Atom Probe Tomography Data Research Articles

Computer vision assisted decomposition analysis of atom probe tomography data

While atom probe tomography (APT) offers 3D spatially-resolved compositional characterization at the nanometer scale, individual inspection of large datasets is time demanding and the interpretation thereof is highly operator-dependent. Recently, the decomposition behavior of (V,Al)N thin films was investigated by APT. The conventional analysis, employing standard tools from commercial software, is laborious and only a small fraction (below 5%) of the total acquired data has been used. In the present work, the analysis workflow is automated, taking into account more than 90% of the total acquired data. As a first step, the decomposition products, including matrix and enriched phases, were identified via deep learning-based semantic image segmentation. Performance of the ‘matrix’ and the decomposition product ‘aluminum-rich’ was very good, achieving F1-scores of 0.994 and 0.943, respectively, while performance for ‘vanadium-rich’ was with 0.854 slightly lower. The segmentation enables the investigation of phase formation evolution over temperature. As a next step, early stages of the spinodal decomposition were probed on the ‘matrix’ phase by a neighborhood analysis. While results were consistent with prior work, the data analysis in this work is statistically more robust by leveraging an order of magnitude larger dataset. In the end, the transferability of the presented phase segmentation workflow of the (V,Al)N-based model is shown and discussed for the decomposition of isostructural (Ti,Al)N.

A Machine Learning Framework for Quantifying Chemical Segregation and Microstructural Features in Atom Probe Tomography Data.

Atom probe tomography (APT) is ideally suited to characterize and understand the interplay of segregation and microstructure in modern multi-component materials. Yet, the quantitative analysis typically relies on human expertise to define regions of interest. We introduce a computationally efficient, multi-stage machine learning strategy to identify compositionally distinct domains in a semi-automated way, and subsequently quantify their geometric and compositional characteristics. In our algorithmic pipeline, we first coarse-grain the APT data into voxels, collect the composition statistics, and decompose it via clustering in composition space. The composition classification then enables the real-space segmentation via a density-based clustering algorithm, thus revealing the microstructure at voxel resolution. Our approach is demonstrated for a Sm-(Co,Fe)-Zr-Cu alloy. The alloy exhibits two precipitate phases with a plate-like, but intertwined morphology. The primary segmentation is further refined to disentangle these geometrically complex precipitates into individual plate-like parts by an unsupervised approach based on principle component analysis, or a U-Net-based semantic segmentation trained on the former. Following the composition and geometric analysis, detailed composition distribution and segregation effects relative to the predominant plate-like geometry can be readily mapped from the point cloud, without resorting to the voxel compositions.

Introducing a Dynamic Reconstruction Methodology for Multilayered Structures in Atom Probe Tomography.

Atom probe tomography (APT) is a powerful three-dimensional nanoanalyzing microscopy technique considered key in modern materials science. However, progress in the spatial reconstruction of APT data has been rather limited since the first implementation of the protocol proposed by Bas et al. in 1995. This paper proposes a simple semianalytical approach to reconstruct multilayered structures, i.e., two or more different compounds stacked perpendicular to the analysis direction. Using a field evaporation model, the general dynamic evolution of parameters involved in the reconstruction of this type of structure is estimated. Some experimental reconstructions of different structures through the implementation of this method that dynamically accommodates variations in the tomographic reconstruction parameters are presented. It is shown both experimentally and theoretically that the depth accuracy of reconstructed APT images is improved using this method. The method requires few parameters in order to be easily usable and substantially improves atom probe tomographic reconstructions of multilayered structures.

Direct recognition of crystal structures via three-dimensional convolutional neural networks with high accuracy and tolerance to random displacements and missing atoms

Computational methods and machine learning algorithms for automatic information extraction are crucial to enable data-driven materials science. These approaches are changing materials characterization and analytics, which often require a user-specified threshold to e.g. detect structure or symmetries in structures with defects. Here, we present a machine learning-based approach that directly works on the original periodic arrangements of atoms based on a three-dimensional convolutional neural network without any transformation of descriptors. Our approach shows a high classification accuracy and tolerance to the presence of random displacements and missing atoms. Experimentally, we successfully reconstruct the ordered L12 precipitates extracted from atom probe tomography data, consistent with segmentation based on isocomposition surfaces. The convolutional layers are essential for the simultaneous identification of compositional and structural information, which also give rise to its high tolerance. Our work advances machine learning-based crystal structure identification for incomplete crystal structural data.

The origin of platinum group minerals in oceanic crust

Abstract Highly siderophile elements (HSEs), including Re and Os, are used extensively as geochemical tracers and geochronometers to investigate the formation and evolution of Earth’s crust and mantle. Mantle rocks are commonly serpentinized, but the effect of serpentinization on the distribution of HSEs is controversial because HSEs are commonly hosted by rare, micrometer- to sub-micrometer-scale grains of platinum group minerals (PGMs) of ambiguous origin that are challenging to identify, characterize, and interpret. In this study, atom probe tomography (APT) is used to characterize two spatially close PGM grains hosted by a partially serpentinized harzburgite from Macquarie Island, Australia. The APT data reveal an extraordinary level of detail that provides insights into the origin of a complex Cu–Pt alloy grain (average composition ~Cu4Pt). The grain hosts Fe-, Ni-, and Pt-rich sub-grains associated with Rh, variably overlapping networks of Pd- and Cd-enrichment, and OH-rich volumes identified as fluid inclusions. Osmium and Ru are hosted by an idioblastic laurite (RuS2) grain. Compositional, textural, and phase-diagram constraints are consistent with a modified pre-serpentinization origin for the PGMs, and a comparison between observed and calculated grain distributions indicate that while Os isotope ratios were probably unaffected by serpentinization, whole-rock and grain-scale HSE and isotopic ratios may have been decoupled during serpentinization.

Focused Ion Beam induced hydride formation does not affect Fe, Ni, Cr-clusters in irradiated Zircaloy-2

Room temperature focused ion beam (FIB) milling is known to potentially promote the formation of hydrides in zirconium and its alloys. We used atom probe tomography (APT) to determine the composition of irradiated and as-produced Zircaloy-2 fuel cladding. We consistently found ∼ 50 at% hydrogen in all room temperature FIB-milled specimens run in voltage pulsing APT measurements. Crystallographic analysis of APT data however showed slightly better agreement with δ-hydride (ZrH2, FCC, ∼ 60–66.7 at% H) than γ-hydride (ZrH, FCT, ∼ 50 at% H). Electron energy loos spectroscopy (EELS) measurements prior to APT analyses confirmed the presence of δ-hydride. Hence, APT gives a systematic underestimation of hydrogen for Zr-hydride. Milling at cryogenic temperatures was found to not cause such hydride formation. However, we did not find significant differences in the clusters formed by segregation of the alloying elements Fe, Cr and Ni to irradiation induced a-loops whether the material was identified as α-Zr or hydride. Therefore, analyzing irradiation-induced redistribution of alloying elements in Zr fuel cladding using APT does not rely on FIB preparation at cryogenic temperatures. However, in conjunction with voltage pulsing APT cryo-FIB can be worthwhile if one aims at investigating hydrogen distribution or hydrides.

Radiogenic Pb in xenotime trapped in nanoscale inclusions of apatite during fluid alteration

This study focuses on the low-temperature mineralogical response of xenotime, a phosphate mineral routinely used as a geochronometer, to fluid-assisted alteration. The studied xenotime grain (z6413) comes from a ∼ 1000 Ma pegmatite from the Grenville Province, Canada, and is commonly used as reference material for UPb analyses. At the microscale, the grain has a mottled texture, sub-micrometer porosity, and small domains dark in backscattered electron (BSE) images that are characterised by curviplanar, sharp boundaries. The small dark BSE domains are associated with Th-U-rich inclusions and larger porosity (2–3 μm) and are interpreted to result from localised fluid-assisted coupled dissolution-reprecipitation. Sensitive high-resolution ion microprobe (SHRIMP) UPb analyses of unaltered and fluid-affected domains yield concordant crystallisation dates, irrespective of the textural domains. The apparently unaltered xenotime domain was characterised at the nanoscale to determine if the grain was affected by fluids beyond the altered domains defined by BSE imaging. Transmission electron microscopy (TEM) imaging results indicate the presence of randomly distributed Ca + Pb nanoscale precipitates. Atom probe tomography (APT) reveals the presence of spherical clusters (4 to 18 nm in size) enriched in radiogenic Pb, Ca, and Si atoms, which, combined with TEM observations, are interpreted as nanoscale inclusions of apatite. In addition to the inclusions, a dislocation enriched in Ca and fluid mobile elements such as Cl, Li, Na, and Mn was imaged from APT data indicating percolating of fluids further than the reaction front. APT 206Pb/238U nanogeochronology indicates that the nanoscale inclusions of apatite formed at 863 ± 28 Ma, 100–150 Ma after crystallisation of the host xenotime, with its formation attributed to fluid metasomatism. This study shows that fluid-xenotime reaction caused Pb* to be redistributed at the nanoscale, recording the timing of metasomatism. However, at the scale of SHRIMP analytical spot (10 μm), xenotime is concordant, indicating that Pb was not mobile at the microscale and fluid-altered xenotime can preserve its crystallisation age. Although the studied grain shows a limited amount of altered domains in BSE imaging, nanoscale analyses reveal a more pervasive re-equilibration of the minerals through the percolation of fluids along dislocations.

Large-Scale Atom Probe Tomography Data Mining: Methods and Application to Inform Hydrogen Behavior.

A large number of atom probe tomography (APT) datasets from past experiments were collected into a database to conduct statistical analyses. An effective way of handling the data is shown, and a study on hydrogen is conducted to illustrate the usefulness of this approach. We propose to handle a large collection of APT spectra as a point cloud and use a city block distance-based metric to measure dissimilarity between spectra. This enables quick and automated searching for spectra by similarity. Since spectra from APT experiments on similar materials are similar, the point cloud of spectra contains clusters. Analysis of these clusters of spectra in this point cloud allows us to infer the sample materials. The behavior of contaminant hydrogen is analyzed and correlated with voltage, electric field, and sample base material. Across several materials, the H2+ /H+ ratio is found to decrease with increasing field, likely an indication of postionization of H2+ ions. The absolute amounts of H2+ and H+ are found to frequently increase throughout APT experiments.

Development of automated tip preparation for atom probe tomography by using script-controlled FIB-SEM

Atom probe tomography (APT) has become a popular technique for microstructural analysis of a wide range of alloys and devices over the past two decades owing to the employment of laser-assisted field evaporation and the development of site-specific tip preparation using a focused ion beam (FIB) with a scanning electron microscopy (SEM) system. In laser-assisted field evaporation, laser irradiation conditions largely influence mass resolution; therefore, recent commercial APT instruments allow strict control of the analysis conditions. However, the mass resolution is affected not only by the laser condition but also by the thermal conductivity of the material and the tip shape. In addition, it is also important to keep the tip shape constant in order to obtain tomography data with good reproducibility since the analytical volume highly depends on the tip shape. In this study, we have developed a method to fabricate the tip with the desired shape automatically by using a script-controlled FIB-SEM system, which has traditionally depended on the skill of the FIB-SEM operator. The tip shape was then intentionally changed by using this method, and its effect on the APT data is also discussed.

Cluster characterization in atom probe tomography: Machine learning using multiple summary functions

In this work, we develop a machine learning-based method to characterize intracluster concentration (ρc), background concentration (ρb), clustering radius (r̄), and radius dispersity (δr) in simulated atom probe tomography data using multiple spatial statistics summary functions to train a Bayesian regularized neural network. We build upon previous work that utilized Ripley’s K-function by incorporating additional features from nearest-neighbor spatial statistics summary functions to better characterize concentration-based metrics. The addition of nearest-neighbor based features allows for highly accurate estimates of ρc and ρb, both with 90% of the predictions within 4.0% of the real value; the root-mean-square errors are reduced by 81.5% and 92.8% from predictions using only K-function based features, respectively. Additionally, including these nearest-neighbor based features improves the ability to differentiate between r̄ and δr.

Crystal plasticity enhances trace element mobility in garnet

Abstract Chemical heterogeneities along grain boundaries in garnet occur across a wide range of metamorphic conditions, yet the processes underlying their development remain poorly understood. Here we integrate electron backscattered diffraction (EBSD) and atom probe tomography (APT) to evaluate the mechanisms driving nanoscale trace element mobility to deformation microstructures in a granulite-facies garnet. This approach shows that low-angle boundaries can be enriched in Ca, Ti, P, Cu, K, Na, Cl, and H. Based on the correlation between EBSD and APT data, we propose that solute ions (Ca, Ti, P, and Cu) were segregated to the interface during the migration of dislocation associated with ductile deformation of the grain. In contrast, elements such as K, Na, Cl, and H are interpreted to reflect diffusion along the low-angle boundary from an externally derived fluid source. These results provide the missing link between chemical heterogeneity and deformation-related microstructures in garnet. Our approach shows that a combination of microstructural and nanoscale geochemical analyses can provide unprecedented insights into mechanisms of element transfer within minerals.

PosgenPy: An Automated and Reproducible Approach to Assessing the Validity of Cluster Search Parameters in Atom Probe Tomography Datasets

Abstract One of the main capabilities of atom probe tomography (APT) is the ability to not only identify but also characterize early stages of precipitation at length scales that are not achievable by other techniques. One of the most popular methods to identify nanoscale clustering in APT data, based on the density-based spatial clustering of applications with noise (DBSCAN), is used extensively in many branches of research. However, it is common that not all of the steps leading to the selection of certain parameters used in the analysis are reported. Without knowing the rationale behind parameter selection, it may be difficult to compare cluster parameters obtained by different researchers. In this work, a simple open-source tool, PosgenPy, is used to justify cluster search parameter selection via providing a systematic sweep through parameter values with multiple randomizations to minimize a false-positive cluster ratio. The tool is applied to several different microstructures: a simulated material system and two experimental datasets from a low-alloy steel . The analyses show how values for the various parameters can be selected to ensure that the calculated cluster number density and cluster composition are accurate.

WITHDRAWN – Improving the Quantification of Deuterium in Zr Alloy Atom Probe Tomography Data

WITHDRAWN – Improving the Quantification of Deuterium in Zr Alloy Atom Probe Tomography Data

Atom Probe Analysis of a Zr-based Bulk Metallic Glass

Abstract Zr-based bulk metallic glasses (BMGs) are amorphous alloys that can exhibit excellent mechanical properties, including high yield strength and fracture toughness. These properties are linked to local microstructural heterogeneities. Whether via microscopy-based techniques, synchrotron techniques, or calorimetric approaches, the amorphous structure of BMGs makes the characterisation of the details of these local structural and chemical heterogeneities extremely challenging. Our focus here is on atom probe tomography (APT), where considerable uncertainty remains in terms of how and when to apply this otherwise powerful technique to amorphous materials. This work reports a systematic evaluation of the experimental parameter space. We report results of BMG composition acquired against various APT operating parameters for Zr63.96Cu13.36Ni10.29Al11.04Nb1.25 (at. %). We demonstrate that a customised peak-based ranging approach yields satisfactory compositional accuracy with absolute errors of <1 at. %. Beyond composition, we have discussed the data quality in terms of attributes of the mass spectra: mass resolution, signal-to-thermal tail ratio, and overlapped peak ratio. We also assess the composition of the well-known clustered evaporation effects, common in APT data of BMGs. We conclude that these regions have negligible differences in composition from the surrounding “matrix” or bulk in these alloys.

Atom Probe Tomography Analysis of Mica.

Laser-assisted atom probe tomography (APT) is a relatively new, powerful technique for sub-nanometric mineral and biomineral analysis. However, the laser-assisted APT analysis of highly anisotropic and chemically diverse minerals, such as phyllosilicates, may prove especially challenging due to the complex interaction between the crystal structure and the laser pulse upon applying a high electric field. Micas are a representative group of nonswelling clay minerals of relevance to a number of scientific and technological fields. In this study, a Mg-rich biotite was analyzed by APT to generate preliminary data on nonisotropic minerals and to investigate the effect of the crystallographic orientation on mica chemical composition and structure estimation. The difference in results obtained for specimens extracted from the (001) and (hk0) mica surfaces indicate the importance of both experimental parameters and the crystallography. Anisotropy of mica has a strong influence on the physicochemical properties of the mineral during field evaporation and the interpretation of APT data. The promising results obtained in the present study open the way to future innovative APT applications on mica and clay minerals and contribute to the general discussion on the challenges for the analysis of geomaterials by atom probe tomography.

Characterization of structure and hardness at aging of the A319 type aluminum alloy with Sn trace addition

The effect of 0.1 wt% (0.02 at%) Sn trace addition on the structure and phase composition of Al-Si-Cu based alloy has been studied using thermodynamic calculations (Thermo-Calc software) and experimental techniques (transmission electron microscopy (TEM), atom probe tomography (APT), hardness and specific electrical conductivity measurements). The Vickers hardness measurement made after aging at 175 °C revealed that the peak hardness of the Sn-containing alloy is about 22% higher than that of the Sn-free alloy (135 HV vs 115 HV). Moreover, the peak hardness is achieved in a much shorter aging time (4 h vs 16 h). Considerably finer precipitates of the θ′ phase (average length less than 60 nm and thickness 3–5 nm) with a substantially higher number density are observed in the Sn-containing alloy. Tiny, rounded particles most of which are associated with the θ′ phase precipitates are also observed. Atom probe tomography analysis confirmed that the observed nanoparticles consist of tin. Quantitative APT analysis revealed that the number density of the copper-containing precipitates is at least twice that of the tin-containing ones (7.6·104 vs 3·104 µm-3). APT data also revealed a clear evidence of noticeable dissolution of Sn (average concentration from 0.05 to 0.10 at%) and silicon (average concentration from 2.2 to 2.8 at%) at the core of the θ′ phase precipitates, which is consistent with the fact that Sn and to a lesser extent Si act as catalysts in the nucleation of this phase. The data obtained using the proxigram technique suggest that Sn atoms show no tendency to segregate at the coherent θ′/(Al) heterophase interface upon aging, whereas localized Sn segregation is observed at the semi-coherent interface. Both Sn-free and Sn-containing alloys in their peak aged states have been subjected to uniaxial compression tests at an elevated temperature of 250 °С. The results suggest that the yield strength of the Sn-containing alloy is about 1.5 times that of the Sn-free alloy (250 MPa vs 175 MPa). The revealed difference is very big and testifies to a high potential of the new alloys as materials for new-generation engine parts.

Bubbles and atom clusters in rock melts: A chicken and egg problem

Bubble and crystal nucleation and growth dynamics play a vital function in determining the eruptive behaviour of a magma. Investigating the initial stages of bubble nucleation and crystallization processes, including their rate and relative timing, can thus give us insight into the eruption parameters of magmas including their rheology. Although many experiments and numerical simulations of decompression-induced vesicle nucleation in magmas have been carried out, the mechanisms of bubble nucleation remain poorly understood. It is widely accepted that there are two ways bubbles can form: either by homogeneous nucleation within a melt, or by heterogeneous nucleation on a surface, interface, or particle. It has been tentatively suggested that what had been interpreted as homogeneous nucleation in magmas was instead heterogeneous nucleation on the surface of elusive nano- or micro-scale crystals present throughout the melt. Herein, we have conducted an experiment whereby bubble nucleation is induced in a synthetic andesitic glass sample using radiation damage from a focused ion beam (FIB) to investigate the initial stages of bubble nucleation. Scanning transmission electron microscopy, electron energy-loss spectroscopy and atom probe tomography (APT) data reveal sub-nanometer scale chemical heterogeneities in the form of Fe-, Na-, and/or Ca-rich clusters at the bubble-glass interface of the newly created bubbles. Comparatively, specimens prepared by cryogenic FIB, in which bubble nucleation is suppressed, are without bubbles; however, APT data shows similar chemical heterogeneities suggesting these clusters existed within the synthetic glass prior to bubble formation. Based on these observations, our data demonstrate the existence of nanoscale chemical heterogeneities within a synthetic silicate glass and at bubble-glass interfaces. Although our experiments do not replicate the conditions of natural processes, our results contribute to the hypothesis that previously described homogeneous nucleation may be re-interpreted as a variety of heterogeneous nucleation. These findings highlight the need to revise the commonly used definition of homogeneous nucleation among magma scientists.

On the fabrication of atom probe tomography specimens of Al alloys at room temperature using focused ion beam milling with liquid Ga ion source.

In this work, a simple rectangular milling technique was demonstrated to prepare needle shape atom probe tomography (APT) specimens from Al alloys by focused-ion-beam (FIB) milling using Ga+ ions at room temperature. Ga has high miscibility in Al owing to which electropolishing technique is preferred over Ga+ ion FIB instruments for the fabrication of APT specimens. Although, site specific sample preparation is not possible by the electropolishing technique. This led to the motivation to demonstrate a new rectangular milling technique using Ga+ FIB instrument that resulted a significant reduction of Ga+ ion impregnation into the specimens. This is attributed to the reduction of milling time (<30 s at 30 kV acceleration voltage) and the use of lower currents (<0.3nA) compared to the conventional annular milling method. The yield of specimens during field evaporation in APT was also significantly increased from around 8 million ions to more than 86 million ions due to the avoidance of Ga+ ion embrittlement. Therefore, the currently demonstrated rectangular milling technique can be used to prepare APT specimens from Al-alloys and obtained accurate compositions of matrix, phases, and hetero-phase interfaces with Ga < 0.1at%. RESEARCH HIGHLIGHTS: Feasibility of using Ga+ ions for the preparation of needle shaped specimens at room temperature from aluminum alloys. Demonstration of using a rectangular milling technique instead of annular milling technique that led to a significant reduction in impregnation of Ga+ ions into the specimen needles. Due to very low Ga+ ion damage, the yield of the APT data increased by 10-12 times.

Advanced quantification of the site-occupancy in ordered multi-component intermetallics using atom probe tomography

A new experimental method is proposed to quantify the site-occupancy of substitutional solute elements in multi-component intermetallics from atom probe tomography data and is applied to the L12 ordered γʹ phase in the top, middle and bottom regions of an electron powder bed fusion produced IN-738LC build. Ti, W and Ta are found to substitute almost exclusively for the β sites. Cr and Mo show mixed behaviour with a higher proportion substituting for the β sites. Co also shows mixed behaviour but with a higher proportion substituting for the α sites. While γʹ maintains an almost constant chemistry and site-preference behaviour throughout the build, the relative site-occupancy ratio is observed to change, particularly for Co and Mo. The results suggest that local changes in thermal history inherent to metal additive manufacturing processes may induce changes to the resultant site-occupancy of γʹ. The method described here improves the experimental quantification of the local atomic site-occupancy, enabling an assessment of the substitutional solute element fractions occurring at the α and β sites in L12 ordered structures. This is important in multi-component intermetallics because of the burgeoning interest in relating the elastic and plastic properties of these structures to their site-occupancy.

Deciphering the Distribution and Crystal-Chemical Environment of Arsenic, Lead, Silica, Phosphorus, Tin, and Zinc in a Porous Ferrihydrite Grain Using Transmission Electron Microscopy and Atom Probe Tomography

The interaction of contaminants and nutrients with soil constituents is controlled by processes in intergranular and intragranular pore spaces of organic matter or/and common secondary minerals such as ferrihydrite, ∼Fe3+10O14(OH)2. This contribution shows that distribution and clustering of the contaminants As, P, Pb, Si, Sn, and Zn in a porous ferrihydrite grain is greatly affected by the heterogeneous size distribution and chemical composition of the pores as well as the ability of their polyhedra to polymerize with the same type of polyhedron. Transmission electron microscopy (TEM) and atom probe tomography (APT) studies are conducted on focused ion beam (FIB) sections extracted from a porous ferrihydrite grain from the smelter-impacted topsoil in Sudbury, Ontario, Canada. The ferrihydrite grain has pore spaces ranging in diameter from tens to hundreds of nanometers. TEM and scanning-TEM studies indicate that the surfaces of the pore walls are enriched in Si. APT data in conjunction with First Near Neighbor (1NN) analyses indicate different degrees of clustering of Pb, As, Sn, Zn, Si, and P within the sample and selected domains. Careful evaluations of 3D atomic plots and 1NN distances indicates the occurrence of polymerized arsenite-, silica-, Sn-, and Zn-polyhedra within pore spaces of the ferrihydrite. Deciphering adsorption, polymerization, and nucleation processes in porous Fe-(hydr)oxides and other soil constituents requires multianalytical approaches and, in this regard, we discuss the advantages and disadvantages of the combination of TEM and APT for characterizing complex environmental samples at the atomic to nanometer scale.