Decoding the Mechanisms of Phase Transitions from In Situ Microscopy Observations

Abstract

Analysis of the temperature‐ and stimulus‐dependent imaging data toward elucidation of the physical transformations is an ubiquitous problem in multiple fields. Here, temperature‐induced phase transition in BaTiO3 is explored using the machine learning analysis of domain morphologies visualized via variable‐temperature scanning transmission electron microscopy (STEM) imaging data. This approach is based on the multivariate statistical analysis of the time or temperature dependence of the statistical descriptors of the system, derived in turn from the categorical classification of observed domain structures or projection on the continuous parameter space of the feature extraction‐dimensionality reduction transform. The proposed workflow offers a powerful tool for the exploration of the dynamic data based on the statistics of image representation as a function of the external control variable to visualize the transformation pathways during phase transitions and chemical reactions. This can include the mesoscopic STEM data as demonstrated here, but also optical, chemical imaging, etc., data. It can further be extended to the higher dimensional spaces, for example, analysis of the combinatorial libraries of materials compositions.

The Effect of Chemical Environment and Temperature on the Domain Structure of Free‐Standing BaTiO3 via In Situ STEM

“…This is likely due to the lamella dimensions and boundary conditions, such as thickness and potential clamping to the bottom chip, as reported elsewhere. [37,40,41] The change in domain configuration, from a-c to a-a between RT and 60 °C, is typically not observed in bulk but it resembles similar behavior to that reported in low-strain BTO thin films [42,43] and is comparable to recent observations for BTO lamellae in situ studies. [36,37,[44][45][46] At this point, it would be tempting to suggest that this domain reconfiguration is due to the material accessing the next low-temperature phase transition, from tetragonal to orthorhombic, [47] allowed perhaps due to the lowstrain regime of the BTO lamellae.…”

“…Although not as precise as other bulk techniques used to measure T C in bulk, the lack of contrast here observed is an indication of the material undergoing the expected phase transition into the cubic phase, corresponding to the disappearance of the symmetry‐breaking tetragonal distortion, similar to previous reports on lamellae BTO in situ heating. [ 36 , 37 ] Upon cooling, the sample displays a hysteretic behavior as a function of temperature, indicative of a first‐order phase transition. [ 38 , 39 ] The domain wall variants are present in the same temperature range and with almost identical area fractions as when heated up.…”

“…This is likely due to the lamella dimensions and boundary conditions, such as thickness and potential clamping to the bottom chip, as reported elsewhere. [ 37 , 40 , 41 ]…”

“…The change in domain configuration, from a‐c to a‐a between RT and 60 °C, is typically not observed in bulk but it resembles similar behavior to that reported in low‐strain BTO thin films [ 42 , 43 ] and is comparable to recent observations for BTO lamellae in situ studies. [ 36 , 37 , 44 , 45 , 46 ] At this point, it would be tempting to suggest that this domain reconfiguration is due to the material accessing the next low‐temperature phase transition, from tetragonal to orthorhombic, [ 47 ] allowed perhaps due to the low‐strain regime of the BTO lamellae. This would suggest an almost transitional character for BTO in a free‐standing form, similar to that reported in low‐strain BTO thin films.…”

Bayesian optimization in continuous spaces via virtual process embeddings

“…[8][9][10][11] For materials synthesis, this is oen the compositional space of selected multicomponent phase diagram, 12,13 or the processing history. 14,15 For imaging, this is the image plane of the object 16 or the parameter space of the microscope control. 17 From the optimization perspective, critical considerations are the dimensionality and completeness of the parameter space, and the properties of the target function dened over it.…”