McLean P. Echlin

Plastic deformation in polycrystalline metals is usually primarily accommodated through dislocation glide, which often organizes into discrete slip bands. The evolution of plasticity is governed by the interactions of these dislocation structures with one another, as well as with grain boundaries and other microstructural interfaces or features. Traditional observations of slip bands are largely limited to free surfaces, constraining our understanding of their three-dimensional behavior within the bulk. The X-ray topotomography and diffraction contrast tomography (DCT) datasets presented here enable direct visualization of the three-dimensional morphology of slip bands in relation to the surrounding grain structure. These data reveal, for instance, in the investigated titanium alloy that slip transmission across grain boundaries is more prevalent at free surfaces, while in the bulk, transmission events are often driven by geometrically constrained regions, such as triple junctions, where elevated local stress heterogeneity promotes dislocation activity across interfaces. This dataset is shared with the community to facilitate broader access to and analysis of correlated 3D deformation and microstructure alloy datasets, which are expected to become increasingly prevalent in the near future.

The response of metals and alloys depends on a wide range of factors ranging from composition and point defects, to larger defects such as dislocations and/or precipitates, and aspects such as the grain size, texture or other microstructural factors. Furthermore, how a material is processed and then subsequently loaded can also strongly influence material behavior. Combined, these aspects span length and time scales that vary by several orders of magnitude, providing a grand challenge to any single modeling approach. Consequently, over the past decades, multiscale modeling (MM) has emerged as a foundational tool for investigating and predicting the performance. While much development has been done, there are still open challenges in how methods bridge across different time and/or length scales, how we inform, validate, and understand complex material behavior, and more recently how emerging machine learning and artificial intelligence tools can enhance our MM approaches. Addressing these challenges and the outlook for MM on the horizon is the focus of this roadmap article. The contributions within are divided broadly into three categories: bridging the length scale gap between models, targeting timescale gaps alongside length scale gaps, and bridging the gap between experiment and simulation.

Gaining understanding of process-structure-property relationships in materials at a mechanistic level relies on correlative microscopy workflows. These workflows, in turn, fundamentally depend on image matching, i.e., a computer vision task with the objective of finding point correspondences in pairs of images. Matching models are difficult to evaluate quantitatively in the materials field due to a shortage of representative benchmark datasets. Nonetheless, prior research indicates that traditional rule-based image matching techniques such as the surface-invariant feature transform currently fall short on such matching tasks. We present a dataset for cross-modal image matching and data fusion in the materials microscopy domain, which we coin AmalgaMatch, to support model benchmarking and fine-tuning efforts. All images are micrographs captured using the most widely applied imaging techniques in materials science including light-optical, scanning electron, and transmission electron microscopy, as well as electron backscatter diffraction (EBSD). Therein, various detectors and imaging modes are employed to capture micrographs of diverse materials. While the majority of images are raw images, some underwent typical processing routes using digital image correlation or EBSD indexing. Common regions in image pairs are populated with hand-annotated keypoint correspondences. While mutual information is limited in cross-modal, multi-scale image pairs, we relied on characteristic defects such as dislocations, grain boundaries, triple junctions, inclusions, pores or topographic features for annotation. Furthermore, the dataset is divided into groups, defined by distinct registration use cases, and further into subsets, defined by the imaged material. The dataset covers many typical use cases for image matching in materials science, including slip partitioning, dislocation characterization, and surface fractography. In total, it comprises 6 groups and 19 subsets with 35 scenes and 187 annotated image pairs to support autonomous multi-modal materials data fusion. For each image, we provide structured metadata to facilitate training of hybrid matching models which process textual alongside image-based inputs to improve the matching quality and robustness. A formal ontological model for correlative microscopy and image matching processes is proposed to express image contents, relationships, and transformations through knowledge graphs and to enable aligning with FAIR data principles.

This work comprehensively assesses the capabilities of recently emerging foundation models for multimodal image matching and registration in the materials science and engineering (MSE) domain. A distinctive feature in correlative materials microscopy is that the images, which need to be spatially associated, commonly not only have limited mutual information but also distinct length scales. To date, it is largely unknown how foundational multimodal matchers, such as MatchAnything RoMa or ELoFTR, which were trained predominantly with macroscale images from the human environment, generalize to image pairs in such an out-of-domain setting-especially given the pronounced information and field of view disparities in image pairs. To evaluate these models, we use the recently published AmalgaMatch dataset which covers 187 image pairs partitioned into six groups which represent distinct image matching tasks commonly faced in correlative materials microscopy and 19 subsets representing different materials. This dataset with its diverse microscopy modalities and broad range of depicted metals, alloys, and ceramics facilitates evaluation of these models in a rather representative manner. We observe that MatchAnything RoMa, as opposed to the ELoFTR variant, attains satisfactory matching results for many matching tasks and materials represented in this dataset. In contrast, image pairs within the dislocation characterization, slip partitioning and multi-scale groups pose a difficult challenge for these models. These pairs often exhibit limited mutual information and strong field of view mismatch, with field of view ratios reaching down to 2%. We propose inference strategies and workflows which embed a foundation model in an image processing pipeline to increase the matching quality and robustness and ultimately overcome aforementioned challenges inherent to MSE matching tasks.

Geometrically necessary dislocation (GND) content is measured from mm³-scaled Ti7Al three-dimensional (3D) microstructural data using a theory extended for hexagonal close packed crystals, which accounts for basal, prismatic and pyramidal ⟨ c + a ⟩ type dislocation content. The Ti7Al samples have been mechanically pre-strained to two different strain levels, and will then be strained along the same axis in uniaxial tension during simulation. Both inter- and intragranular GNDs across the microstructures have been characterized, with a large contribution of pyramidal ⟨ c + a ⟩ GNDs, consistent with the relative slip activity involved in pre-straining. The spatially resolved crystallographic GND distributions within the 3D microstructures are used to instantiate a microstructure model for forward modeling deformation simulations by a dislocation density hardening elasto-viscoplastic fast Fourier transform framework. Coarsening the voxel resolution during the initial microstructure construction procedure is shown to strongly impact both the magnitude and spatial distribution of the GNDs and in turn the forward deformation response of the pre-strained material. This study indicates that the voxel resolution desired when transferring from measured to model microstructures need not only be proportionally scaled with the microstructure but also sufficiently fine to capture the subgranular orientation gradients that may already be present in the material.