Symposium F.MT05—Advancing Materials Characterization Through Atom Probe Tomography

Fabrication of steel structures requires joining by fusion welding. As requirements for weight and cost savings drive increased demand for advanced high-strength, high-toughness structural steels, the ability to balance mechanical performance and microstructural robustness of steels for the spectrum of welding processes poses a significant challenge. One aspect of this challenge is material responses to reheating during multi-pass welding. Thermal transients induce extensive microstructural changes in prior weld passes, whose nature and magnitude are highly dependent on the specifics of the chosen welding processes and the steel’s composition. Herein, we present a new high-nickel steel, developed at the Naval Surface Warfare Center Carderock Division, which exhibits a positive response to the intrinsic heat treatment imposed during multi-pass welding processes. Because of the non-equilibrium solidification conditions, the weldments exhibit a hierarchical microstructure with several unique features, including: (i) cellular/dendritic substructure due to constitutional super-cooling; (ii) micro-segregation; (iii) high oxygen and hydrogen concentrations; and (iv) different amounts of lath martensite, and retained austenite across the fusion zone. Here, site-specific atom-probe tomography has become an essential tool that enables us to study the different regions of this hierarchical microstructure in great detail. The atom-probe tomography experiments are complemented by differential interference contrast optical microscopy, TEM, X-ray diffraction, and EBSD to study the microstructural features over hierarchical length scales. We have demonstrated that it is possible to produce a fine martensitic microstructure, which is reinforced by a high number density of V(N,C) nanoprecipitates, without post-weld heat treatments, leading to both high-strength and toughness. The presence of the V(N,C) nanoprecipitates has a twofold benefit: (i) provides dispersion strengthening; and (ii) removes nitrogen and diffusible hydrogen from the microstructure, which improves the ductility of the weldments and increases resistance against hydrogen embrittlement. The fundamental knowledge acquired in this study is also utilized to optimize the alloy system for the fabrication of structural components through additive manufacturing processes.

Cr-rich alpha prime (α’) precipitates are detrimental to the mechanical properties of FeCr based ferritic-martensitic steels. Understanding α’ precipitation under irradiation conditions is essential for the application of this steel in nuclear reactor environments. To study the effect of dose rate (ballistic dissolution) and irradiation temperature (radiation-enhanced diffusion) on α’ formation, ultra-high purity Fe-Cr alloys with 18 wt.% Cr (in either solid solution or thermally aged to form pre-existing α’precipitates) were irradiated with 8 MeV Fe ions to a midrange (~1 um) dose of 0.35-3.5 displacements per atom (dpa) between 300-450 °C at 10^-3, 10^-4 and 10^-5 dpa/s. Following irradiation, atom probe tomography (APT) was employed to characterize the number density, radius and Cr concentration of Cr-rich clusters. Homogeneously distributed α’ precipitates were revealed when radiation-enhanced diffusion dominates over ballistic dissolution (>300°C). All clusters are identified and quantified through the maximum separation method. The cluster Cr concentrations decrease as irradiation temperature decreases or dose rate increases. This is consistent with the less solute enrichment revealed by the proxigram across the iso-concentration interface. However, despite the irradiation conditions, all of the measured values are well below the equilibrium Cr concentration for α’ phase suggested by the phase diagram. The radius and number density of α’ clusters are similar and within statistical errors for different specimens. The segregation of Cr, C and N atoms to dislocation loops was also observed.

This research was sponsored by the Office of Fusion Energy Sciences, U.S. Department of Energy under contract DE-AC05-00OR22725 with UT-Battelle, LLC and grant # DE-SC0006661 with the University of Tennessee. The fabrication of the Fe-Cr binary alloys has been carried out within the framework of the EUROfusion Consortium and has received funding from the Euratom research and training program 2019–2020 under Grant Agreement No. 633053.

Focused-ion beam (FIB) target preparation is the gold standard for creating site-specific specimen for atom probe tomography (APT). However, FIB preparation can introduce artifacts, such as contamination, amorphization, and heating. One way to minimize sample damage is the usage of cryo-FIB. Cooling the specimen during FIB milling reduces heating related damage and limits the diffusion of Ga. Cryo-FIB sample preparation as part of a complete cryo-APT workflow enables the study of soft and unstable materials, e.g., organic solar cells or high-mobility hydrogen impurities.
However, FIB lift-out preparation requires modifications to succeed at low temperatures. Conventional techniques for in-situ lift-out use beam induced deposition to temporarily attach the specimen to a manipulator and permanently mount the specimen on a sample support structure. These "soldering" techniques use metal-organic precursor gases injected into the microscope chamber and locally decomposed using the electron or ion beam. Unfortunately, these gas-phase deposition processes do not work at low temperatures because the precursors form thick condensates on all cold surfaces. Alternative techniques, like beam cured adhesives, are also not feasible at cryogenic temperature.
Here, we demonstrate a gas-free cryo-liftout workflow using a low-temperature microgripper. The gripper assembly consists of a 3-axis micromanipulator equipped with an insulated gripper thermally anchored to the microscope's cold shield. Piezo motors position the system and actuate the gripper jaws with nanometer precision. The specimen is mounted inside the microscope on a cryo-stage. The region of interest is partially released from the bulk material using FIB cutting. The gripper grabs the specimen and — after a final release cut — maneuvers it onto a micro-post. After the specimen is attached to the post using redeposition welding, the gripper releases the specimen. In the last preparation step, annular FIB milling gives the sample its final shape.
In contrast to using conventional manipulation, there is no need to temporarily join the probe and specimen and subsequently cut them free. This simplification makes the process faster and more reliable, overall increasing its yield.

Conventional aluminum alloys cannot be utilized in the 250–450 °C temperature range due to a lack of microstructural stability, creating a technological gap currently filled by steels, titanium alloys, and Ni-base superalloys. Replacement of these materials with aluminum alloys has significant implications for the energy and transportation sectors (e.g., increased engine operating temperature and efficiency, lightweighting, cost reduction), but has yet to be achieved. In this talk, we will give examples of APT analysis aiding in the design of new generations of high-temperature, additively manufactured aluminum alloys specifically targeted for use in the 250–450 °C temperature range. In particular, we will show how APT analysis reveals Zr and Mn segregation to the coherent and semi-coherent interfaces of θ’ precipitates in a novel additively manufactured Al-Cu-Zr-Mn alloy, leading to thermal stability up to 350 °C. We will also discuss how the accurate measurement of matrix solute supersaturations facilitated by APT aid in the design of heat treatments for precipitation-strengthened alloys with thermally stable phases and the observations of refined, nanoscale microstructural features formed during the additive manufacturing process. APT was conducted at ORNL's Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility.

The unique ability of Atom probe tomography (APT) to study three-dimensional atomic distribution relies on the field evaporation of ions onto a position-sensitive detector. Multiple ion events reaching the detector within the same pulse period can result in compositional and spatial inaccuracies. Ideally, in APT analysis, acquisition parameters and conditions are modified so that the occurrence of multi-hit events is limited. Some material systems have a higher tendency for multi-hit events during field evaporation in APT, such as those containing boron, carbon, oxygen, nitrogen, and compound semiconductor materials. Under these unavoidable circumstances, these unique events have shown the potential to provide insight on field evaporation and material structure [1, 2]. Here, we define a metric by tracking the distance between coincident detector events within a heterostructure to gain insight on the unique evaporation behavior and crystallographic origin of nitride species. The distributions of pair distances are investigated for each species and region, to understand the influence of elemental composition on evaporated ion separation.

Al_0.3Ga_0.7N/AlN/GaN heterostructures, grown via molecular beam epitaxy (MBE) and metal-organic chemical vapor deposition (MOCVD), were utilized as test structures. Trends in mean and median pair separation distances were consistent between the two preparation methods. Additionally, individual pair species showed agreement within 0.2 mm in detector space. This finding allows for the distinction of Al-x and Ga-x events based on their pair separation. Interestingly, Al-x events show higher separation values than Ga-x events. This is in disagreement with the known lattice constants of Al-N and GaN. This result suggests that crystallographic spacing is not the dominant influence on pair separation. A more likely mechanism is related to the unique field evaporation of the species as well as characteristics of the surface field during acquisition. The crystallographic orientation upon evaporation may also play a role in the resulting pair separation. Distributions of pair separation distance were utilized to investigate variability in the spread of multi-hit events across heterostructure layers. The most prominent feature was the consistent bimodality of Al²⁺-N₂, suggesting two separate evaporation phenomena. This unique behavior may originate from a combination of dissociation events and correlated events occurring for the same species.

This in-depth analysis presents a statistical protocol for interpreting multi-hit events from APT data. The reported findings help define species-specific evaporation trends within a compound semiconductor heterostructure in an aim to expand the application of APT to complex systems. Defining the relation between multi-hit events and structural chemistry will extend the current capabilities of APT to gain insight on atomic spacing that was previously unattainable without the aid of additional characterization methods. Furthermore, this methodology can be applied to other material systems to qualify the likelihood of correlation, potential dissociation trends, and crystallographic nature.


References

1. De Geuser, F., Gault, B., Bostel, A., & Vurpillot, F. (2007). Correlated field evaporation as seen by atom probe tomography. Surface science, 601(2), 536-543.

2. Müller, M., Saxey, D. W., Smith, G. D., & Gault, B. (2011). Some aspects of the field evaporation behaviour of GaSb. Ultramicroscopy, 111(6), 487-492.

Efficient n-type doping in ultra-wide bandgap (Al_xGa_1-x)₂O₃ is a key for the realization of high frequency transistors with theoretical performance limit. Group IV elements (Si, Sn and Ge) are being considered as efficient donors in (Al_xGa_1-x)₂O₃ as they are predicted to substitute in cation sites and contribute in conductivity [1]. First principal calculations by density functional theory (DFT) proposed that Si is the most promising candidate for achieving n-type conductivity in (Al_xGa_1-x)₂O₃ as it remains a shallow donor at the highest range of Al composition while other donors are likely to transform into deep level defects at high Al-(Al_xGa_1-x)₂O₃ [1]. Incorporation of Si in (Al_xGa_1-x)₂O₃ has been reported experimentally. The dopant behavior in these films including dopant distribution, layer homogeneity, segregation/clustering with varying alloy composition is not reported so far. It has been reported for similar wide bandgap system, AlGaN that n-type doping introduces point defects such as cation vacancies (V_Ga and V_Al). These defect sites act as charge trapping centers by interacting with dopants and forms complex defects (V_Ga-Si_Ga or V_Al-Si_Al) [2]. This results in reduced carrier mobility at high doping regime and is termed as “compensating knee” [2]. Similar investigation on dopant behavior in (Al_xGa_1-x)₂O₃ is not reported yet. It is crucial to have in depth understanding of dopant interaction down to atomic level structural chemistry of (Al_xGa_1-x)₂O₃ films is pivotal to achieve excellent doping profile.
Atom probe tomography (APT) was employed to investigate the atomic level dopant behavior and dopant interaction in (Al_xGa_1-x)₂O₃ with varying Al content (x = 0.10-1.0). APT reveals that at low Al content, Si distribution is homogeneous while some high Si density island-like regions appears at high Al content (x80%). The statistical radial distribution function (RDF) was used to understand the dopant interaction in by exploring the nearest neighbor chemistry in the bulk regions of each (Al_xGa_1-x)₂O₃ layer. RDF in layers with low Al contents (x = 10-20%) showed high Ga concentration in regions surrounding the Si atoms with Al concentration being very low, indicating Si is substituting in Ga sites. Ga concentration drops at the position of Si atoms implying the presence of V_Ga. This result indicates the potential V_Ga-Si_Ga defect complex formation. At high Al (x = 60-100%), high Al concentration was observed around Si atoms with Al concentration dropping at the location where Si atoms reside, indicating Al site occupancy with possible V_Al-Si_Al complexes. The initial interpretation of APT results coincided with what was observed in case of GaN, AlGaN and Ga₂O₃ and would account for compensating effect in (Al_xGa_1-x)₂O₃.
The information obtained from APT would provide a guide to the first principal DFT study of Si doped (Al_xGa_1-x)₂O₃. DFT calculation will be performed to confirm the site specific substitution of dopants in (Al_xGa_1-x)₂O₃ at low and high Al content. DFT will confirm the vacancy, interstitial or vacancy-interstitial complex defects by evaluating the formation enthalpy of these defects and defect complexes at different Al contents that would explain carrier compensation in wide bandgap semiconductors.
We acknowledge Dr. Hongping Zhao’s group in Dept. of ECE at the Ohio State University for providing the samples.
[1] J. B. Varley et. al., Appl. Phys. Lett. 116, 172104 (2020)
[2] J. S. Harris et. al., Appl. Phys. Lett. 112, 152101 (2018)

Typical AlCu-alloys used for automotive applications can only withstand temperatures up to 250 ^oC. Increasing this temperature will improve fuel efficiency. The presence of the θ' (Al₂Cu) precipitate in these alloys is critical for the alloy hardness; however, θ' is a metastable precipitate that disappears (transforms) at higher temperatures, destroying the alloy properties. Recently, an affordable and castable Al-Cu alloy that retains θ' and maintains its strength after a lengthy exposure to 350 ^oC was developed (> 200 Hrs). The addition of Mn in combination with Zr and their segregation to precipitate interfaces play a significant role in stabilizing the metastable θ' precipitates, while adding Zr and Mn separately only improves the stability to 200 °C and 300 °C, respectively. A series of atom probe tomography (APT) and scanning transmission electron microscopy (STEM) experiments on Al-Cu-Mn-Zr/Ti-containing alloys subjected to long-term annealing (up to 2,100 h) in the critical temperature range, 300 °C and 350 °C, were performed to reveal how Mn and Zr work synergistically to stabilize the strengthening phase (θ'). The experimentally observed solute segregation to θ' precipitate interfaces (APT) and atomic structure at the interfaces (STEM) were used to guide density functional theory (DFT) simulations to understand how Mn and Zr additions change the energetics of θ' in an Al-matrix. The APT and STEM results reveal that Mn additions reduce interfacial energy, and thus, stabilize θ' long enough for the slower diffusing Zr atoms to segregate to coherent θ' interfaces that eventually create a θ'/ L1₂-Al₃(Zr_x,Ti_1-x) co-precipitate structure. The co-precipitate is highly stable, is a key factor that governs microstructural stability beyond 300 °C. This study reveals how solute additions with different stabilization mechanisms can work in concert to stabilize a desired microstructure, and the results provide insights that can be applied to other high-temperature alloy systems.

This research was supported by the DOE Office of Energy Efficiency and Renewable Energy, Vehicle Technologies Office, Propulsion Materials Program and the DOE Basic Energy Sciences, Materials Sciences and Engineering Division. APT was conducted at ORNL's Center for Nanophase Materials Sciences (CNMS), which is a US DOE Office of Science User Facility. The authors appreciate the support provided by the Oak Ridge Leadership Computing Facility at the ORNL. The authors thank Ray Unocic and Allen Haynes (ORNL) for reviewing the manuscript. The authors would also like to thank Dana McClurg for heat treatments and hardness measurements and James Burns for APT sample preparation and APT operation.

Atom probe tomography (APT) is a powerful characterization technique for obtaining three-dimensional structure and materials composition at the atomic scale and is complementary to transmission electron microscopy (TEM). In tandem, the two techniques provide detailed characterization of structure, composition, and chemistry. However, the full potential of both of these techniques is typically hindered by artifacts from specimen preparation (e.g., Ga implantation and amorphization) and environmental conditions that degrade the specimen’s surface (e.g., surface oxidation). Dual beam FIB (DB-FIB), which is an efficient tool for removing a substantial amount of material and in situ electron beam imaging, allows more control when shaping the APT specimen tip and is typically used to prepare APT specimens. However, Ga-induced damage and amorphization from FIB milling can result in ambiguous results, especially for Al/Al alloys and Ga-containing materials. In fact, it was recently shown that amorphization due to the high energy Ga beam during FIB milling dramatically affected the shape of the mass spectrum due to tip heating during APT acquisitions in Si APT specimens. Studies such as this provoke this study of using low energy Ar ion milling post-FIB milling to obtain high-quality APT specimens. Low energy (< 1 keV) Ar ion milling has been shown to improve TEM specimen quality by removing amorphization and Ga implantation from FIB prepared specimens. We present a post-FIB specimen preparation of APT specimens under controlled environments using concentrated beam (< 1 μm), low energy Ar⁺ milling for the removal of manufacturing damage, oxides and FIB-induced damage for TEM and APT analyses. This controlled environment is integrable to current APT environmentally controlled transfer system.

Al and Mg metals and their alloys were chosen as test specimens due to their highly oxidizing properties. The APT specimens were prepared in a FIB system using standard lift-out methods and annular milling. The APT specimen preparation workflow from FIB to an Ar⁺ milling system using a FIB vacuum transfer system was interfaced with a glovebox in between transfers. For ease of specimen handling and transfer, the half-grid with the APT specimens was secured in the cartridge of a TEM vacuum transfer specimen holder that is compatible with the ion milling system. Ar⁺ milling was performed by using a concentrated ion beam that is rastered longitudinally within a defined area on the APT specimen at decreasing milling energies. TEM, energy dispersive X-ray spectroscopy (EDS), and APT characterization were performed under controlled environments before and after ion milling to determine the removal of surface oxides and FIB-induced damage. Quantification of the removal of FIB-induced damage and the workflow’s ability to prevent specimen oxidation and contamination will be presented.

Recent studies of lithium-ion batteries suggest that the use of binary CoSn alloys as anodes should provide an improvement over currently used anode materials [1]. However, the implementation of CoSn alloys is challenging due to uncertainties regarding the phase transformations within this system. In order to understand these, we evaluate the compositions of different intermetallic compounds produced via the peritectic reactions, nucleate and grow within the microstructure of binary Sn -25 at.% Co by employing atom probe tomography (APT). Owing to its capability, APT with its exceptional capabilities of spatially resolution and quantitative chemical analyses is presented as a sophisticated analytical tool to elucidate the underlying process of the phase transition behaviour of different Co-Sn intermetallic compounds and the local structure of the solid solution. The APT results revealed the presence of Co clusters with (54.1 ± 0.1) at.% Co , (23.34 ± 0.3) at.% Co and (0.6 ± 0.1) at.% Co in pure Sn phase [2], which is consistent with the equilibrium phase diagram that indicates there is limited solubility of Co in Sn (up to 0.6%) at a temperature of 1196 °C . This limited solubility of Co atoms induced the formation of different observed intermetallic phases that are CoSn and CoSn₃. This study demonstrates the unique capabilities of APT to characterize a series of peritectic reactions within the binary Co-Sn alloy. Moreover, crystallization kinetics of the crystallization reactions in Co-Sn alloy system were investigated under non-isothermal as well as isothermal conditions. These investigations were performed by using of model -free (isoconversion) methods. The results show decreasing on the activation energy for the crystallization kinetics for CoSn and CoSn₃ phases with increasing temperature, which reveals the involvement of two processes of nucleation and diffusion during the crystallization of these intermetallic phases.
References

[1] G. Mulas, S. Enzo, C. Bonatto Minella, E. Arca, C. Gerbaldi, N. Penazzi, S. Bodoardo, J. Hassoun and S. Panero, Mechanochemical synthesis and electrochemical properties of nanostructured electrode materials for Li ion batteries. Journal of Solid State Electrochemistry, 2009. 13(2): p. 239-243.
[2] M. Khushaim , F. Alahmari , N. Kattan, D. Chassaing and T. Boll, Microstructural properties and peritectic reactions in a binary Co–Sn alloy by means of scanning electron microscopy and atom probe tomography, Mater. Res. Express 7 (2020) 086508

In this contribution, we review and demonstrate various methods by which we push our characterization limits to better understand nanometric features in the smallest – possibly the most influential – features in nanostructurally engineered materials. Oxide dispersion strengthening of steels for example and, more generally, materials relying on multi-phase strengthening, precipitation hardening, and cluster-strengthening all require an initial catalytic core for the formation of the desired precipitates or mesoscale phases. There is thus a thrust in looking for these pre-precipitation clusters and nano-clouds, which occur in a wide variety of materials research projects. For their analysis we take advantage of correlative techniques between Atom Probe Tomography (APT), Transmission Electron Microscopy, and electron backscatter diffraction, some in combination with modeling & simulation and others utilizing our original cryo-transfer enabling developments.
We will first focus on FeCr alloys, which model ferritic steels relevant for the fusion community. Here we find intriguing details in the cluster and interfacial chemistry in its α/α’ decomposed high-purity binary system [1]. We applied multiple analysis methods to determine their precipitate core compositions, as their structurally coherent boundaries have often shown to be relatively broad compared to other matrix/precipitate systems.
In the field of ferritic steels for advanced nuclear facilities, the nascent irradiation induced clustering and precipitation of nuclear transmutation byproducts is generally investigated to better understand the alloys’ evolution of mechanical properties over extended lifetimes and thus irradiation doses. In our near-atomic-scale study we are able to identify the isotopic specificity and thus the particular ions of interest that influence the steel’s mechanical properties.
Focusing on SmCo high-temperature supermagnets, we are able to determine the geometry, packing density and local chemistry between Zr-rich nano-platelets and Cu-rich cells, which prove to strongly influence the magnetic properties, such as coercivity [2]. Among these phases, even the finest clustering features may have a strong impact on the macroscopic magnetic properties. With respect to light metals, we show results of room-temperature (‘natural’) aging of Al alloys, where we have also investigated further alloy classes in Al and Mg that require specimens to remain at cryogenic temperatures from after their creation – to focused ion beam sharpening of the APT tips – to transferring them to the atom probe and enable interrogating their arrested nanostructures [3,4,5].
These examples demonstrate the impact of near-atomic to nano-scale analyses not only to better understand the initial ‘gatherings’ of atoms (nano-clouds) within a structure, but to stress the continued need for pushing boundaries of highest-resolution characterization methods, the correlation amongst each other, and deploying cryo-vacuum transfer methods.

[1] S. Küchler, V. Vojtech, S.S.A. Gerstl, R.E. Schäublin, J.F. Löffler, submitted to Acta Mater.
[2] L. Pierobon, András Kovács, R.E. Schäublin, S. S. A. Gerstl, J. Caron, U. Wyss, R.E. Dunin-Borkowski, J.F. Löffler, M. Charilaou, arXiv:1901.01922, 2019
[3] S. Pogatscher,1,2 H. Antrekowitsch,2 M. Werinos,2 F. Moszner,1 S. S. A. Gerstl,3 M. F. Francis,4 W. A. Curtin,4 J. F. Loffler,1 and P. J. Uggowitzer, Phys. Rev. Letters 112 (2014) 225701
[4] M. Cihova, R.E. Schäublin, L.B. Hauser, S.S.A. Gerstl, C. Simson, P.J. Uggowitzer, J.F. Loffler, Acta Mater. 158 (2018) 214-229.
[5] R.E. Schäublin, M. Becker, M. Cihova, S.S.A. Gerstl, D. Deiana, C. Hébert, S. Pogatscher, P.J. Uggowitzer and J.F. Löffler, submitted to Acta Mater.

A paradox in APT is that while on the one hand, it provides an unprecedented level of imaging resolution in three dimensions, it is very difficult to obtain an accurate perspective of morphology or shape outlined by atoms of similar chemistry and microstructure. Hence, unlike scattering techniques such as electron microscopy, interfaces appear diffuse, not sharp. This, in turn, makes it challenging to visualize and quantitatively interpret the microstructure at the “meso” scale, where one is interested in the shape and form of the interfaces and their associated chemical gradients. In this presentation we provide an overview of the computational and machine learning approaches to characterize interface chemistry and morphology directly from atom probe images.

CAMECA’s recent innovations in the areas of automated acqusition, reconstruction, and data analyis have greatly increased the throughput and utilizaton of the Local Electrode Atom Probe™ [1]. These advances position the commercial atom probe as a process monitoring tool for nanoscale devices, providing routine characterization for suitable applications for the microelectronics industry.
CAMECA’s Chain Acquisition module enables atom probe data to be acquired from multiple samples, without user intervention, using a recipe-based architecture [2]. Image processing technology is used to do automatic alignment of specimens on a microtip coupon, enabling the user to build a list of specimens that can be analyzed according to individal experimental plans. Live calibration and automated ranging can be perfomed as the instrument is acquiring data, enabling both accurate charge state ratios and reduced user input to mass spectrum assignments. Automated local electrode testing can be performed between analyses, preventing a compromised electrode from affecting the quality of planned experiments, keeping a record of the electrode condtion over time, and notifying a user when an electrode may have decreased in quality.
The experimenal conditions can also be dynamically changed with CAMECA’s Scripted Acqusition module. Triggers can be defined to modify acqusition conditions in response to composition, background level, number of ions, or a variety of other experimental outputs. Scripted Acqusition can be used in conjunction with Chain Acqusition to further increase throughput; for example, the acqusition can be stopped after the region of interest is collected and automatically begin acquisition from the next specimen in the chain. This decreases acquisition overall time and increases the value of the dataset by only collecting the necessary volumes. Project management allows scripts to easily be shared between users so that repeatable experiments can be conducted in laboratories around the world.
In addition to automated acqusitions, CAMECA’s Atom Probe (AP) Suite enables recipe-based reconstructions and analysis. Reconstructions are completed automatically via user-entered parameters, so that the data is ready for review shortly after collection. Automated reconstruction parmeters can also be selected from a pre-existing list of experiment recipes. Further automated analysis can be queued with application of standard or custom analysis recipes applied to the finished reconstructions. This allows the analyst to repeatably extract the most important information from a dataset without manually applying the analysis steps. Together, these advances have made it faster and easier to collect data and reduce the experience requirements for new users to successfully collect and analyze data with atom probe tomography and reduce time-to-knowledge in the experimental process.

References
1.Larson, D. J. et al. Local Electrode Atom Probe Tomography (Springer Publishing, 2013).
2. Reinhard, D. A. et al. Atom Probe Tomography Productivity Enhancements. Microscopy and Microanalysis 25, 522–523 (2019).

Clustering of atoms and molecules significantly impacts material properties and is therefore important to characterize accurately. In this work, we use machine learning and spatial statistics to quantify clustering behavior in simulated and experimental APT data. We base our technique on Ripley's K-function analysis, K(r), which is a sensitive measure of spatial correlation (e.g. clustering). We use features from the K(r) analysis to train a Bayesian regularized neural network (BRNN) to extract radius metrics, intra-cluster density, and background density.

For our test case, training data was simulated for a fixed solute density with clusters of various mean radius, radius variance, intra-particle density, background density, and number of particles. The system was verified on a set of similarly generated data sets, with over 90% of the BRNN-based estimates falling within 10% of the true radius values. Comparing the accuracy of our model against values obtained from the maximum separation algorithm (MSA) in our simulated APT data, we find that the MSA algorithm exhibits large variations based on user-defined inputs while the BRNN model quantifies morphologies accurately. We also apply this analysis to an experimental APT reconstruction with MgZn clusters in a 7000 series aluminum alloy, easily extracting cluster parameters from real APT data.

A unique advantage of APT is its ability to provide precise mass information (i.e. < 1 Da) over volumes of thousands of cubic nanometers with a three-dimensional spatial resolution of < 1 nm. This resolution completely depends on the reconstruction process used to convert the raw data into its final form for analysis: the method must be accurate yet fast enough to work on a typical APT dataset containing millions of ions. Any assessment of spatial signal (e.g. clustering) in APT relies on the fidelity of the reconstruction.

Currently, the vast majority of reconstructions fundamentally rely on point-projection. This method is fast but suffers from spatial resolution inaccuracies—such as ions being placed unphysically close or unphysical voids or variations in the sample shape—because of its assumptions about both the sample geometry and ion trajectories. Furthermore, current reconstruction techniques and analyses derived from them do not have an associated uncertainty because they usually generate only a single, user-directed reconstruction. This makes it laborious to compare different sets of inputs even though it is known that these parameters can cause the reconstructions to vary substantially even when constrained by other knowledge of the sample (e.g. TEM). As a result, APT experiments generally base conclusions off of a single reconstruction without stating a statistical confidence.

We have developed a reconstruction method that provides a solution envelope, increasing confidence in the conclusions drawn from the data regardless of the exact solution. This method does so without needing measurements of the sample other than APT. We explicitly account for the lower x-y resolution of APT by shifting points in x & y, which incorporates this uncertainty into subsequent analyses. This adjustment is determined by solving the well-studied Assignment and Exact Cover Problems for slices of data in detector space; this process can also constrain the reconstruction to maintain density and minimum particle spacing. We may then use the resultant solution sets to determine the robustness of the morphological signal as compared to a null hypothesis of random labeling using the tools of spatial statistics.

We evaluate this method using simulated APT data containing small (0.7 nm diameter) clusters of boron in a random close packed (RCP) matrix of silicon, a system in which there is an approximately 2:1 evaporation field difference that leads to chromatic aberrations. A conventional reconstruction completely loses the clustering signal (p = 0.224), even when optimized to minimize the positioning error with the known input pattern. In comparison, our method still recovers a clear clustering signature across the ensemble of reconstruction solutions (p < 0.001). These results demonstrate that this reconstruction method can significantly improve the accuracy of reconstructions and increase confidence in any morphological conclusions.

Nanosized voids, including pores and bubbles, are a common type of defects that can either improve or degrade the material properties. Voids are essentially empty cavities. Accurate measurement of the material chemistry around voids is important for understanding the void-material interactions. Atom probe tomography (APT) is a promising technique for this task because of its capability to reconstruct 3D positions of elements in materials with a high spatial resolution and chemical sensitivity. However, there has been a knowledge gap regarding how voids are imaged using APT. By performing a correlative scanning transmission electron microscopy (STEM) and APT analysis on void-containing materials, we demonstrated that voids could lead to obvious aberrations in APT reconstructions. Specifically, the local atomic density near voids in an APT dataset can either increase or decrease compared to the matrix. Field evaporation simulations combined with experiments revealed the physical mechanisms for these aberrations and identified the key factors that determined the nature of the aberrations. Based on this study, a general approach for quantifying elemental concentrations near voids using APT has been proposed that has higher accuracy than traditional STEM-based techniques. This approach has been applied to characterize elemental segregations near bubbles in medium and high-entropy alloys (HEAs). The accurate APT measurement shows that in Ni-based HEAs, both Cr and Fe become depleted near bubbles, but Cr depletes more, indicating higher vacancy mobility via Cr. This finding is consistent with the defect energies calculated based on density functional theory and explains the faster bubble growth rate in Ni-based HEAs containing Cr.

Acknowledgment
This work was supported by the Energy Dissipation to Defect Evolution (EDDE) Center, an Energy Frontier Research Center funded by the U.S. Department of Energy (DOE), Office of Science, Basic Energy Sciences under contract number DE-AC05-00OR22725. Electron microscopy and APT were conducted at Oak Ridge National Laboratory's Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility. We acknowledge the financial support of the Region Normandie–FEDER and ANR / EMC3 Labex, DYNAMITE project and support from Semiconductor Research Corporation (SRC) under task ID 2679.001for the field evaporation simulation.

Atom probe tomography (APT) is a powerful tool for 3D materials characterization due to its desirable combination of high spatial resolution and high analytical sensitivity. In state-of the-art laser-pulsed APT, a near ultraviolet (NUV) laser (E ≈ 3.5 eV to 3.6 eV; ≈ 355 nm to 343 nm) incident on the specimen provides the thermal energy to overcome the activation barrier for field ion evaporation. NUV laser-pulsed APT has been used to successfully analyze a wide range of materials including metals, semiconductors, insulators, biological materials, and even liquids. However, the data quality is often degraded due to artifacts from the thermal pulsing mechanism. These various artifacts manifest in the form of increased uncertainty in the quantification of specimen composition from the time-of-flight mass spectrometry measurement.

Photon energies in the extreme ultraviolet region of the electromagnetic spectrum (E ≈ 10 eV to 100 eV; ≈ 124 nm to 12 nm) are significantly higher than those in the near ultraviolet and as such, have the potential to trigger entirely new ionization and desorption mechanisms that are not possible with the much lower photon energies in the NUV band. These include various athermal ionization pathways and/or highly local heating that is entirely unlike that achieved using NUV sources.

Initial results from the world’s first EUV (41.85 eV; λ = 29.6 nm) radiation-pulsed atom probe microscope will be presented. This instrument uses wavelength-tunable femtosecond-pulsed coherent EUV radiation from phase-matched high harmonic generation in a hollow-core capillary waveguide. EUV radiation-pulsed field ion emission in a variety of materials types will be discussed with respect to the quality of the mass spectrometry data and compared to results obtained using NUV sources. Specifically, the measured composition, time-independent background, time-delayed evaporation, charge state ratios, and multiple hit counts will be compared and related.

This project was made possible with material support from CAMECA Instruments through a Cooperative Research and Development Agreement.

Heat treatment of materials to improve performance has been employed for thousands of years. Control of the final size, composition, and distribution of the material’s phases is of paramount importance and can be influenced by small changes in initial compositions, temperature, and annealing time. Therefore, knowing what occurs at the earliest stages of the transformation process would shed insight into how to modify the materials for better control of their properties. It has been only within the last one hundred years that the characterization technologies for probing the nanometer-scale mechanisms behind the improved material performance resulting from heat treatments have been developed. While these tools have immensely expanded our understanding of materials and the mechanisms acting on them during thermal processing, there are still aspects that remain challenging to evaluate. In particular, sensitive quantitative compositional measurement with the combination of nano-scale spatial resolution and fractions of a second temporal resolution would provide needed insights into the earliest steps in diffusional processes.
Presented here is a demonstration of how the laser in a laser-pulsed atom probe instrument has been used to thermally induce crystallization and subsequently probe the resulting material structure for a Fe-Si-Nb-Cu-B alloy. Methods of estimating the laser induced temperature profile, determining the cumulative annealing time, and evaluating the resultant three-dimensional nanoscale crystallization are explored. Correlated transmission electron microscopy imaging and diffraction are used to supplement the atom probe results. The distribution of copper-rich cluster sizes and cluster densities measured by atom probe tomography are compared with Thermo-Calc modeling results. Through this example, advantages and limitations of using the atom probe as a processing tool are evaluated.

Metallic surfaces undergo a series of surface and subsurface structural and chemical transformations that drastically change their surface and mechanical properties. Understanding such dynamics from a fundamental science standpoint is particularly important in the case of heterogeneous catalysis, as such fundamental understanding is needed to build rational links between chemical/structural surface properties and the desired catalytic performance. This work addresses the dynamics of O₂ hydrogenation over rhodium (Rh), platinum (Pt) and PtRh alloy (10% wt Rh) revealed with Field Emission/Ion Microscopy (FIM/FEM). A comparative study is conducted with the use of other oxygen carrier gases: N₂O, CO and CO₂. Finally, Atom Probe Tomography (APT) is used to map the path of the oxygen atoms on and through the Rh surface.
FIM and FEM are capable to image the apex of sharp needles with nanometric lateral resolution. At the nanoscale, this apex is in the form of a hemisphere with a variety of exposed atomic planes and facets that can be simultaneously probed. Moorphology of such sample mimicks a single metal nanoparticle by its size and shape. FEM exhibits image contrast which is mainly determined by crystallographic plane-dependent variations of the work function. Local brightness variations on video-FEM is used to probe dynamic surface reactions¹. The APT in turn provides a 3D atom-by-atom map of the entire apex volume with near atomic scale spatial resolution and part per million chemical sensitivity. By mapping the 3D distribution of Rh oxide and the morphology of the oxide/metal interface within the bulk after a controlled O₂(g) exposure, APT is used to explore the nature of oxygen atom penetration in the bulk.
Crystallographically-dependent chemical and structural dynamics of Rh and Pt surface oxidation during an ongoing oxygen exposure is directly observed with FIM/FEM. Focusing on Rh, we observe a mechanism of inter-facet cooperation between Rh{012} and Rh{113} crystallographic surface facets. Clean Rh surface (i.e. no oxide) exhibits bright contrast over {012} facets while {113} facets remain dark. However, upon exposure to oxygen originating from O₂, N₂, or CO₂, we observe simultaneously, an increase brightness over the {113} facets and a decrease in brightness over the {012} facets. This tells us that dissociative adsorption of O_ads occurs on the {012} and is followed by oxygen penetration into the subsurface through the {113}. Similar dynamics are observed on Pt and PtRh with exposure to other oxygen carrier gases of N₂O and CO₂. Additionally, the addition of H₂ recovers the original FEM pattern reflecting the reaction between H_ads and O_ads. This mechanism of inter-facet cooperation leads to a unique distribution of oxygen atoms within the bulk, which can be directly observed by APT. Following a 600s exposure of O₂ (2.0×10^-2Pa) at 700K of a pristine Rh needle (i.e. no native surface oxide), APT analysis reveals an anisotropic distribution of Rh oxides (RhO_x) and the complex morphology of the oxide/metal interface reflecting a specific crystallographic dependence². An enhanced density of RhO_Xⁿ⁺ (x = 1 or 2 and n = 1 or 2) species are found both under the (111) central plane and along only one of the two three-fold symmetric crystal zones comprising the {113} facets. An iso-density surface of the APT reconstruction reveals a central accumulation of RhO_X species along the [111] direction, connected to the {113} vicinal facets. In summary, we conclude O_ads occurs on the {012} and is followed by oxygen penetration into the subsurface through the {113} towards the [111].
The combination of FEM and APT analysis delivers new insight into the Rh oxide formation and offers exciting perspectives leading a better and more fundamental understanding of the heterogenous catalysis as well as corrosion dynamics.

¹Lambeets, S.V. et al. , J.Phys. Chem. C 121, 16238 (2017)
²Lambeets, S.V. et al. , J.Phys. Chem. Lett. 11, 3144 (2020)

Estimations of evaporation field values in atom probe tomography are sparse in the literature despite their importance in the reconstruction and data analysis process. This work describes a method for estimating the zero-barrier evaporation field (ZBEF) using a straightforward measurement series of voltage versus laser pulse energy at a constant evaporation rate. This estimate depends on the sample radius of curvature and its specific heat. If a similar measurement is made of the voltage versus base temperature for a given evaporation rate, direct extraction of the material's specific heat can be made, leaving only the sample radius of curvature as an input parameter.

The method is applied to extract ZBEF in a previously published voltage versus laser pulse energy dataset for CdTe (18.07(87) V/nm).[1] Furthermore, using the published voltage versus base temperature sweep of CdTe permits extraction of a specific heat (11.27(254) J/K/mol @ 23.1 K) in good agreement with the literature (11.14 J/K/mol @ 22.17 K).[2] This method is then applied to a number of organic small molecules to understand how ZBEF depends on molecular properties.

[1] Diercks & Gorman, J. Phys. Chem. C, 119(35), 20623–20631 (2015).
[2] Birch, J. Phys. C, 8(13), 2043–2047 (1975).

In the methanol-to-hydrocarbons (MTH) process, methanol is converted to more valuable products, such as olefins and aromatics, using zeolite or zeotype catalysts. Methanol can be obtained from conventional as well as from more renewable sources, such as biomass, carbon containing waste and CO₂, which explains the great commercial interest of this catalytic process. [1][2] The exact end-product of the process is determined by the zeolite pore structure and the number and nature of the acidic active sites. A key step in this MTH mechanism is the formation of a so-called hydrocarbon pool in which olefins and aromatic compounds function as activating and deactivating reaction intermediates. [3] Most of the unmodified proton form zeolites are extensively studied for their performance and deactivation in the MTH process. However, protons can be exchanged by other cations, which can drastically influence catalytic performance. [3] These influences on the catalytic behaviour are nonetheless not well understood.

In this work, magnesium was introduced in the chabazite structure SSZ-13 to (partially) replace the protons in the zeolite framework. This resulted in differences in both catalytic performance as well as catalyst deactivation. Using operando UV-Vis spectroscopy and operando X-ray diffraction, differences between reaction intermediates and catalyst characteristics can be linked to the catalytic performance.

Zeolites are notoriously difficult to study at the nano-scale, because of their instability under e.g. electron beams. [4] Atom probe tomography (APT) is uniquely positioned among all tomographic techniques; it can provide 3D chemical information with sub-nm resolution, making it the only technique capable of finding nanometre scale relationships in these materials, and especially for studying elements that offer no significant z-contrast differences (e.g. Al and Si). With this technique we were able to visualize magnesium clusters, obtain information about the distribution of the magnesium with respect to the aluminium, as well as the correlation between magnesium and coke molecules after the MTH process.

To conclude, in this work both operando as well as APT studies on (un)-modified zeolites are described to be able to contribute to the unravelling of the MTH mechanism in the presence of e.g. Mg.

References
[1] Ji, Y.; Deimund, et al. ACS Catal. 2015, 5, 4456–4465.
[2] Vogt, C.; et al. Nat. Catal. 2019, 2, 188–197.
[3] Yarulina, I.; et al. Nat. Catal. 2018, 1, 398–411
[4] Schmidt, J. E.; et al. Angew. Chem. Int. Ed. 2018, 57, 10422–10435

Yttria-stabilized tetragonal zirconia polycrystal (Y-TZP) is of great interest due to its superior corrosion, wear-resistance, and high flexural strength. Y-TZP exhibits a unique toughening mechanism when a stimulus is applied, in which it transforms from tetragonal (t) to monoclinic (m) phase. Over time, this transformation can lead to an aging phenomenon known as low-temperature degradation (LTD)[1]. The aging process is influenced by microstructural features, such as porosity, impurities, grain size, and stabilizer content. These features are also interlinked with one another, making the route for process optimization less clear. Grain boundaries (GBs) are thought to serve as the nucleation site for t-m transformation as well as the path for transformation propagation [2,3]. In an attempt to strengthen the GBs and limit LTD, commercial processing methods have introduced dopants and varying stabilizers to the zirconia microstructure. For every pair of yttria ions in Y-TZP, one oxygen vacancy is introduced (to neutralize the excess charge of Y³⁺)[4]. The annihilation of oxygen vacancies is widely accepted as the root cause of hydrothermal aging in Y-TZP; however, the underlying mechanism and its connection to the destabilization of the tetragonal phase is not fully understood. In this work, we will combine atomic-level insights on chemical distribution from atom probe tomography (APT) with graph-based materials representation to quantify the extent of elemental heterogeneity influence on microstructural features linked with the t-m transformation.

APT analysis was performed on a set of commercial 3Y-TZP ceramics (3 mol.% Y₂O₃), revealing trace levels of Aluminum (Al), Hafnium (Hf), and Silicon (Si), which were introduced in the raw material and during processing steps. The three-dimensional atom maps from APT demonstrated Y-rich GBs, with a limited number of GBs exhibiting preferential segregation of Al impurities, as well. Concentration profiles taken perpendicularly across the GBs indicated 5 at.% Y and 2 at.% Al within the GB and ~1.5 at.% Y throughout the matrix. APT analysis provides the three-dimensional positions and chemical identity for each atom. The positional data from regions of interest were converted to a graph representation, allowing for the definition and extraction of descriptors to understand the underlying physical phenomena. For example, the nearest neighbor distance between ions of the same species was used to define a descriptor that helps identify the number and distribution of phases present within the sample. Another descriptor was based on the shortest path consisting of Al to help assess the propensity of diffusion. These methods were further extended to understand the likelihood of an O vacancy in a particular region.

Our combined approach utilizes rigorous materials characterization techniques and informatics to quantify critical variations in microstructure related to the t-m transformation. Through graph-based representation, we define new metrics to correlate the microstructure with potential routes for degradation. These understandings will aid in the improved design and processing of zirconia-based ceramics to limit the susceptibility to LTD aging while maintaining their valuable mechanical properties. Our method enhances the capabilities of APT data to quantify the underlying phenomena of material degradation.

References:
1. Chevalier, J. et al., Acta Biomaterialia, 7(7), 2986-2993 (2011).
2. Mecartney, M. L., & Rühle, M., Acta metallurgica, 37(7), 1859-1863 (1989).
3. Chevalier, J., Biomaterials, 27(4), 535-543 (2006).
4. Krishnamurthy R., et al., Journal of the American Ceramic Society 87(10), 1821-1830 (2004).
5. Acknowledgement: APT was conducted at ORNL's Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility.

Quantification of carbon in materials using Atom Probe Tomography (APT) is a common issue. Large underestimation in the carbon content is often observed [1,2]. One well known reason arises from multiple events that occur when several ions strike the detector nearly simultaneously and close to each other. When this happens, a part of ions may be lost due to the so-called electronic “dead time” and “dead zone” existing between each hit. This phenomenon is very likely when analyzing carbon because the high electric field needed to induce its evaporation promotes correlated departures of carbon ions at the tip surface. Dissociation of molecular ions during their flight from the tip to the detector is another source of carbon multiple events.

Carbon biases can also arise from mass-to-charge ratio overlaps in APT mass spectra. For example, (¹²C¹²C)²⁺ and ¹²C⁺ ions are overlapped as they share the same mass-to-charge ratio (12 Da). In this case, the proportion of each overlapped ion is needed to make a correct quantification. These quantities can sometimes be derived using isotopes abundance. In the above example, the true number of (¹²C¹²C)²⁺ counts at 12 Da can be derived from (¹²C¹³C)²⁺ counts at 12.5 Da. However, since the stable isotope ¹³C only account for 1.07% of the total carbon amount, ions containing¹³C are not systematically detected. This issue arises in materials containing a low amount of carbon (less than 1%), for instance in the matrix of steels or in semicondcutors.

The overlap issue faced in materials containing a low carbon content was addressed in this study using (¹²C)⁺ and (¹³C)⁺ implanted in equal proportion. Doses, energies and tilts used to introduce (¹²C)⁺ and (¹³C)⁺ in a 300 mm silicon wafer were 1 ± 0.01 x 10¹⁵ cm^-2, 30 keV and 3° respectively. The wafer was then annealed to restore silicon crystallinity (750°C for 5 hours). Carbon solubility in slicon being very low (~2 ppb at 750°C ) compared to the implanted content (~0.2 at. %), carbon clusters were found by APT. Most peaks were indeed related to molecular ions originating from these clusters. Statistics was successfully used to identify these peaks and quantify the contribution of molecular ions to the carbon content.

Carbon quantification using APT was found more accurate using low electric fields conditions (high laser powers applied at the tip surface). This was attributed to the hindering of molecular ion dissociation at low electric fields. Furthermore, the atomic fraction of carbon in clusters was found close to that of the expected SiC phase, only when using low electric field conditions (high laser power). This gives a mean number of trapped self-interstitials per carbon atom close to one, in agreement with other studies [3,4].

[1] Thuvander, Mattias, et al. "Quantitative atom probe analysis of carbides." Ultramicroscopy 111.6 (2011): 604-608.
[2] Estivill, Robert, et al. "Quantitative analysis of Si/SiGeC superlattices using atom probe tomography”. Ultramicroscopy 159 (2015): 223-231.
[3] Cristiano et al. "Interstitial trapping efficiency of c+ implanted into preamorphised silicon—control of EOR defects." Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 127 (1997): 22-26.
[4] Cacciato et al. "Dislocation formation and B transient diffusion in C coimplanted Si." Journal of applied physics 79.5 (1996): 2314-2325.

Perovskite metal oxides (ABO₃) are highly promising for artificial superlattice fabrications due to their exceptional physical properties and chemical stabilities. Among perovskite metal oxide superlattice, SrTiO₃ (STO) and LaMnO₃ (LMO) based heterostructures has gained significant attention due to enhanced magnetic properties compared to bulk LMO, which is desirable for spintronic devices. Previous scanning transmission electron microscopy (STEM) results demonstrated that the oxygen octahedral rotation (the angle between Mn-O-Mn bonding) at the STO/LMO interface plays a vital role in tuning the magnetic moments in the LMO/STO superlattice. Diffusion of elements across these interfaces can also play a role in the magnetic properties; however, diffusion across these interfaces has not been well documented. STEM has been heavily used to study these materials because it is ideal for observing interfacial structure, but a small amount of interfacial elemental diffusion is difficult to quantify using analytical STEM techniques, such as electron energy loss spectroscopy (EELS).
Atom probe tomography (APT) is a good candidate for quantifying the chemistry through these interfaces because it is a sub-nm resolution technique with a maximum sensitivity of ~10 ppm. To this end, APT was employed to map the interfacial chemistry of the STO/LMO superlattice structure. Two different superlattice structures were tested to isolate the top and bottom STO/LMO interfaces. STO/LMO/LASO (5x) and LASO/LMO/STO (5x) superstructures were grown by atomic layer deposition (ALD) on STO supstrates to isolate LMO grown on top of STO (LMO→STO) and STO grown on top of LMO (STO→LMO). The saturation magnetic moment per Mn ion was significantly higher for LMO→STO (M_S~1.67 µ_B per Mn ion) compared to STO→LMO (M_S~1.23 µ_B per Mn ion). Other samples grown in the same manner showed the same behavior. The APT results reveal differences in interfacial elemental diffusion that can explain the observed differences in magnetic behavior. In this presentation, the impact of growth order on elemental diffusion using ALD and how elemental diffusion influences magnetic properties will be discussed. This study provides a direct insight on how changes in interfacial structure and chemistry of LMO/STO superlattice can modify the subsequent magnetic functionalities when the growth sequence is reversed. This information is crucial for understand the ALD growth technique of complex oxides and how to optimize ferromagnetism in future spintronic devices. APT was conducted at ORNL's Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility.

Oxide Dispersion Strengthened (ODS) steels are promising candidates for future nuclear reactors, partly due to the fine and dense dispersion of the nanoparticles they contain. Until now, there was no consensus as to the nature of the nanoparticles, because their analysis pushed the techniques to their limits and in consequence, introduced some artefacts.
Atom probe tomography (APT) analysis of an ODS steel reveal a dense dispersion (10²³–10²⁴ m^-3) of small (≈ 2 nm) nanoparticles enriched in Y, Ti and O. There are also numerous evidences of APT artefacts: high amount of matrix atoms in the particles (i.e. Fe and Cr), reconstructed morphology variations and so atomic density fluctuations, all depending of particles size. It appears that the current understanding of APT artefacts cannot explain all these observations. According to field evaporation simulations, composition and morphological biases result from a cross-over of ions trajectories. It then appears that the particles are pures in Y, Ti and O and therefore all the Fe and Cr atoms, observed in the particles (> 80 at.%), are coming for trajectories aberrations. This conclusion is supported by an analytical model (CCC), that can be applied to other materials, and by complementary experimental analysis (TEM measurements on extractive replica). Moreover, the proportion of Fe and Cr, artificially introduced in the particles, provides the composition of the Cr rich shell around the particles.
From this observation, a literature review shows a significant dispersion of the APT measurement in Y, Ti and O in the particles (considering only those elements), which makes precise characterization difficult. Several factors may be at the origin of this dispersion: the ODS steel (14YWT, MA957, 12YWT, MKCR…), the particles size, the data treatment protocol (and so the user), the APT used (ECOWATAP, LAWATAP, LEAP…) and its analysis conditions (pulse type, pulse fraction, temperature…). One of the objectives is to understand this measurement dispersion by limiting itself to the influence of the APT used. To do this, selected ODS steels was analyzed on different APT, available to the Material Physics Group and to the Norwegian University of Science and Technology and the data was processed by a single user with a quasi-automatic data treatment protocol.

Nickel superalloys are employed in the hottest part of jet engines and industrial gas turbines for power generation. In these harsh environments, sulphur contamination can occur in-service, causing reductions in the predicted component lifetime. Our work uses environmental exposures and state of the art characterisation to explore the effect of low-level silicon (0.5 wt%) and manganese (1 wt%) additions on the sulphidation and type-2 hot corrosion response of a polycrystalline nickel superalloy.

Silicon was added promote formation of a compact dual-layer chromia-alumina scale that reduced oxidation rate in long-term experiments. The silicon-containing alloy was more resistant to hot corrosion from coupled NaCl + SOx exposure.

Manganese improved oxidation resistance by creating MnCr2O4, which reduced the oxidation rate but did not change the oxide scale morphology. The presence of Mn (a known sulphur scavenger) proved less effective than silicon, reducing the sulphidation and corrosion damage extent but not the overall depth.

When it comes to structures, steel has an ability like no other material to provide outstanding and versatile combinations of strength and toughness, but also to boast with durability and cost efficiency. The properties of steel, as in most structural materials, are attainable by selection of the composition and adjustment of thermomechanical sequence and heat treatment. The migration and final distribution of solute atoms into the iron lattice and its inherent defects determines the performance of the end product. Atom probe tomography (APT) allows to image and quantify nano-scale phenomena relevant to steel metallurgy that have been known to exist for long but could not be directly observed, such as the first stages of tempering in carbon steels or solute decoration of dislocations, the so-known as Cottrell atmospheres. From the viewpoint of the properties, APT helps to understand in deep detail the mechanisms leading to precipitation strengthening, improved ductility, grain boundary pinning, toughness maximization or deterioration and grain boundary embrittlement, among others.
Steel is fun to play with due to its infinite combinations of interstitial and substitutional alloying elements, the various equilibrium and non-equilibrium reactions of the matrix and its complex and sometimes abundant lattice defect substructure. We will review here application cases of APT in steel product development carried out by ArcelorMittal in collaboration with leading institutions.

Combining mechanical properties with chemical information on the nanometer scale have become a key approach in recent characterization techniques of metals and their alloys. The application of advanced high-resolution characterization techniques such as local electrode atom probe tomography enables a detailed investigation of the material's interfaces on an atomic level. Molybdenum has advantageous structural as well as functional properties leading to a broad range of applications. As a structural material, however, the application of technically pure molybdenum at low temperatures is limited because of its tendency for brittle intergranular failure. The direct correlation of strength and ductility with segregation content at the boundaries significantly helps to understand the intergranular fracture of molybdenum. The deliberate introduction of cohesion enhancing elements lead to a change in fracture mode, a lower brittle-to-ductile transition temperature and a higher ductility of the material. In this work, the site-specific preparation of the region of interest for atom probe measurement is performed using focused ions and so-called transmission Kikuchi diffraction. The grain boundaries of molybdenum alloys, which are showing a fracture mode change, are analyzed and compared to pure molybdenum. An outlook for improved ductility of molybdenum focusing on segregated elements is given.

It is well known that environmental exposure of materials can lead to alteration, such as for example, the formation of an unintended surface oxide layer, that in some cases can affect the intended analysis. To prevent unwanted morphological or chemical alteration of specimens prior to analysis, control of the specimen environment during transfer between instruments is required for specimens sensitive to air or thermal exposure. At PNNL, we have developed a suite of hardware and experimental protocols for the preparation and handling of environmentally sensitive materials centered around the combination of a unique environmental transfer hub chamber, modified specimen suitcase device, and modified cryo FIB/SEM capabilities. The hub chamber design also allows the connection of ancillary tools, such as a gas phase reactor chamber for controlled surface reaction studies and UHV specimen transfer. The ability to prepare, manipulate, and transfer specimens under either cryogenic or vacuum conditions provides a means to uniquely apply APT analysis to address new science domains in biological, nuclear, and catalytic materials. Here I will describe in detail the design of our system and our developed workflows focusing on specific examples of APT analysis applied to temperature sensitive material systems requiring cryo specimen preparation and transfer, as well as air sensitive materials systems requiring vacuum or inert vacuum transfer.

The presence of hydrogen in steels can lead to catastrophic embrittlement/early-fracture. This is a serious issue for hydrogen transportation and storage. However, consensus has not been reached on the exact mechanism of hydrogen embrittlement, mainly due to the difficulty to provide direct evidence of the hydrogen-materials interactions that underpins the hypotheses [1]. In addition, a proposed solution to hydrogen embrittlement by using steels that contain hydrogen traps such as carbide precipitates [2], is limited in its effectiveness, due to the inability to directly observe the proposed hydrogen trapping at microstructural features.
As such, we used atom probe to study the hydrogen distribution at key features, including dislocations [3], grain boundaries [3], and both incoherent [3] and coherent [4] carbide precipitates in BCC/BCT iron matrix. To enable these studies, we charged the sample with deuterium (a hydrogen isotope) to avoid ambiguity from background hydrogen, and utilised a cryogenic sample transfer protocol to allow sufficient signal to be retained for observations. These efforts lead to the confirmations of: i) hydrogen enrichment at dislocations, providing a concrete validation of the hydrogen-enhanced dislocation mobility theory of embrittlement; ii) hydrogen enrichment at grain boundaries, underpinning the hydrogen-enhanced grain boundary decohesion theory; iii) the hydrogen at the interface between large, incoherent precipitates and the surrounding steel matrix, settling a long-standing debate around whether hydrogen trapping is an interfacial effect; and iv) the hydrogen at the interior of small, coherent carbides, suggesting hydrogen can internalise into carbides under certain conditions.
Reference:
[1] I. M. Robertson, et al. Metall. Mater. Trans. A 46a(6), 2323-2341 (2015)
[2] H. K. D. H. Bhadeshia, ISIJ Int. 56, 24-36 (2016)
[3] Y.-S. Chen et al., Science 367, 171-175 (2020)
[4] Y.-S. Chen et al., Science 355, 1196-1199 (2017)

Stress corrosion cracking (SCC) is an important degradation mechanism in a number of key engineering applications, including water-cooled nuclear reactors. The mechanism behind SCC requires a susceptible material, a local stress and a corrosive environment, but the local behaviour and chemistry at the crack tip is complex and still not fully understood. For improved understanding of the SCC mechanism, it is necessary to better characterise the mechanical behaviour of the material during the formation of cracks, as well as the local chemistry at the crack tip. Atom probe tomography (APT) has been used previously to image the chemical composition of the grain boundary ahead of a crack tip in SUS316 stainless steel, showing that segregation of Ni plays an important role in sensitising the grain boundary to corrosive attack [1]. The high-speed atomic force microscope (HS-AFM) is a contact mode form of atomic force microscopy that can image the topography of surfaces at frames-per-second acquisition rates, enabling live images of chemical processes. Recent work with HS-AFM has observed real-time formation of cracks in 20Cr-25Ni-Nb stabilised stainless steel [2] and AISI Type 304 stainless steel [3] in a corrosive solution known to accelerate SCC. The ability to watch a grain boundary corrode and crack at video frame rates is a powerful tool to link SCC behaviour to the underlying mechanisms behind the process. In this study, the atomic-scale chemical characterisation capability of APT is utilised on crack tips of a sensitised AISI type 304 stainless steel that was observed cracking in an in-situ HS-AFM experiment in an aggressive thiosulphate solution. An APT liftout was obtained from a region close to the crack tip using a specimen preparation method adapted from that used to lift out site-specific grain boundaries in silicon [4]. This preparation method was able to site the crack interface routinely within the APT region of interest, and APT data of the metal-oxide interface showed is a depletion of Cr and Ni enrichment at the boundary, with a thin porous oxide rich in sodium. This was confirmed by focused ion beam cross-sectioning showing large voids at grain boundaries below the surface with a thin Na-rich oxide only at the edges of the voids, indicating a cracking mechanism dominated by the stress. The preparation process will be detailed, as well as correlative data between the two techniques to further the understanding of SCC behaviour in steels.

[1] M. Meisnar, M.P. Moody & S. Lozano-Perez, Atom probe tomography of stress corrosion crack tips in SUS316 stainless steel, Corrosion Science 98, pp 661-671, 2015
[2] S. Moore et al, A study of dynamic nanoscale corrosion initiation events using HS-AFM, Faraday Discussions, 210, 2018
[3] A Laferrere et al, In situ imaging of corrosion processes in nuclear fuel cladding, Corros. Eng., Sci. Technol., 52 , 596 —604, 2017
[4] C Lotharukpong et al, Specimen preparation methods for elemental characterisation of grain boundaries and isolated dislocations in multicrystalline silicon using atom probe tomography, Materials Characterisation, 131, pp 472-479, 2017

CAMECA’s recent innovations in the areas of automated acqusition, reconstruction, and data analyis have greatly increased the throughput and utilizaton of the Local Electrode Atom Probe™ [1]. These advances position the commercial atom probe as a process monitoring tool for nanoscale devices, providing routine characterization for suitable applications for the microelectronics industry.
CAMECA’s Chain Acquisition module enables atom probe data to be acquired from multiple samples, without user intervention, using a recipe-based architecture [2]. Image processing technology is used to do automatic alignment of specimens on a microtip coupon, enabling the user to build a list of specimens that can be analyzed according to individal experimental plans. Live calibration and automated ranging can be perfomed as the instrument is acquiring data, enabling both accurate charge state ratios and reduced user input to mass spectrum assignments. Automated local electrode testing can be performed between analyses, preventing a compromised electrode from affecting the quality of planned experiments, keeping a record of the electrode condtion over time, and notifying a user when an electrode may have decreased in quality.
The experimenal conditions can also be dynamically changed with CAMECA’s Scripted Acqusition module. Triggers can be defined to modify acqusition conditions in response to composition, background level, number of ions, or a variety of other experimental outputs. Scripted Acqusition can be used in conjunction with Chain Acqusition to further increase throughput; for example, the acqusition can be stopped after the region of interest is collected and automatically begin acquisition from the next specimen in the chain. This decreases acquisition overall time and increases the value of the dataset by only collecting the necessary volumes. Project management allows scripts to easily be shared between users so that repeatable experiments can be conducted in laboratories around the world.
In addition to automated acqusitions, CAMECA’s Atom Probe (AP) Suite enables recipe-based reconstructions and analysis. Reconstructions are completed automatically via user-entered parameters, so that the data is ready for review shortly after collection. Automated reconstruction parmeters can also be selected from a pre-existing list of experiment recipes. Further automated analysis can be queued with application of standard or custom analysis recipes applied to the finished reconstructions. This allows the analyst to repeatably extract the most important information from a dataset without manually applying the analysis steps. Together, these advances have made it faster and easier to collect data and reduce the experience requirements for new users to successfully collect and analyze data with atom probe tomography and reduce time-to-knowledge in the experimental process.

References
1.Larson, D. J. et al. Local Electrode Atom Probe Tomography (Springer Publishing, 2013).
2. Reinhard, D. A. et al. Atom Probe Tomography Productivity Enhancements. Microscopy and Microanalysis 25, 522–523 (2019).

A paradox in APT is that while on the one hand, it provides an unprecedented level of imaging resolution in three dimensions, it is very difficult to obtain an accurate perspective of morphology or shape outlined by atoms of similar chemistry and microstructure. Hence, unlike scattering techniques such as electron microscopy, interfaces appear diffuse, not sharp. This, in turn, makes it challenging to visualize and quantitatively interpret the microstructure at the “meso” scale, where one is interested in the shape and form of the interfaces and their associated chemical gradients. In this presentation we provide an overview of the computational and machine learning approaches to characterize interface chemistry and morphology directly from atom probe images.

Clustering of atoms and molecules significantly impacts material properties and is therefore important to characterize accurately. In this work, we use machine learning and spatial statistics to quantify clustering behavior in simulated and experimental APT data. We base our technique on Ripley's K-function analysis, K(r), which is a sensitive measure of spatial correlation (e.g. clustering). We use features from the K(r) analysis to train a Bayesian regularized neural network (BRNN) to extract radius metrics, intra-cluster density, and background density.

For our test case, training data was simulated for a fixed solute density with clusters of various mean radius, radius variance, intra-particle density, background density, and number of particles. The system was verified on a set of similarly generated data sets, with over 90% of the BRNN-based estimates falling within 10% of the true radius values. Comparing the accuracy of our model against values obtained from the maximum separation algorithm (MSA) in our simulated APT data, we find that the MSA algorithm exhibits large variations based on user-defined inputs while the BRNN model quantifies morphologies accurately. We also apply this analysis to an experimental APT reconstruction with MgZn clusters in a 7000 series aluminum alloy, easily extracting cluster parameters from real APT data.

Heat treatment of materials to improve performance has been employed for thousands of years. Control of the final size, composition, and distribution of the material’s phases is of paramount importance and can be influenced by small changes in initial compositions, temperature, and annealing time. Therefore, knowing what occurs at the earliest stages of the transformation process would shed insight into how to modify the materials for better control of their properties. It has been only within the last one hundred years that the characterization technologies for probing the nanometer-scale mechanisms behind the improved material performance resulting from heat treatments have been developed. While these tools have immensely expanded our understanding of materials and the mechanisms acting on them during thermal processing, there are still aspects that remain challenging to evaluate. In particular, sensitive quantitative compositional measurement with the combination of nano-scale spatial resolution and fractions of a second temporal resolution would provide needed insights into the earliest steps in diffusional processes.
Presented here is a demonstration of how the laser in a laser-pulsed atom probe instrument has been used to thermally induce crystallization and subsequently probe the resulting material structure for a Fe-Si-Nb-Cu-B alloy. Methods of estimating the laser induced temperature profile, determining the cumulative annealing time, and evaluating the resultant three-dimensional nanoscale crystallization are explored. Correlated transmission electron microscopy imaging and diffraction are used to supplement the atom probe results. The distribution of copper-rich cluster sizes and cluster densities measured by atom probe tomography are compared with Thermo-Calc modeling results. Through this example, advantages and limitations of using the atom probe as a processing tool are evaluated.

Nanosized voids, including pores and bubbles, are a common type of defects that can either improve or degrade the material properties. Voids are essentially empty cavities. Accurate measurement of the material chemistry around voids is important for understanding the void-material interactions. Atom probe tomography (APT) is a promising technique for this task because of its capability to reconstruct 3D positions of elements in materials with a high spatial resolution and chemical sensitivity. However, there has been a knowledge gap regarding how voids are imaged using APT. By performing a correlative scanning transmission electron microscopy (STEM) and APT analysis on void-containing materials, we demonstrated that voids could lead to obvious aberrations in APT reconstructions. Specifically, the local atomic density near voids in an APT dataset can either increase or decrease compared to the matrix. Field evaporation simulations combined with experiments revealed the physical mechanisms for these aberrations and identified the key factors that determined the nature of the aberrations. Based on this study, a general approach for quantifying elemental concentrations near voids using APT has been proposed that has higher accuracy than traditional STEM-based techniques. This approach has been applied to characterize elemental segregations near bubbles in medium and high-entropy alloys (HEAs). The accurate APT measurement shows that in Ni-based HEAs, both Cr and Fe become depleted near bubbles, but Cr depletes more, indicating higher vacancy mobility via Cr. This finding is consistent with the defect energies calculated based on density functional theory and explains the faster bubble growth rate in Ni-based HEAs containing Cr.

Acknowledgment
This work was supported by the Energy Dissipation to Defect Evolution (EDDE) Center, an Energy Frontier Research Center funded by the U.S. Department of Energy (DOE), Office of Science, Basic Energy Sciences under contract number DE-AC05-00OR22725. Electron microscopy and APT were conducted at Oak Ridge National Laboratory's Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility. We acknowledge the financial support of the Region Normandie–FEDER and ANR / EMC3 Labex, DYNAMITE project and support from Semiconductor Research Corporation (SRC) under task ID 2679.001for the field evaporation simulation.

Atom probe tomography (APT) is a powerful tool for 3D materials characterization due to its desirable combination of high spatial resolution and high analytical sensitivity. In state-of the-art laser-pulsed APT, a near ultraviolet (NUV) laser (E ≈ 3.5 eV to 3.6 eV; ≈ 355 nm to 343 nm) incident on the specimen provides the thermal energy to overcome the activation barrier for field ion evaporation. NUV laser-pulsed APT has been used to successfully analyze a wide range of materials including metals, semiconductors, insulators, biological materials, and even liquids. However, the data quality is often degraded due to artifacts from the thermal pulsing mechanism. These various artifacts manifest in the form of increased uncertainty in the quantification of specimen composition from the time-of-flight mass spectrometry measurement.

Photon energies in the extreme ultraviolet region of the electromagnetic spectrum (E ≈ 10 eV to 100 eV; ≈ 124 nm to 12 nm) are significantly higher than those in the near ultraviolet and as such, have the potential to trigger entirely new ionization and desorption mechanisms that are not possible with the much lower photon energies in the NUV band. These include various athermal ionization pathways and/or highly local heating that is entirely unlike that achieved using NUV sources.

Initial results from the world’s first EUV (41.85 eV; λ = 29.6 nm) radiation-pulsed atom probe microscope will be presented. This instrument uses wavelength-tunable femtosecond-pulsed coherent EUV radiation from phase-matched high harmonic generation in a hollow-core capillary waveguide. EUV radiation-pulsed field ion emission in a variety of materials types will be discussed with respect to the quality of the mass spectrometry data and compared to results obtained using NUV sources. Specifically, the measured composition, time-independent background, time-delayed evaporation, charge state ratios, and multiple hit counts will be compared and related.

This project was made possible with material support from CAMECA Instruments through a Cooperative Research and Development Agreement.

Metallic surfaces undergo a series of surface and subsurface structural and chemical transformations that drastically change their surface and mechanical properties. Understanding such dynamics from a fundamental science standpoint is particularly important in the case of heterogeneous catalysis, as such fundamental understanding is needed to build rational links between chemical/structural surface properties and the desired catalytic performance. This work addresses the dynamics of O₂ hydrogenation over rhodium (Rh), platinum (Pt) and PtRh alloy (10% wt Rh) revealed with Field Emission/Ion Microscopy (FIM/FEM). A comparative study is conducted with the use of other oxygen carrier gases: N₂O, CO and CO₂. Finally, Atom Probe Tomography (APT) is used to map the path of the oxygen atoms on and through the Rh surface.
FIM and FEM are capable to image the apex of sharp needles with nanometric lateral resolution. At the nanoscale, this apex is in the form of a hemisphere with a variety of exposed atomic planes and facets that can be simultaneously probed. Moorphology of such sample mimicks a single metal nanoparticle by its size and shape. FEM exhibits image contrast which is mainly determined by crystallographic plane-dependent variations of the work function. Local brightness variations on video-FEM is used to probe dynamic surface reactions¹. The APT in turn provides a 3D atom-by-atom map of the entire apex volume with near atomic scale spatial resolution and part per million chemical sensitivity. By mapping the 3D distribution of Rh oxide and the morphology of the oxide/metal interface within the bulk after a controlled O₂(g) exposure, APT is used to explore the nature of oxygen atom penetration in the bulk.
Crystallographically-dependent chemical and structural dynamics of Rh and Pt surface oxidation during an ongoing oxygen exposure is directly observed with FIM/FEM. Focusing on Rh, we observe a mechanism of inter-facet cooperation between Rh{012} and Rh{113} crystallographic surface facets. Clean Rh surface (i.e. no oxide) exhibits bright contrast over {012} facets while {113} facets remain dark. However, upon exposure to oxygen originating from O₂, N₂, or CO₂, we observe simultaneously, an increase brightness over the {113} facets and a decrease in brightness over the {012} facets. This tells us that dissociative adsorption of O_ads occurs on the {012} and is followed by oxygen penetration into the subsurface through the {113}. Similar dynamics are observed on Pt and PtRh with exposure to other oxygen carrier gases of N₂O and CO₂. Additionally, the addition of H₂ recovers the original FEM pattern reflecting the reaction between H_ads and O_ads. This mechanism of inter-facet cooperation leads to a unique distribution of oxygen atoms within the bulk, which can be directly observed by APT. Following a 600s exposure of O₂ (2.0×10^-2Pa) at 700K of a pristine Rh needle (i.e. no native surface oxide), APT analysis reveals an anisotropic distribution of Rh oxides (RhO_x) and the complex morphology of the oxide/metal interface reflecting a specific crystallographic dependence². An enhanced density of RhO_Xⁿ⁺ (x = 1 or 2 and n = 1 or 2) species are found both under the (111) central plane and along only one of the two three-fold symmetric crystal zones comprising the {113} facets. An iso-density surface of the APT reconstruction reveals a central accumulation of RhO_X species along the [111] direction, connected to the {113} vicinal facets. In summary, we conclude O_ads occurs on the {012} and is followed by oxygen penetration into the subsurface through the {113} towards the [111].
The combination of FEM and APT analysis delivers new insight into the Rh oxide formation and offers exciting perspectives leading a better and more fundamental understanding of the heterogenous catalysis as well as corrosion dynamics.

¹Lambeets, S.V. et al. , J.Phys. Chem. C 121, 16238 (2017)
²Lambeets, S.V. et al. , J.Phys. Chem. Lett. 11, 3144 (2020)

Conventional aluminum alloys cannot be utilized in the 250–450 °C temperature range due to a lack of microstructural stability, creating a technological gap currently filled by steels, titanium alloys, and Ni-base superalloys. Replacement of these materials with aluminum alloys has significant implications for the energy and transportation sectors (e.g., increased engine operating temperature and efficiency, lightweighting, cost reduction), but has yet to be achieved. In this talk, we will give examples of APT analysis aiding in the design of new generations of high-temperature, additively manufactured aluminum alloys specifically targeted for use in the 250–450 °C temperature range. In particular, we will show how APT analysis reveals Zr and Mn segregation to the coherent and semi-coherent interfaces of θ’ precipitates in a novel additively manufactured Al-Cu-Zr-Mn alloy, leading to thermal stability up to 350 °C. We will also discuss how the accurate measurement of matrix solute supersaturations facilitated by APT aid in the design of heat treatments for precipitation-strengthened alloys with thermally stable phases and the observations of refined, nanoscale microstructural features formed during the additive manufacturing process. APT was conducted at ORNL's Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility.

Yttria-stabilized tetragonal zirconia polycrystal (Y-TZP) is of great interest due to its superior corrosion, wear-resistance, and high flexural strength. Y-TZP exhibits a unique toughening mechanism when a stimulus is applied, in which it transforms from tetragonal (t) to monoclinic (m) phase. Over time, this transformation can lead to an aging phenomenon known as low-temperature degradation (LTD)[1]. The aging process is influenced by microstructural features, such as porosity, impurities, grain size, and stabilizer content. These features are also interlinked with one another, making the route for process optimization less clear. Grain boundaries (GBs) are thought to serve as the nucleation site for t-m transformation as well as the path for transformation propagation [2,3]. In an attempt to strengthen the GBs and limit LTD, commercial processing methods have introduced dopants and varying stabilizers to the zirconia microstructure. For every pair of yttria ions in Y-TZP, one oxygen vacancy is introduced (to neutralize the excess charge of Y³⁺)[4]. The annihilation of oxygen vacancies is widely accepted as the root cause of hydrothermal aging in Y-TZP; however, the underlying mechanism and its connection to the destabilization of the tetragonal phase is not fully understood. In this work, we will combine atomic-level insights on chemical distribution from atom probe tomography (APT) with graph-based materials representation to quantify the extent of elemental heterogeneity influence on microstructural features linked with the t-m transformation.

APT analysis was performed on a set of commercial 3Y-TZP ceramics (3 mol.% Y₂O₃), revealing trace levels of Aluminum (Al), Hafnium (Hf), and Silicon (Si), which were introduced in the raw material and during processing steps. The three-dimensional atom maps from APT demonstrated Y-rich GBs, with a limited number of GBs exhibiting preferential segregation of Al impurities, as well. Concentration profiles taken perpendicularly across the GBs indicated 5 at.% Y and 2 at.% Al within the GB and ~1.5 at.% Y throughout the matrix. APT analysis provides the three-dimensional positions and chemical identity for each atom. The positional data from regions of interest were converted to a graph representation, allowing for the definition and extraction of descriptors to understand the underlying physical phenomena. For example, the nearest neighbor distance between ions of the same species was used to define a descriptor that helps identify the number and distribution of phases present within the sample. Another descriptor was based on the shortest path consisting of Al to help assess the propensity of diffusion. These methods were further extended to understand the likelihood of an O vacancy in a particular region.

Our combined approach utilizes rigorous materials characterization techniques and informatics to quantify critical variations in microstructure related to the t-m transformation. Through graph-based representation, we define new metrics to correlate the microstructure with potential routes for degradation. These understandings will aid in the improved design and processing of zirconia-based ceramics to limit the susceptibility to LTD aging while maintaining their valuable mechanical properties. Our method enhances the capabilities of APT data to quantify the underlying phenomena of material degradation.

References:
1. Chevalier, J. et al., Acta Biomaterialia, 7(7), 2986-2993 (2011).
2. Mecartney, M. L., & Rühle, M., Acta metallurgica, 37(7), 1859-1863 (1989).
3. Chevalier, J., Biomaterials, 27(4), 535-543 (2006).
4. Krishnamurthy R., et al., Journal of the American Ceramic Society 87(10), 1821-1830 (2004).
5. Acknowledgement: APT was conducted at ORNL's Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility.

It is well known that environmental exposure of materials can lead to alteration, such as for example, the formation of an unintended surface oxide layer, that in some cases can affect the intended analysis. To prevent unwanted morphological or chemical alteration of specimens prior to analysis, control of the specimen environment during transfer between instruments is required for specimens sensitive to air or thermal exposure. At PNNL, we have developed a suite of hardware and experimental protocols for the preparation and handling of environmentally sensitive materials centered around the combination of a unique environmental transfer hub chamber, modified specimen suitcase device, and modified cryo FIB/SEM capabilities. The hub chamber design also allows the connection of ancillary tools, such as a gas phase reactor chamber for controlled surface reaction studies and UHV specimen transfer. The ability to prepare, manipulate, and transfer specimens under either cryogenic or vacuum conditions provides a means to uniquely apply APT analysis to address new science domains in biological, nuclear, and catalytic materials. Here I will describe in detail the design of our system and our developed workflows focusing on specific examples of APT analysis applied to temperature sensitive material systems requiring cryo specimen preparation and transfer, as well as air sensitive materials systems requiring vacuum or inert vacuum transfer.

The presence of hydrogen in steels can lead to catastrophic embrittlement/early-fracture. This is a serious issue for hydrogen transportation and storage. However, consensus has not been reached on the exact mechanism of hydrogen embrittlement, mainly due to the difficulty to provide direct evidence of the hydrogen-materials interactions that underpins the hypotheses [1]. In addition, a proposed solution to hydrogen embrittlement by using steels that contain hydrogen traps such as carbide precipitates [2], is limited in its effectiveness, due to the inability to directly observe the proposed hydrogen trapping at microstructural features.
As such, we used atom probe to study the hydrogen distribution at key features, including dislocations [3], grain boundaries [3], and both incoherent [3] and coherent [4] carbide precipitates in BCC/BCT iron matrix. To enable these studies, we charged the sample with deuterium (a hydrogen isotope) to avoid ambiguity from background hydrogen, and utilised a cryogenic sample transfer protocol to allow sufficient signal to be retained for observations. These efforts lead to the confirmations of: i) hydrogen enrichment at dislocations, providing a concrete validation of the hydrogen-enhanced dislocation mobility theory of embrittlement; ii) hydrogen enrichment at grain boundaries, underpinning the hydrogen-enhanced grain boundary decohesion theory; iii) the hydrogen at the interface between large, incoherent precipitates and the surrounding steel matrix, settling a long-standing debate around whether hydrogen trapping is an interfacial effect; and iv) the hydrogen at the interior of small, coherent carbides, suggesting hydrogen can internalise into carbides under certain conditions.
Reference:
[1] I. M. Robertson, et al. Metall. Mater. Trans. A 46a(6), 2323-2341 (2015)
[2] H. K. D. H. Bhadeshia, ISIJ Int. 56, 24-36 (2016)
[3] Y.-S. Chen et al., Science 367, 171-175 (2020)
[4] Y.-S. Chen et al., Science 355, 1196-1199 (2017)

In this contribution, we review and demonstrate various methods by which we push our characterization limits to better understand nanometric features in the smallest – possibly the most influential – features in nanostructurally engineered materials. Oxide dispersion strengthening of steels for example and, more generally, materials relying on multi-phase strengthening, precipitation hardening, and cluster-strengthening all require an initial catalytic core for the formation of the desired precipitates or mesoscale phases. There is thus a thrust in looking for these pre-precipitation clusters and nano-clouds, which occur in a wide variety of materials research projects. For their analysis we take advantage of correlative techniques between Atom Probe Tomography (APT), Transmission Electron Microscopy, and electron backscatter diffraction, some in combination with modeling & simulation and others utilizing our original cryo-transfer enabling developments.
We will first focus on FeCr alloys, which model ferritic steels relevant for the fusion community. Here we find intriguing details in the cluster and interfacial chemistry in its α/α’ decomposed high-purity binary system [1]. We applied multiple analysis methods to determine their precipitate core compositions, as their structurally coherent boundaries have often shown to be relatively broad compared to other matrix/precipitate systems.
In the field of ferritic steels for advanced nuclear facilities, the nascent irradiation induced clustering and precipitation of nuclear transmutation byproducts is generally investigated to better understand the alloys’ evolution of mechanical properties over extended lifetimes and thus irradiation doses. In our near-atomic-scale study we are able to identify the isotopic specificity and thus the particular ions of interest that influence the steel’s mechanical properties.
Focusing on SmCo high-temperature supermagnets, we are able to determine the geometry, packing density and local chemistry between Zr-rich nano-platelets and Cu-rich cells, which prove to strongly influence the magnetic properties, such as coercivity [2]. Among these phases, even the finest clustering features may have a strong impact on the macroscopic magnetic properties. With respect to light metals, we show results of room-temperature (‘natural’) aging of Al alloys, where we have also investigated further alloy classes in Al and Mg that require specimens to remain at cryogenic temperatures from after their creation – to focused ion beam sharpening of the APT tips – to transferring them to the atom probe and enable interrogating their arrested nanostructures [3,4,5].
These examples demonstrate the impact of near-atomic to nano-scale analyses not only to better understand the initial ‘gatherings’ of atoms (nano-clouds) within a structure, but to stress the continued need for pushing boundaries of highest-resolution characterization methods, the correlation amongst each other, and deploying cryo-vacuum transfer methods.

[1] S. Küchler, V. Vojtech, S.S.A. Gerstl, R.E. Schäublin, J.F. Löffler, submitted to Acta Mater.
[2] L. Pierobon, András Kovács, R.E. Schäublin, S. S. A. Gerstl, J. Caron, U. Wyss, R.E. Dunin-Borkowski, J.F. Löffler, M. Charilaou, arXiv:1901.01922, 2019
[3] S. Pogatscher,1,2 H. Antrekowitsch,2 M. Werinos,2 F. Moszner,1 S. S. A. Gerstl,3 M. F. Francis,4 W. A. Curtin,4 J. F. Loffler,1 and P. J. Uggowitzer, Phys. Rev. Letters 112 (2014) 225701
[4] M. Cihova, R.E. Schäublin, L.B. Hauser, S.S.A. Gerstl, C. Simson, P.J. Uggowitzer, J.F. Loffler, Acta Mater. 158 (2018) 214-229.
[5] R.E. Schäublin, M. Becker, M. Cihova, S.S.A. Gerstl, D. Deiana, C. Hébert, S. Pogatscher, P.J. Uggowitzer and J.F. Löffler, submitted to Acta Mater.

A unique advantage of APT is its ability to provide precise mass information (i.e. < 1 Da) over volumes of thousands of cubic nanometers with a three-dimensional spatial resolution of < 1 nm. This resolution completely depends on the reconstruction process used to convert the raw data into its final form for analysis: the method must be accurate yet fast enough to work on a typical APT dataset containing millions of ions. Any assessment of spatial signal (e.g. clustering) in APT relies on the fidelity of the reconstruction.

Currently, the vast majority of reconstructions fundamentally rely on point-projection. This method is fast but suffers from spatial resolution inaccuracies—such as ions being placed unphysically close or unphysical voids or variations in the sample shape—because of its assumptions about both the sample geometry and ion trajectories. Furthermore, current reconstruction techniques and analyses derived from them do not have an associated uncertainty because they usually generate only a single, user-directed reconstruction. This makes it laborious to compare different sets of inputs even though it is known that these parameters can cause the reconstructions to vary substantially even when constrained by other knowledge of the sample (e.g. TEM). As a result, APT experiments generally base conclusions off of a single reconstruction without stating a statistical confidence.

We have developed a reconstruction method that provides a solution envelope, increasing confidence in the conclusions drawn from the data regardless of the exact solution. This method does so without needing measurements of the sample other than APT. We explicitly account for the lower x-y resolution of APT by shifting points in x & y, which incorporates this uncertainty into subsequent analyses. This adjustment is determined by solving the well-studied Assignment and Exact Cover Problems for slices of data in detector space; this process can also constrain the reconstruction to maintain density and minimum particle spacing. We may then use the resultant solution sets to determine the robustness of the morphological signal as compared to a null hypothesis of random labeling using the tools of spatial statistics.

We evaluate this method using simulated APT data containing small (0.7 nm diameter) clusters of boron in a random close packed (RCP) matrix of silicon, a system in which there is an approximately 2:1 evaporation field difference that leads to chromatic aberrations. A conventional reconstruction completely loses the clustering signal (p = 0.224), even when optimized to minimize the positioning error with the known input pattern. In comparison, our method still recovers a clear clustering signature across the ensemble of reconstruction solutions (p < 0.001). These results demonstrate that this reconstruction method can significantly improve the accuracy of reconstructions and increase confidence in any morphological conclusions.

In the methanol-to-hydrocarbons (MTH) process, methanol is converted to more valuable products, such as olefins and aromatics, using zeolite or zeotype catalysts. Methanol can be obtained from conventional as well as from more renewable sources, such as biomass, carbon containing waste and CO₂, which explains the great commercial interest of this catalytic process. [1][2] The exact end-product of the process is determined by the zeolite pore structure and the number and nature of the acidic active sites. A key step in this MTH mechanism is the formation of a so-called hydrocarbon pool in which olefins and aromatic compounds function as activating and deactivating reaction intermediates. [3] Most of the unmodified proton form zeolites are extensively studied for their performance and deactivation in the MTH process. However, protons can be exchanged by other cations, which can drastically influence catalytic performance. [3] These influences on the catalytic behaviour are nonetheless not well understood.

In this work, magnesium was introduced in the chabazite structure SSZ-13 to (partially) replace the protons in the zeolite framework. This resulted in differences in both catalytic performance as well as catalyst deactivation. Using operando UV-Vis spectroscopy and operando X-ray diffraction, differences between reaction intermediates and catalyst characteristics can be linked to the catalytic performance.

Zeolites are notoriously difficult to study at the nano-scale, because of their instability under e.g. electron beams. [4] Atom probe tomography (APT) is uniquely positioned among all tomographic techniques; it can provide 3D chemical information with sub-nm resolution, making it the only technique capable of finding nanometre scale relationships in these materials, and especially for studying elements that offer no significant z-contrast differences (e.g. Al and Si). With this technique we were able to visualize magnesium clusters, obtain information about the distribution of the magnesium with respect to the aluminium, as well as the correlation between magnesium and coke molecules after the MTH process.

To conclude, in this work both operando as well as APT studies on (un)-modified zeolites are described to be able to contribute to the unravelling of the MTH mechanism in the presence of e.g. Mg.

References
[1] Ji, Y.; Deimund, et al. ACS Catal. 2015, 5, 4456–4465.
[2] Vogt, C.; et al. Nat. Catal. 2019, 2, 188–197.
[3] Yarulina, I.; et al. Nat. Catal. 2018, 1, 398–411
[4] Schmidt, J. E.; et al. Angew. Chem. Int. Ed. 2018, 57, 10422–10435

Perovskite metal oxides (ABO₃) are highly promising for artificial superlattice fabrications due to their exceptional physical properties and chemical stabilities. Among perovskite metal oxide superlattice, SrTiO₃ (STO) and LaMnO₃ (LMO) based heterostructures has gained significant attention due to enhanced magnetic properties compared to bulk LMO, which is desirable for spintronic devices. Previous scanning transmission electron microscopy (STEM) results demonstrated that the oxygen octahedral rotation (the angle between Mn-O-Mn bonding) at the STO/LMO interface plays a vital role in tuning the magnetic moments in the LMO/STO superlattice. Diffusion of elements across these interfaces can also play a role in the magnetic properties; however, diffusion across these interfaces has not been well documented. STEM has been heavily used to study these materials because it is ideal for observing interfacial structure, but a small amount of interfacial elemental diffusion is difficult to quantify using analytical STEM techniques, such as electron energy loss spectroscopy (EELS).
Atom probe tomography (APT) is a good candidate for quantifying the chemistry through these interfaces because it is a sub-nm resolution technique with a maximum sensitivity of ~10 ppm. To this end, APT was employed to map the interfacial chemistry of the STO/LMO superlattice structure. Two different superlattice structures were tested to isolate the top and bottom STO/LMO interfaces. STO/LMO/LASO (5x) and LASO/LMO/STO (5x) superstructures were grown by atomic layer deposition (ALD) on STO supstrates to isolate LMO grown on top of STO (LMO→STO) and STO grown on top of LMO (STO→LMO). The saturation magnetic moment per Mn ion was significantly higher for LMO→STO (M_S~1.67 µ_B per Mn ion) compared to STO→LMO (M_S~1.23 µ_B per Mn ion). Other samples grown in the same manner showed the same behavior. The APT results reveal differences in interfacial elemental diffusion that can explain the observed differences in magnetic behavior. In this presentation, the impact of growth order on elemental diffusion using ALD and how elemental diffusion influences magnetic properties will be discussed. This study provides a direct insight on how changes in interfacial structure and chemistry of LMO/STO superlattice can modify the subsequent magnetic functionalities when the growth sequence is reversed. This information is crucial for understand the ALD growth technique of complex oxides and how to optimize ferromagnetism in future spintronic devices. APT was conducted at ORNL's Center for Nanophase Materials Sciences (CNMS), which is a U.S. DOE Office of Science User Facility.

Nickel superalloys are employed in the hottest part of jet engines and industrial gas turbines for power generation. In these harsh environments, sulphur contamination can occur in-service, causing reductions in the predicted component lifetime. Our work uses environmental exposures and state of the art characterisation to explore the effect of low-level silicon (0.5 wt%) and manganese (1 wt%) additions on the sulphidation and type-2 hot corrosion response of a polycrystalline nickel superalloy.

Silicon was added promote formation of a compact dual-layer chromia-alumina scale that reduced oxidation rate in long-term experiments. The silicon-containing alloy was more resistant to hot corrosion from coupled NaCl + SOx exposure.

Manganese improved oxidation resistance by creating MnCr2O4, which reduced the oxidation rate but did not change the oxide scale morphology. The presence of Mn (a known sulphur scavenger) proved less effective than silicon, reducing the sulphidation and corrosion damage extent but not the overall depth.

Oxide Dispersion Strengthened (ODS) steels are promising candidates for future nuclear reactors, partly due to the fine and dense dispersion of the nanoparticles they contain. Until now, there was no consensus as to the nature of the nanoparticles, because their analysis pushed the techniques to their limits and in consequence, introduced some artefacts.
Atom probe tomography (APT) analysis of an ODS steel reveal a dense dispersion (10²³–10²⁴ m^-3) of small (≈ 2 nm) nanoparticles enriched in Y, Ti and O. There are also numerous evidences of APT artefacts: high amount of matrix atoms in the particles (i.e. Fe and Cr), reconstructed morphology variations and so atomic density fluctuations, all depending of particles size. It appears that the current understanding of APT artefacts cannot explain all these observations. According to field evaporation simulations, composition and morphological biases result from a cross-over of ions trajectories. It then appears that the particles are pures in Y, Ti and O and therefore all the Fe and Cr atoms, observed in the particles (> 80 at.%), are coming for trajectories aberrations. This conclusion is supported by an analytical model (CCC), that can be applied to other materials, and by complementary experimental analysis (TEM measurements on extractive replica). Moreover, the proportion of Fe and Cr, artificially introduced in the particles, provides the composition of the Cr rich shell around the particles.
From this observation, a literature review shows a significant dispersion of the APT measurement in Y, Ti and O in the particles (considering only those elements), which makes precise characterization difficult. Several factors may be at the origin of this dispersion: the ODS steel (14YWT, MA957, 12YWT, MKCR…), the particles size, the data treatment protocol (and so the user), the APT used (ECOWATAP, LAWATAP, LEAP…) and its analysis conditions (pulse type, pulse fraction, temperature…). One of the objectives is to understand this measurement dispersion by limiting itself to the influence of the APT used. To do this, selected ODS steels was analyzed on different APT, available to the Material Physics Group and to the Norwegian University of Science and Technology and the data was processed by a single user with a quasi-automatic data treatment protocol.

When it comes to structures, steel has an ability like no other material to provide outstanding and versatile combinations of strength and toughness, but also to boast with durability and cost efficiency. The properties of steel, as in most structural materials, are attainable by selection of the composition and adjustment of thermomechanical sequence and heat treatment. The migration and final distribution of solute atoms into the iron lattice and its inherent defects determines the performance of the end product. Atom probe tomography (APT) allows to image and quantify nano-scale phenomena relevant to steel metallurgy that have been known to exist for long but could not be directly observed, such as the first stages of tempering in carbon steels or solute decoration of dislocations, the so-known as Cottrell atmospheres. From the viewpoint of the properties, APT helps to understand in deep detail the mechanisms leading to precipitation strengthening, improved ductility, grain boundary pinning, toughness maximization or deterioration and grain boundary embrittlement, among others.
Steel is fun to play with due to its infinite combinations of interstitial and substitutional alloying elements, the various equilibrium and non-equilibrium reactions of the matrix and its complex and sometimes abundant lattice defect substructure. We will review here application cases of APT in steel product development carried out by ArcelorMittal in collaboration with leading institutions.

Stress corrosion cracking (SCC) is an important degradation mechanism in a number of key engineering applications, including water-cooled nuclear reactors. The mechanism behind SCC requires a susceptible material, a local stress and a corrosive environment, but the local behaviour and chemistry at the crack tip is complex and still not fully understood. For improved understanding of the SCC mechanism, it is necessary to better characterise the mechanical behaviour of the material during the formation of cracks, as well as the local chemistry at the crack tip. Atom probe tomography (APT) has been used previously to image the chemical composition of the grain boundary ahead of a crack tip in SUS316 stainless steel, showing that segregation of Ni plays an important role in sensitising the grain boundary to corrosive attack [1]. The high-speed atomic force microscope (HS-AFM) is a contact mode form of atomic force microscopy that can image the topography of surfaces at frames-per-second acquisition rates, enabling live images of chemical processes. Recent work with HS-AFM has observed real-time formation of cracks in 20Cr-25Ni-Nb stabilised stainless steel [2] and AISI Type 304 stainless steel [3] in a corrosive solution known to accelerate SCC. The ability to watch a grain boundary corrode and crack at video frame rates is a powerful tool to link SCC behaviour to the underlying mechanisms behind the process. In this study, the atomic-scale chemical characterisation capability of APT is utilised on crack tips of a sensitised AISI type 304 stainless steel that was observed cracking in an in-situ HS-AFM experiment in an aggressive thiosulphate solution. An APT liftout was obtained from a region close to the crack tip using a specimen preparation method adapted from that used to lift out site-specific grain boundaries in silicon [4]. This preparation method was able to site the crack interface routinely within the APT region of interest, and APT data of the metal-oxide interface showed is a depletion of Cr and Ni enrichment at the boundary, with a thin porous oxide rich in sodium. This was confirmed by focused ion beam cross-sectioning showing large voids at grain boundaries below the surface with a thin Na-rich oxide only at the edges of the voids, indicating a cracking mechanism dominated by the stress. The preparation process will be detailed, as well as correlative data between the two techniques to further the understanding of SCC behaviour in steels.

[1] M. Meisnar, M.P. Moody & S. Lozano-Perez, Atom probe tomography of stress corrosion crack tips in SUS316 stainless steel, Corrosion Science 98, pp 661-671, 2015
[2] S. Moore et al, A study of dynamic nanoscale corrosion initiation events using HS-AFM, Faraday Discussions, 210, 2018
[3] A Laferrere et al, In situ imaging of corrosion processes in nuclear fuel cladding, Corros. Eng., Sci. Technol., 52 , 596 —604, 2017
[4] C Lotharukpong et al, Specimen preparation methods for elemental characterisation of grain boundaries and isolated dislocations in multicrystalline silicon using atom probe tomography, Materials Characterisation, 131, pp 472-479, 2017