↗

arXiv2026

Towards Self-Optimizing Electron Microscope: Robust Tuning of Aberration Coefficients via Physics-Aware Multi-Objective Bayesian Optimization

Realizing high-throughput aberration-corrected Scanning Transmission Electron Microscopy (STEM) exploration of atomic structures requires rapid tuning of multipole probe correctors while compensating for the inevitable drift of the optical column. While automated alignment routines exist, conventional approaches rely on serial, gradient-free searches (e.g., Nelder-Mead) that are sample-inefficient and struggle to correct multiple interacting parameters simultaneously. Conversely, emerging deep learning methods offer speed but often lack the flexibility to adapt to varying sample conditions without extensive retraining. Here, we introduce a Multi-Objective Bayesian Optimization (MOBO) framework for rapid, data-efficient aberration correction. Importantly, this framework does not prescribe a single notion of image quality; instead, it enables user-defined, physically motivated reward formulations (e.g., symmetry-induced objectives) and uses Pareto fronts to expose the resulting trade-offs between competing experimental priorities. By using Gaussian Process regression to model the aberration landscape probabilistically, our workflow actively selects the most informative lens settings to evaluate next, rather than performing an exhaustive blind search. We demonstrate that this active learning loop is more robust than traditional optimization algorithms and effectively tunes focus, astigmatism, and higher-order aberrations. By balancing competing objectives, this approach enables "self-optimizing" microscopy by dynamically sustaining optimal performance during experiments.

cs.LG↗

arXiv2025

Random Combinatorial Libraries and Automated Nanoindentation for High-Throughput Structural Materials Discovery

Accelerating the discovery of structural materials is essential for applications in hard and refractory alloys, hypersonic platforms, nuclear systems, and other extreme environment technologies. Progress is often constrained by slow synthesis and characterization cycles and the need for extensive mechanical testing across large compositional spaces. Here, we propose a rapid screening strategy based on random material libraries, in which thousands of distinct compositions are embedded within a single specimen, mapped by EDS, and subsequently characterized. Using nanoindentation as a representative case, we show that such libraries enable dense composition property mapping while reducing the number of samples required to explore high dimensional composition spaces compared to traditional synthesis and test workflows. An experimentally calibrated Monte Carlo framework is developed to quantify practical limits, including particle size, EDS noise and resolution, positional accuracy, and nanoindenter motion costs. The simulations identify regimes where random libraries provide orders of magnitude acceleration over classical workflows. Finally, we demonstrate experimental navigation of these libraries using automated indentation. Together, these results establish random libraries as a general route to high throughput characterization in structurally critical material systems.

cond-mat.mtrl-sci↗

arXiv2025

Building Trustworthy AI for Materials Discovery: From Autonomous Laboratories to Z-scores

Accelerated material discovery increasingly relies on artificial intelligence and machine learning, collectively termed "AI/ML". A key challenge in using AI is ensuring that human scientists trust the models are valid and reliable. Accordingly, we define a trustworthy AI framework GIFTERS for materials science and discovery to evaluate whether reported machine learning methods are generalizable, interpretable, fair, transparent, explainable, robust, and stable. Through a critical literature review, we highlight that these are the trustworthiness principles most valued by the materials discovery community. However, we also find that comprehensive approaches to trustworthiness are rarely reported; this is quantified by a median GIFTERS score of 5/7. We observe that Bayesian studies frequently omit fair data practices, while non-Bayesian studies most frequently omit interpretability. Finally, we identify approaches for improving trustworthiness methods in artificial intelligence and machine learning for materials science by considering work accomplished in other scientific disciplines such as healthcare, climate science, and natural language processing with an emphasis on methods that may transfer to materials discovery experiments. By combining these observations, we highlight the necessity of human-in-the-loop, and integrated approaches to bridge the gap between trustworthiness and uncertainty quantification for future directions of materials science research. This ensures that AI/ML methods not only accelerate discovery, but also meet ethical and scientific norms established by the materials discovery community. This work provides a road map for developing trustworthy artificial intelligence systems that will accurately and confidently enable material discovery.

cond-mat.mtrl-sci↗

arXiv2025

Accelerated Materials Discovery through Cost-Aware Bayesian Optimization of Real-World Indentation Workflows

Accelerating the discovery of mechanical properties in combinatorial materials requires autonomous experimentation that accounts for both instrument behavior and experimental cost. Here, an automated nanoindentation (AE-NI) framework is developed and validated for adaptive mechanical mapping of combinatorial thin-film libraries. The method integrates heteroskedastic Gaussian-process modeling with cost-aware Bayesian optimization to dynamically select indentation locations and hold times, minimizing total testing time while preserving measurement accuracy. A detailed emulator and cost model capture the intrinsic penalties associated with lateral motion, drift stabilization, and reconfiguration-factors often neglected in conventional active-learning approaches. To prevent kernel-length-scale collapse caused by disparate time scales, a hierarchical meta-testing workflow combining local grid and global exploration is introduced. Implementation of the workflow is shown on a experimental Ta-Ti-Hf-Zr thin-film library. The proposed framework achieves nearly a thirty-fold improvement in property-mapping efficiency relative to grid-based indentation, demonstrating that incorporating cost and drift models into probabilistic planning substantially improves performance. This study establishes a generalizable strategy for optimizing experimental workflows in autonomous materials characterization and can be extended to other high-precision, drift-limited instruments.

cond-mat.mtrl-sci↗

arXiv2025

DIVIDE: A Framework for Learning from Independent Multi-Mechanism Data Using Deep Encoders and Gaussian Processes

Scientific datasets often arise from multiple independent mechanisms such as spatial, categorical or structural effects, whose combined influence obscures their individual contributions. We introduce DIVIDE, a framework that disentangles these influences by integrating mechanism-specific deep encoders with a structured Gaussian Process in a joint latent space. Disentanglement here refers to separating independently acting generative factors. The encoders isolate distinct mechanisms while the Gaussian Process captures their combined effect with calibrated uncertainty. The architecture supports structured priors, enabling interpretable and mechanism-aware prediction as well as efficient active learning. DIVIDE is demonstrated on synthetic datasets combining categorical image patches with nonlinear spatial fields, on FerroSIM spin lattice simulations of ferroelectric patterns, and on experimental PFM hysteresis loops from PbTiO3 films. Across benchmarks, DIVIDE separates mechanisms, reproduces additive and scaled interactions, and remains robust under noise. The framework extends naturally to multifunctional datasets where mechanical, electromagnetic or optical responses coexist.

cs.LG↗

arXiv2025

Polarization switching on the open surfaces of the wurtzite ferroelectric nitrides: ferroelectric subsystems and electrochemical reactivity

Binary ferroelectric nitrides are promising materials for information technologies and power electronics. However, polarization switching in these materials is highly unusual. From the structural perspective, polarization reversal is associated with the change of the effective polarity at the surfaces and interfaces from N-to-M terminated, suggesting strong coupling between ferroelectric and chemical phenomena. Phenomenologically, macroscopic studies demonstrate the presence of complex time dependent phenomena including wake-up. Here, we explore the polarization switching using the multidimensional high-resolution piezoresponse force microscopy (PFM) and spectroscopy, detecting both the evolution of induced ferroelectric domain, electromechanical response, and surface deformation during first-order reversal curve measurements. We demonstrate the presence of two weakly coupled ferroelectric subsystems and the bias-induced electrochemical reactivity. The observed behaviors are very similar to the recent studies of other wurtzite system but additionally include electrochemical reactivity, suggesting the universality of these behaviors for the wurtzite binary ferroelectrics. These studies suggest potential of high-resolution multimodal PFM spectroscopies to resolve complex coupled polarization dynamics in materials. Furthermore, these PFM based studies are fully consistent with the recent electron microscopy observations of the shark-teeth like ferroelectric domains in nitrides. Hence, we believe that these studies establish the universal phenomenological picture of polarization switching in binary wurtzite.

cond-mat.mtrl-sci↗

arXiv2025

Automating Nanoindentation: Optimizing Workflows for Precision and Accuracy

Nanoindentation is vital for probing mechanical properties, yet traditional grid-based workflows are inefficient for targeting specific microstructural features. We present an automated nanoindentation framework that integrates machine learning, real-time alignment, and adaptive indentation strategies. The system operates in three modes: standard automation, feature-based indentation via image-to-coordinate mapping, and large-scale indentation with full x, y, and z axis alignment. A key challenge (precise sample positioning across imaging and indentation stages) was met by correcting initial travel-distance errors (2.5-6 micrometres) through pixel-to-micron calibration, reducing alignment errors to the submicron level. Benchmark tests demonstrate phase-specific and orientation-guided indentation enabled by self-organising maps and macro imaging. The framework markedly improves precision, minimises user intervention, and enables efficient, targeted characterisation of complex materials. By providing a direct interface between nanoindentation instruments and Python-based automation frameworks, it can be adopted on most existing platforms. This work lays the foundation for next-generation autonomous mechanical testing tailored to microstructurally complex materials.

physics.ins-det↗

arXiv2024

Invariant Discovery of Features Across Multiple Length Scales: Applications in Microscopy and Autonomous Materials Characterization

Physical imaging is a foundational characterization method in areas from condensed matter physics and chemistry to astronomy and spans length scales from atomic to universe. Images encapsulate crucial data regarding atomic bonding, materials microstructures, and dynamic phenomena such as microstructural evolution and turbulence, among other phenomena. The challenge lies in effectively extracting and interpreting this information. Variational Autoencoders (VAEs) have emerged as powerful tools for identifying underlying factors of variation in image data, providing a systematic approach to distilling meaningful patterns from complex datasets. However, a significant hurdle in their application is the definition and selection of appropriate descriptors reflecting local structure. Here we introduce the scale-invariant VAE approach (SI-VAE) based on the progressive training of the VAE with the descriptors sampled at different length scales. The SI-VAE allows the discovery of the length scale dependent factors of variation in the system. Here, we illustrate this approach using the ferroelectric domain images and generalize it to the movies of the electron-beam induced phenomena in graphene and topography evolution across combinatorial libraries. This approach can further be used to initialize the decision making in automated experiments including structure-property discovery and can be applied across a broad range of imaging methods. This approach is universal and can be applied to any spatially resolved data including both experimental imaging studies and simulations, and can be particularly useful for exploration of phenomena such as turbulence, scale-invariant transformation fronts, etc.

physics.comp-ph↗

arXiv2023

Learning and Controlling Silicon Dopant Transitions in Graphene using Scanning Transmission Electron Microscopy

We introduce a machine learning approach to determine the transition dynamics of silicon atoms on a single layer of carbon atoms, when stimulated by the electron beam of a scanning transmission electron microscope (STEM). Our method is data-centric, leveraging data collected on a STEM. The data samples are processed and filtered to produce symbolic representations, which we use to train a neural network to predict transition probabilities. These learned transition dynamics are then leveraged to guide a single silicon atom throughout the lattice to pre-determined target destinations. We present empirical analyses that demonstrate the efficacy and generality of our approach.

cond-mat.mes-hall↗

arXiv2023

Physics and Chemistry from Parsimonious Representations: Image Analysis via Invariant Variational Autoencoders

Electron, optical, and scanning probe microscopy methods are generating ever increasing volume of image data containing information on atomic and mesoscale structures and functionalities. This necessitates the development of the machine learning methods for discovery of physical and chemical phenomena from the data, such as manifestations of symmetry breaking in electron and scanning tunneling microscopy images, variability of the nanoparticles. Variational autoencoders (VAEs) are emerging as a powerful paradigm for the unsupervised data analysis, allowing to disentangle the factors of variability and discover optimal parsimonious representation. Here, we summarize recent developments in VAEs, covering the basic principles and intuition behind the VAEs. The invariant VAEs are introduced as an approach to accommodate scale and translation invariances present in imaging data and separate known factors of variations from the ones to be discovered. We further describe the opportunities enabled by the control over VAE architecture, including conditional, semi-supervised, and joint VAEs. Several case studies of VAE applications for toy models and experimental data sets in Scanning Transmission Electron Microscopy are discussed, emphasizing the deep connection between VAE and basic physical principles. All the codes used here are available at https://github.com/saimani5/VAE-tutorials and this article can be used as an application guide when applying these to own data sets.

cs.LG↗

arXiv2021

Robust Feature Disentanglement in Imaging Data via Joint Invariant Variational Autoencoders: from Cards to Atoms

Recent advances in imaging from celestial objects in astronomy visualized via optical and radio telescopes to atoms and molecules resolved via electron and probe microscopes are generating immense volumes of imaging data, containing information about the structure of the universe from atomic to astronomic levels. The classical deep convolutional neural network architectures traditionally perform poorly on the data sets having a significant orientational disorder, that is, having multiple copies of the same or similar object in arbitrary orientation in the image plane. Similarly, while clustering methods are well suited for classification into discrete classes and manifold learning and variational autoencoders methods can disentangle representations of the data, the combined problem is ill-suited to a classical non-supervised learning paradigm. Here we introduce a joint rotationally (and translationally) invariant variational autoencoder (j-trVAE) that is ideally suited to the solution of such a problem. The performance of this method is validated on several synthetic data sets and extended to high-resolution imaging data of electron and scanning probe microscopy. We show that latent space behaviors directly comport to the known physics of ferroelectric materials and quantum systems. We further note that the engineering of the latent space structure via imposed topological structure or directed graph relationship allows for applications in topological discovery and causal physical learning.

physics.data-an↗

arXiv2016

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

Electronic interactions present in material compositions close to the superconducting dome play a key role in the manifestation of high-Tc superconductivity. In many correlated electron systems, however, the parent or underdoped states exhibit strongly inhomogeneous electronic landscape at the nanoscale that may be associated with competing, coexisting, or intertwined chemical disorder, strain, magnetic, and structural order parameters. Here we demonstrate an approach based on a combination of scanning tunneling microscopy/spectroscopy (STM/S) and advanced statistical learning for an automatic separation and extraction of statistically significant electronic behaviors in the spin density wave (SDW) regime of a lightly (~1%) gold-doped BaFe2As2. We show that the decomposed STS spectral features have a direct relevance to fundamental physical properties of the system, such as SDW-induced gap, pseudogap-like state, and impurity resonance states.

cond-mat.mtrl-sci↗

arXiv2016

Data Mining Graphene: Correlative Analysis of Structure and Electronic Degrees of Freedom in Graphenic Monolayers with Defects

The link between changes in the material crystal structure and its mechanical, electronic, magnetic, and optical functionalities - known as the structure-property relationship - is the cornerstone of the contemporary materials science research. The recent advances in scanning transmission electron and scanning probe microscopies (STEM and SPM) have opened an unprecedented path towards examining the materials structure-property relationships on the single-impurity and atomic-configuration levels. Lacking, however, are the statistics-based approaches for cross-correlation of structure and property variables obtained in different information channels of the STEM and SPM experiments. Here we have designed an approach based on a combination of sliding window fast Fourier transform, Pearson correlation matrix, and linear and kernel canonical correlation, to study a relationship between lattice distortions and electron scattering from SPM data on graphene with defects. Our analysis revealed that the strength of coupling to strain is altered between different scattering channels, which can explain coexistence of several quasiparticle interference patterns in nanoscale regions of interest. In addition, the application of kernel functions allowed us to extract a non-linear component of the relationship between the lattice strain and scattering intensity in graphene. The outlined approach can be further utilized towards analyzing correlations in various multi-modal imaging techniques where the information of interest is spatially distributed and generally has a complex multidimensional nature.

cond-mat.mtrl-sci↗

arXiv2010

Ferroelectric domain scaling and electronic properties in ultrathin BiFeO3 films on vicinal substrates

We report electrically switchable polarization and ferroelectric domain scaling over a thickness range of 5-100 nm in BiFeO3 films deposited on [110] vicinal substrates. The BiFeO3 films of variable thickness were deposited with SrRuO3 bottom layer using pulsed laser deposition technique. These films have fractal domain patterns and the domain width scales closely with the square root of film thickness, in accordance with the Landau-Lifschitz-Kittel (LLK) law. The Switching Spectroscopy Piezo-response Force Microscopy provides clear evidence for the ferroelectric switching behavior in all the films. Using Quasi-particle Self-consistent GW (QPGW) approximation we have investigated physical parameters relevant for direct tunneling behavior, namely the effective mass and effective barrier height of electrons. For rhombohedral BFO, we report a large effective barrier height value of 3.6 eV, which is in reasonable agreement with nanoscale transport measurements. QPGW investigations into the tetragonal BFO structure with P4mm symmetry revealed a barrier height of 0.38 eV, significantly lower compared to its rhombohedral counterpart. This difference has very significant implications on the transport properties of nearly tetragonal BFO phase.

cond-mat.mtrl-sci↗