A new synthesis pathway for Fe–Cr oxide dispersion alloys enabled by ultrasonic atomization

1 A new synthesis pathway for Fe–Cr oxide dispersion alloys enabled by ultrasonic atomization Sandipan Sen1, Shubhashis Dixit1, Marcel Muench1, Liu Yang1, Christoph Somsen2, Sascha Seils1, Daniel Schliephake1, Alexander Kauffmann2, Martin Heilmaier1 1Institute for Applied Materials (IAM-WK), Karlsruhe Institute of Technology (KIT), EngelbertArnold-Str. 4, 76131 Karlsruhe, Germany 2 Institute for Materials (IM), Ruhr University Bochum (RUB), Universitätsstr. 150 44780 Bochum, Germany * corresponding author mail: alexander.kauff[email protected] (A. Kauffmann) phone: +49 234 32 18430 Abstract The study explores a novel route of manufacturing for oxide dispersion strengthened (ODS) alloys using ultrasonic atomization (UA) as the crucial step. A ferritic ODS alloy with the composition Fe–14Cr– 0.4Ti–0.5Y (wt.%) was synthesized by arc melting and suction casting using either elemental Y directly or a Fe₂Y master alloy as the Y source. This was followed by ultrasonic atomization under continuous Ar flow. The residual O in the gas flow during atomization is sufficient to create an O supersaturation in the ferritic matrix which subsequently provides the foundation for the formation of oxide dispersoids via internal oxidation. The manufactured powders contain fine Y42Ti10O45 dispersoids, formed via nucleation and growth from the supersaturated liquid. These factors lead to a uniform distribution of dispersoids in individual powder particles. An inhomogeneous distribution of dispersoids among the powder particles is observed, resulting from compositional variations in Y across individual powder particles during the atomization process. Keywords ODS steel, ultrasonic atomization, powder metallurgy, dispersoids, atom probe tomography 2 1. Introduction Fe-based oxide dispersion strengthened (ODS) alloys have been promising candidates for high temperature and nuclear applications. These alloys can be further subdivided into ferritic [1–4] and austenitic alloys [5,6]. Their applicability to fission and fusion reactor applications is attributed to their excellent swelling resistance, ability to reduce point defect concentration by trapping irradiation induced vacancies, and good irradiation creep resistance [1–4,7,8]. Additionally, they possess remarkable room temperature (RT) strength and excellent creep resistance above 600 °C [1,2,9–11]. These attractive properties can be attributed to homogenously dispersed oxide or more specifically Y-Ti-O type dispersoids of roughly 2–10 nm in size [10,12–17]. Fine dispersoids may effectively hinder dislocation motion by pinning dislocations [18,19], which enhances material’s strength, work hardening ability, and eventually even ductility [11,20,21]. However, coarsening of these dispersoids reduces their effectiveness in hindering dislocations due to increased interparticle spacing [22,23]. Moreover, they act as pinning centers for the grain boundaries, thereby suppressing grain growth and resulting in finer grains grain (compared to non-ODS counterparts) which contributes to Hall-Petch strengthening [12,18,21,24]. Historically, ODS alloys were suggested to be manufactured via powder metallurgy routes, which include mechanical alloying (MA) [2,5,6]. MA of ODS alloys involves high-energy ball milling of metal powders mixed with fine oxide particles (typically Y2O3), which leads to repeated fracturing, cold welding, and dissolution of oxides into the metal matrix [25]. These powders produced by MA are then typically consolidated via hot extrusion (HE) [26,27], hot isostatic pressing (HIP) [14,16], or field assisted sintering technique (FAST) [5,6,23,28]. Despite of its wide adaptation for powder production [25,29–31], the MA route has some critical drawbacks, namely higher O and N absorption during ball milling [32,33] and considerable weight loss during degassing process [33]. Higher O and N content can potentially have detrimental effect on the ductility. For example, in Fe-22Cr-5Al ODS alloy increasing O content from 0.04 to 0.16 wt.% form stripe and chain precipitates near grain boundaries causing partial intergranular fracture and reduced ductility [34]. Moreover, the MA process results in an inhomogeneous distribution of elements along grain boundaries which can lead to oxide particles preferentially concentrating at grain boundaries and this inhomogeneity is strongly dependent on the duration of MA [35,36]. Regions with high number density of oxides and oxide-free channels were reported in mechanically alloyed and hot extruded Fe-9Cr and Fe-14Cr (wt.%) [37,38]. The reason behind such non-uniform distribution was identified as uncontrolled O partial pressure during MA. Besides the non-uniform distribution, a large variation in size of the oxides was also reported, which negatively affected the mechanical properties [38]. For example, large 20–100 nm TiO2 particles were reported in a ferritic ODS alloy, manufactured by MA of Fe–14Cr–2W–0.3Ti and Y2O3 powders in Ar atmosphere followed by HIP [33]. To mitigate the drawbacks of classical MA routes, a novel synthesis route to produce ODS alloys is proposed in the present work. In this method, elemental Y is used to form dispersoids by oxidation instead of using Y2O3 powder in the production of the ODS alloy. Since most of the ferritic alloys described earlier are Fe-Cr based, with a Cr concentration between 13-14 wt.%, the Cr content in the alloys studied here is set to 14 wt.%. Ti is essential to adjust the size and potentially form Y-Ti-O type clusters. However, to prevent the formation of any coarse TiO2, the Ti content is limited to 0.4 wt.% Ti. ODS powder produced by MA mentioned earlier have a Y2O3 concentration of 0.25-0.5 wt.% [1–4]. We, therefore, set the Y content in the alloy to 0.5 wt.% and the final composition is Fe-14Cr-0.4Ti0.5Y (in wt.%)/Fe-15Cr-0.5-0.3Y (at.%). 3 This study aims to answer the following research questions: 1. Is it thermodynamically and kinetically possible to form Y-rich oxide dispersoids through internal oxidation of Y during or after ultrasonic atomization? At which stage of the processing route do dispersoids form and what is the mechanism behind the formation of these dispersoids? 2. Do the dispersoids form uniformly with respect to chemical composition, size and spatial distribution? 2. Results 2.1 Impurity concentration of raw materials Since the formation of oxide particles via internal oxidation is being investigated in this study, it is of paramount importance that the impurity content in the alloys is tracked throughout the entire processing route, beginning with the raw materials. The O and N concentration of the raw materials are given in Tab. 1. Tab. 1: O and N content in the raw materials via CGHE. Raw materials O / wt.% N / wt.% Fe 0.0136 ± 0.0011 0.0003 ± 0.0010 Cr 0.0118 ± 0.0020 0.0003 ± 0.0001 Ti 0.0181 ± 0.0016 0.0022 ± 0.0002 Y 0.8230 ± 0.0420 0.0070 ± 0.0010 It can be seen from Tab. 1 that O content in Fe, Cr and Ti used is low (<0.02 wt.%), while the elemental Y used in the study has a significantly higher concentration of O (0.823 wt.%). N contamination is generally found to be negligible. 2.2 As-cast condition To atomize the powders in ATO machine, rods were drop cast. To get a significant amount of powder at the end of the atomization process, four identically sized rods were drop-cast. A schematic of the drop cast rods is shown in Fig. 1a. The microstructure and composition were carefully investigated for both ends (top and bottom) of each of the four rods to ensure compositional homogeneity prior to UA. This is crucial because, as the alloy is remelted and allowed to fill the suction mold under the influence of gravity, there is a chance of elements with different densities (Fe: 7.87 g/cm3, Cr: 7.19 g/cm3, Y: 4.47 g/cm3 and Ti: 4.54 g/cm3 [39]) to separate during drop casting. This could cause compositional inhomogeneity in the rods which could be then carried over in atomized powders. 4 Fig. 1: a) Schematic of the drop cast rod along with SEM-BSE micrograph of the top and bottom section of the drop-cast rod showing a dendritic microstructure. b) SEM-EDS maps showing uniform distribution of Fe, Cr, Ti, O, and Y at a representative section of the microstructure showing a concentration of Y in the interdendritic regions and c) SEM-EBSD Kikuchi pattern of the Y-enriched phase in the interdendritic regions indicating a Fe10Cr2Y type of crystal structure (ThMn12 prototype, space group 139). The representative microstructure of the top and bottom section of one rod is also shown in Fig. 1a. SEM-BSE micrographs acquired from both sections of the rod reveal a microstructure characteristic of dendritic solidification. The microstructure is characterized by large grains that exhibit a darker contrast. A bright phase appears in interdendritic regions, both along the grain boundaries and within the grain interiors, suggesting the presence of elements with relatively high atomic number. As shown in the SEM-EDS elemental distribution map in Fig. 1b, the regions exhibiting bright contrast are enriched in Y, while Fe, Cr, and Ti are uniformly distributed across the microstructure. The dendrites, in contrast, are depleted in Y but maintain a relatively uniform distribution of Fe, Cr, and Ti. The homogeneous presence of O further eliminates the possibility that the phase in the interdendritic region is an oxide. The chemical compositions of the dendrites and the phase in the interdendritic regions, determined by standard-free SEM-EDS, are summarized in Tab. 2. Based on these observations, an intermetallic phase with an Fe:Cr:Y atomic ratio of approximately 80:13:7 in the interdendritic regions is confirmed. The crystal structures were investigated by X-ray diffraction (XRD) analysis performed on the rod samples. The XRD patterns confirmed the formation of the body-centered cubic (BCC) solid solution (W prototype, space group 229) corresponding to the matrix. The respective diffraction patterns are provided in Supplementary Material Fig. S2. The volume fraction of the Y-enriched intermetallic phase appears too small to be detected by XRD. To further identify the structure of the intermetallic phase, SEM-EBSD analyses were conducted in the Y-rich interdendritic regions. The corresponding Kikuchi diffraction pattern and its indexed solution are shown in Fig. 1c. The indexing result indicates a crystal structure like in Fe10Cr2Y (ThMn12 prototype, space group 139). This ThMn12 prototype phase has been reported in both Fe-Ti-Y and Fe-Cr-Y ternary phase diagrams as the τ-phase with a composition of YTixFe12-x and YCrxFe12-x, respectively. The most common phases reported are Fe10.8Ti1.2Y [40], Fe11TiY [41], Fe10Cr2Y [42] and Fe10.5Cr1.5Y [43]. This allows for a wide range of possible compositions that can crystallize in the ThMn12. Therefore, although, the composition of the intermetallic phase reported in Tab. 2 does not exactly match the ideal Fe10Cr2Y stoichiometry, it is nonetheless closely aligned. The minor deviations observed may be attributed to the large probe size of SEM-EDS analysis 5 compared to the microstructural details probed when applied to the relatively small dendritic arms present in the microstructure. The presence of a dendritic microstructure in the cast rods is not particularly concerning for the intended processing scheme, as the rods undergo remelting during the atomization process, wherein Y is fully miscible with both Fe and Cr in the liquid state. More importantly, the microstructure being consistent in the top and bottom sections of each rod indicates a high degree of chemical homogeneity. To further substantiate this observation, chemical analyses of samples collected from the top and bottom portions of all four rods are presented in Tab. 2. It can be seen from Tab. 2 that the overall alloy composition is close to the desired composition of Fe14Cr-0.4Ti-0.5Y in wt.%. The maximum deviation from nominal composition is observed for Fe (~2 wt.%), relatively small deviations are observed for Cr (0.2 wt.%) and Y (0.03 wt.%) while almost no deviation can be observed for Ti. The O concentration from the top and bottom of all four rods seems to be rather uniform raging between 0.005–0.020 wt.%. This slight variation can be expected since the four rods were cast in different casting experiments. Notably, the concentration N in the rods is low at this stage, being still < 0.002 wt.% in the worst case. Comparing these values to the alloys produced by MA is not trivial, since these alloys are produced using Y2O3 powders. Not many studies report accurate O/N contents measured via CHGE and if so, they usually do not distinguish the O content measured into O from the Y2O3 powders and excess O picked up during the process. However, this value was reported by García-Rodríguez et al. [44]. They reported the excess O content (obtained by removing the O content equivalent to the Y2O3 from the total measured O content. The Fe–14Cr–5Al–3W ODS alloy contained 0.297 wt.% excess O and 0.220 wt.% N, while the Fe–20Cr–5Al–3W alloy had lower levels of 0.104 wt.% O and 0.033 wt.% N. A comparison of these values clearly indicates that impurity uptake, at least during the arc melting stage, is significantly lower than that observed in the mechanical alloying process. 6 Tab. 2: Chemical composition of four the top and bottom of each of the 4 rods obtained by ICP-OES and CGHE after arc melting and drop casting. The nominal composition is also added for comparison. The composition of the matrix and intermetallic dendritic arms (observed in Fig. 1a) obtained by SEM-EDS is also included. 2.3 Ultrasonic atomized powders In this section, the morphology, microstructure and chemical composition of the powder samples produced by UA are shown. The results are categorized into two distinct powder conditions: the asatomized state (AA), representing powder collected directly from the atomization process, and the heattreated state (HT), referring to powder subjected to annealing at 1100 °C for 1 h. The heat treatment was performed to investigate one of two possible scenarios. In the first scenario, the heat treatment may facilitate their nucleation and growth if the desired oxide dispersoids are not formed during atomization. In the second scenario, the heat treatment enables examination of possible changes in their composition and size if such dispersoids are already present after atomization. Fig. 2a presents the top-view SEM-SE micrographs of the AA powder. The powder particles exhibit a predominantly spherical morphology with an average diameter of (59 ± 10) µm and a high sphericity of (0.90 ± 0.17). To investigate the internal microstructure, cross-sectional SEM-BSE images of AA powder particles were obtained, as shown in Fig. 2b-c. Two notable observations can be made, i.e. (i) The powder particles appear dense and nearly pore-free, with no discernible cavities or voids. This is a clear advantage over gas atomization processes where hollow powder particles are common [45,46]. (ii) More importantly, the powder particles as shown for example in Fig. 2b-c contain uniformly distributed dark spots throughout the SEM-BSE micrographs. To distinguish the nomenclature, these dark spots will be referred to as dispersoids that are in the rest of the manuscript. These dispersoids, ranging from approximately 20–200 nm, are located both along grain boundaries and within grain interiors as marked out by the orientation contrast of the matrix in the powder particle. The dark contrast of the dispersoids in the SEM-BSE images suggests enrichment in lighter elements, most likely Specimen Condition Average composition / wt.% Fe Cr Ti Y O N Desired 85.1 14 0.4 0.5 – – Rod 1 Top 84.1 ± 1.0 13.6 ± 0.2 0.4 ± 0.002 0.45 ± 0.01 0.004 ± 0.002 <0.0005 Bottom 84.3 ± 0.8 13.8 ± 0.3 0.40 ± 0.002 0.48 ± 0.01 0.005 ± 0.001 <0.0005 Rod 2 Top 83.0 ± 0.1 13.9 ± 0.1 0.41 ± 0.004 0.45 ± 0.02 0.022 ± 0.002 0.002 ± 0.0007 Bottom 82.7 ± 0.1 13.8 ± 0.1 0.40 ± 0.004 0.41 ± 0.02 0.007 ± 0.001 <0.001 Rod 3 Top 82.8 ± 0.2 13.9 ± 0.1 0.40 ± 0.002 0.44 ± 0.02 0.015 ± 0.001 0.002 ± 0.0003 Bottom 82.7 ± 0.4 13.8 ± 0.1 0.43 ± 0.005 0.51 ± 0.09 0.005 ± 0.003 <0.001 Rod 4 Top 82.4 ± 0.1 13.8 ± 0.1 0.40 ± 0.001 0.53 ± 0.01 0.013 ± 0.001 <0.001 Bottom 82.6 ± 0.3 13.9 ± 0.1 0.40 ± 0.001 0.49 ± 0.01 0.005 ± 0.001 <0.001 Avg. 83.1 ± 0.6 13.8 ± 0.1 0.39 ± 0.02 0.47 ± 0.02 0.009 ± 0.003 <0.001 Matrix (SEM-EDS) 85.2 ± 0.5 14.2 ± 0.1 0.36 ± 0.03 0 ± 0.10 – – Intermetallic (SEM-EDS) 80.9 ± 0.5 12.1 ± 0.1 0.42 ± 0.04 6.56 ± 0.10 – – 7 corresponding to oxides or nitrides. These type of dispersoids are marked with yellow arrow in Fig. 1c. In addition to the darker-contrast oxides or nitrides, small bright specks are also visible in the micrographs in Fig. 2b-c. Based on the microstructural observations of the rods in Fig. 1, these bright specks are inferred to have a higher concentration of Y compared to the surrounding matrix. Therefore, three different types of dispersoids are possible: (i) oxides, (ii) nitrides and (iii) Y rich intermetallic phase. This intermetallic phase is marked with blue arrows in Fig. 2c. Fig. 2d presents the top-view SEM-SE images of the HT powder particles. The particles appear partially agglomerated, which can be attributed to diffusion bonding occurring between adjacent particles during the heat treatment process. Fig.s 2e-f displays the microstructure of an individual HT powder particle. It is evident from these images that the type of dispersoids (dark and bright specks) observed in the AA condition are also present in the HT state, indicating a high degree of thermal stability of these dispersoids. The two types of dispersoids are marked in Fig. 2f with yellow (dark specks) and blue (bright specks) arrows similar Fig. 2c. Fig. 2: SEM micrographs of: (a–c) AA powder and (d–f) HT powder. (a,c) top view SE images, (b,e) BSE cross-section images, and (c,f) their magnified regions. The yellow and blue arrows highlight the dark and bright contrast dispersoids. The dispersoids are however, distributed non-uniformly across the powder particles. In both the AA and HT conditions, significant variations are observed in their number, size, and spatial distribution. To categorize the powder particles, they are classified into three groups based on the dispersoid density within a given micrograph area. For statistical consistency, cross-section SEM-BSE images of 100 AA powder particles were analyzed at identical magnification (20,000 x) with an imaged area of 130 µm2. The dispersoid size, number, and inter-dispersoid spacing were quantified using ImageJ. Based on the number of dispersoids per imaged area (130 µm2), the particles are grouped as follows: low density (< 50 dispersoids), medium density (50–250 dispersoids), and high density (> 250 dispersoids). Representative microstructures and corresponding quantitative data for each category are presented in 8 Tab. 3. The dispersoid size was determined by considering at least 200 dispersoids for each of the three categories. The inter-dispersoid distance referred the distance between the center of 2 dispersoids. Tab. 3: Quantitative classification of AA powder particles based on dispersoid density, showing the representative microstructures, average dispersoid size, and inter-dispersoid spacing for the low, medium, and high dispersoid density categories. As shown in Tab. 3, among the 100 analyzed powder particles, the majority exhibit medium dispersoid density (51 %), followed by low (38 %), and high dispersoid density particles forming the smallest fraction (10 %). The inter-dispersoid spacing is the largest for particles with low dispersoid density and the smallest for those with high density. This trend is expected, as the dispersoids are homogeneously distributed within each individual particle, and the analyzed area (130 µm2) remains constant across all observations. Consequently, the inverse relationship between dispersoid density and inter-dispersoid spacing is consistent. Notably, the average dispersoid size remains nearly identical across all three categories, implying a similar dispersoid formation mechanism in all powder particles. The chemical composition (determined by ICP-OES) of the atomized and heat-treated powders are shown in Tab. 4. Most notably, a loss of around 0.2 wt.% of Y loss is recorded in the atomized powders (0.30 ± 0.01) wt.% compared to drop cast rods (0.47 ± 0.02) wt.%. The O content increased to (0.079 ± 0.001) wt.% which indicates a pickup of O during the atomization process despite the < 10 ppm level O maintained in the Ar atmosphere of the ATO device during atomization. However, this content is still lower than O picked up during the MA process as described in the previous section. Powders from both AA and HT conditions still show a ferritic (BCC) matrix with no peaks corresponding to the dispersoid phase are observed. These XRD results are shown in the Supplementary Material Fig. S2. Tab. 4: Chemical composition of AA and HT powder conditions obtained by ICP-OES and CGHE. The nominal composition has also been added for comparison. Specimen Condition Average composition / wt.% Fe Cr Ti Y O N Desired 85.1 14 0.4 0.5 - - AA 83.4 ± 0.2 13.2 ± 0.1 0.43 ± 0.004 0.29 ± 0.005 0.079 ± 0.001 0.003 ± 0.0001 HT 84.7 ± 0.5 13.0 ± 0.1 0.41 ± 0.003 0.29 ±0.002 0.102 ± 0.012 0.006 ± 0.001 Dispersoid density (number of particles per 130 µm2) Low (< 50) Medium (50–250) High (> 250) Number of powder particles 38 % 52 % 10 % Inter-dispersoid distance (µm) 2.2 ± 1.3 1.3 ± 0.5 0.9 ± 0.3 Dispersoid size (nm) 100 ± 53 108 ± 65 101 ± 70 9 2.3.1 Chemical composition of dispersoids To determine the chemical composition of the dispersoids, APT samples were prepared from powder particles from both AA and HT conditions. To maintain consistency, all the lift-outs for APT were done from particles with a medium dispersoid density powder particle. Two tips from each of AA and HT condition are shown in Fig. 3. Fig. 3: Reconstructed map of APT dataset containing elemental distribution maps, isosurface constructions and onedimensional concentration profiles for dispersoids for a, b) AA powder and c, d) HT powder. Fig. 3a and b illustrate the APT results obtained from the AA powder condition, showing the spatial distribution of six elements: Fe, Cr, Ti, Y, O, and N. While Fe, Cr, and Ti exhibit relatively uniform distributions, noticeable inhomogeneities are evident for N, O, and Y. To further examine these regions of compositional variation, isosurfaces were constructed at 4 at.% for Ti and 4 at.% for N, as displayed in the enlarged view of Fig. 3a. It should be noted that the generation of a Y isosurface encompassed both types of dispersoids, indicating that Y is enriched in both phases. For Fig. 3b, an isosurface for 9 at.% Y was made, since no clear inhomogeneities were observed for Ti and N. Isosurfaces corresponding to a particular element delineate the regions of high concentration of this particular element within the tip, revealing distinct interfaces between the element rich regions and the surrounding matrix. These 16 2. Precipitation from the solid phase after solidification of the powder To determine which of these mechanisms is likely active, we first examine the spatial distribution of the dispersoids. Precipitation of the oxides from the solid phase (which could be expected to be incoherent with the body-centered cubic matrix), would be expected to proceed via heterogeneous nucleation at the grain boundaries of the solidified powder particle. However, since the micrographs shown in Fig. 2 clearly show a uniform distribution of dispersoids with no preference over grain boundaries, this could be the first indication that the formation of the dispersoids may indeed proceed via primary solidification from the liquid. The second hint towards the formation of dispersoids via primary solidification from the liquid is the variance in size of the dispersoids. The average size of the dispersoids is ~100 nm. For an oxide nucleus to grow to 100 nm, the diffusivity of Y in the matrix (either solid or liquid) must be fast enough to enable diffusion length of 100 nm. Whether dispersoids form from in the solid or liquid phase can be determined by calculating the diffusion lengths that Y must achieve under each condition to grow to 100 nm. It must be noted that the colling rate in the atomization process is extremely high. Notably, as soon as the liquid droplets are detached from the melt pool due to capillary waves and are suspended in the atomization chamber, they are immediately cooled by the flowing Ar stream. Data on exact values of cooling rates are not yet reported in literature, however cooling rate from gas atomization (106 K/sec [50]) might be a good estimate also for the ATO system. For the following calculation, a few reasonable assumptions are made: 1. The solidus temperature of the alloy is assumed to be similar to that of Fe–14Cr, which is approximately 1535 °C (1808 K) according to the Fe–Cr phase diagram [51]. 2. The diffusivity of Y in the molten alloy at 2410 °C is assumed to be comparable to that of Fe in molten Fe at the same temperature, with an estimated value of about 8.9 ∙ 10⁻⁸ m²/s [52]. This assumption is made because diffusivity data for Y in molten Fe–Cr alloys are not available in the literature. 3. The diffusivity of Y in the solid alloy is approximated by the diffusivity of Y in BCC Fe at 1500 °C, which is around 10⁻¹⁵ m²/s. This temperature is selected based on the assumption that solid-state precipitation, if it occurs, initiates immediately after solidification and continues until the temperature decreases by approximately 500 K [53]. Taking the above parameters, we can calculate the diffusion length scales for the two mechanisms mentioned above. For both cases, we would consider that the growth takes place within a temperature drop of 500 K. This assumption along with the assumption of the cooling rate, yields a growth time of 5 ∙ 10-4 s. The diffusion distance for precipitate growth can be calculated by the equation 𝑥 ≈ 2√𝐷𝑡, where 𝐷 is the diffusivity and 𝑡 refers to the time. In the solid phase the diffusion distance comes out to be approximately 1.4 nm. This reflects the extremely limited atomic mobility within solids over the available diffusion time. Given such a short diffusion distance, the growth of dispersoids to sizes on the order of 100 nm is effectively unfeasible under these conditions. In contrast, the calculated diffusion distance in the liquid phase extends into the micrometer range, approximately 13.3 µm, indicative of significantly enhanced atomic mobility and mass transport. This substantial difference in diffusion lengths clearly demonstrates that nucleation and growth of dispersoids to larger sizes, such as 200 nm, can be readily achieved in the liquid phase due to its higher diffusivity and more favorable kinetic conditions. 17 Based on the preceding arguments and considering that the melting point of the Y42Ti10O45 oxide is expected to be comparable to that of Y2O3, approximately 2400 °C [57], it is highly probable that the oxide dispersoids will solidify prior to the solidification of the surrounding matrix. While it is confirmed that the formation of oxide dispersoids through Y internal oxidation is thermodynamically feasible, the reasons for the uneven particle size and distribution must be explored. During the ultrasonic atomization process, the alloy on top of the sonotrode is maintained in a molten state for approximately 3–5 min while the atomization process is underway. Though the material is continuously heated with a moving electric arc, local temperature drop is not completely avoidable. This might have caused the lighter element, e.g. Y, to enrich towards the top of the melt pool under the influence of convection diffusion and gravity. This led to a variation in Y concentration between powders atomized from the top and bottom regions of a single melt pool. This can be verified by the compositional analysis of the left-over material in the atomization chamber (material that was not atomized and cooled down slowly), shown in Fig. S3 in the Supplementary Material. The microstructure of the left-over materials closely resembles those observed in the drop cast rods (Fig. 1). The Y concentration in the left-over material (2.3 ± 1.7 wt.%) is considerably higher than the average composition, which corroborates with the description on the Y segregation in the melt pool and further explains the drop in Y content after atomization (Supplementary Material Tab. S2). 4. Conclusions The Fe-14Cr-0.5Ti-0.5Y ferritic ODS alloy is manufactured via powder metallurgy route using a novel ultrasonic atomization (UA) technique with elemental Y rather than Y2O3. The formation mechanism of dispersoids and their distribution in the powders are investigated. Based on experimental observations, the following conclusions can be made. • The proposed UA technique successfully produces oxide dispersoids in the atomized powders and these dispersoids remain chemically stable when subjected to thermal exposure during heat treatment of powders or during sintering via FAST. In addition to the oxides, nitrides and intermetallic dispersoids are also formed. • Although formation of dispersoids is achieved during the UA process, their number, size, and spatial distribution vary significantly from one powder particle to another and strongly depends on the Y content present in each powder particle at the time of atomization. • The inhomogeneity in the dispersoid density across different powder particles is attributed to the non-uniform Y concentration in powders atomized from the top and bottom part of the melt pool in the ATO. • The formation of dispersoids occurs through homogeneous nucleation and growth directly from the liquid phase. This is supported by evidence based on the spatial distribution of the dispersoids as well as the distances Y needs to diffuse for the dispersoids to grow to 100nm or larger. The successful demonstration that nanoscale oxide dispersoids can be produced both thermodynamically and kinetically paves the way for new opportunities in the design and development of advanced ODS alloys. This breakthrough enables innovative alloy engineering strategies aimed at optimizing mechanical properties for demanding application. 18 5. Materials and Experimental methodology 5.1 Material synthesis During the initial arc melting stage, two distinct synthesis approaches were employed to cast the alloy: one utilizing elemental Y and the other employing a Fe2Y master alloy. This manuscript primarily focuses on the processing route involving elemental Y, while results for the Fe2Y master alloy route (referred to as ODS-Fe2Y) are provided in the Supplementary Material Fig. S1. The alloys were synthesized through repetitive arc melting on a water-cooled copper mould in an Ar (99.998 % purity) atmosphere using an AM/0.5 furnace (Edmund Bühler GmbH, Germany). The purities of the constituents (Fe, Cr, Ti, and Y/Fe2Y) were 99.95 %. The arc-melted buttons were then drop cast into cylindrical rods of 10 mm diameter and 120 mm height in a water-cooled suction mould, which were further used for ultrasonic atomization (UA). The UA process was caried out using the ATOLab+ (3DLab Ltd.) system. The chamber was evacuated till the O content in the chamber was about 10 ppm, then filled with Ar (99.998 % purity). A crucible with steel inset was used. The machine was operated at a vibration frequency of 35 kHz, an amplitude of 60 % and an arc current of 100 A (at arc light up) up to 165 A. The entire atomization process was carried out under continuously flowing Ar atmosphere (99.998% purity) with a flow rate of 25 l/min. The rods were systematically fed into the atomization chamber; a small part of the rod was then melted by an electric arc and deposited on the steel inset. The deposited material is maintained in a molten state using the movable arc. Further, the ultrasonic vibration is transferred from the cold end of the sonotrode towards the hot end, which forms capillary waves on the surface of the molten pool. Once the critical vibration amplitude is reached, the droplets were ejected to the flowing Ar stream, which cools them instantly and carried them to the cyclone chamber. Further details of the ATO machine can be found in Ref. [54]. The powders were collected in the sealed chamber and transferred to a glove box all under Ar atmosphere. To heat treat the atomized powder, a Mo foil envelope was created, the powder was then placed inside this Mo envelope. This powder-containing Mo envelope was then heat treated oat 1100 °C for 1 h in a were a HTRT 70-600/18 resistance tube furnace (Carbolite Gero GmbH & Co. KG, Germany), employing three evacuation–backfilling cycles followed by continuous Ar flow (99.998 % purity ) to minimize further oxidation. Then, atomized powders were consolidated via FAST at 1130 °C with force of 24 kN and dwell time of 5 mins. 5.2 Compositional analysis The composition of the drop cast rods, atomized and heat-treated powders as well as the FAST sample were obtained using inductively coupled plasma optical emission spectrometry (ICP-OES) for Fe, Cr, Ti, and Y. Similarly, carrier gas hot extraction (CGHE) was used to analyze O and N concentration. Furthermore, the O and N concentration of raw materials were measured using CGHE. At least five measurements were recorded for each condition to produce reasonable statistical data. 2.3. Microstructural characterization The microstructural characterization of the alloy was carried out at distinct stages of the process. For scanning electron microscopy (SEM) investigation of the drop cast rod (cross-section view), the rods were sectioned using a diamond wire saw, ground with SiC abrasive sheets up to P4000 followed by polishing using 3 and 1 µm diamond suspension (5 min) and Buehler ITW (Germany) colloidal silica (10 min) on semiautomatic LaboPol-60 machine (Struers, Germany). The final polishing was carried 19 out in a Vibromet machine (Buehler ITW, Germany) using a non-crystallizing oxide suspension (Struers, Germany) for 10 h. Similarly, the powders were first cold embedded mounted in a resin to obtain the cross-section specimen of the atomized and heat-treated powders (Epoclear, Schmitz-Metallographie GmbH, Germany). Once cured, the pellets were further hot embedded in EpoMet F (Buehler ITW, Germany). In the case of the consolidated sample, the specimens were cut using a diamond wire saw to reveal the cross-section. These pieces were then hot embedded like the powder samples. The final specimen preparation was carried out following the metallographic procedures analogous to those employed for the rods. Furthermore, the specimens for the top-view of the powders were prepared by simply adhering the powders on a carbon tape. Secondary electron (SEM-SE) and backscattered electron (SEM-BSE) micrographs were obtained at an acceleration voltage of 20 kV using a Zeiss LEO 1530 scanning electron microscope (SEM by Zeiss, Germany). Additionally, SEM-EDS (Energy Dispersive X-Ray Spectroscopy) of the rod, powders and consolidated specimens were obtained using a Zeiss Auriga 60 system operating at 20 kV with Octane Super-A detector (Ametek, USA). Specimens for X-ray diffraction (XRD) analysis were prepared by mounting the powders on a stub with resin, dried adequately, and polished to ensure a flat surface. The samples were then analyzed using a D2 Phaser device (Bruker Crop.) equipped with a Cu K-α source and a 1D LynxEye line detector. The XRD measurement was carried out over a span of 10 to 145° at a step size of 0.01° and the operating voltage and current of the diffractometer were kept at 30 kV and 10 mA. The accumulated acquisition time is 384 s per step. Transmission electron microscopy (TEM) sample preparation was done with a Helios G4 dual beam FIB-SEM system by (ThermoFisher, USA). TEM investigations were conducted on a Tecnai F20 G2 SuperTwin microscope (ThermoFisher, USA) operated with a nominal acceleration voltage of 200 kV. High angle annular dark field images (HAADF) were obtained with a Fischione detector of type M3000 (E.A. Fischione Instruments, Inc., USA). Energy dispersive X-ray (STEM-EDS) spectra were acquired with an EDAX detector of type Elite, T Super with a sensor area of 70 mm2 (Ametek, USA). Atom probe tomography (APT) specimens were prepared directly from the powder samples using the conventional lift-out method in a dual-beam Strata 400S (ThermoFisher, USA. The APT measurements were conducted with a LEAP 4000X HR instrument (Cameca SAS) equipped with a UV laser operating at a wavelength of 355 nm. The experimental parameters were set to a specimen temperature of 50 K, a laser pulse energy of 50 pJ, a pulse frequency of 200 kHz, and a target detection rate between 0.3% and 1%. Three-dimensional reconstruction of the data was carried out with IVAS software version 3.6.14 (Cameca SAS), using SEM micrographs for guidance. 20 References [1] M. Praud, F. Mompiou, J. Malaplate, D. Caillard, J. Garnier, A. Steckmeyer, B. Fournier, Study of the deformation mechanisms in a Fe–14% Cr ODS alloy, Journal of Nuclear Materials 428 (2012) 90–97. https://doi.org/10.1016/j.jnucmat.2011.10.046. [2] C. Zakine, C. Prioul, D. François, Creep behaviour of ODS steels, Materials Science and Engineering: A 219 (1996) 102–108. https://doi.org/10.1016/S0921-5093(96)10415-9. [3] D.K. Mukhopadhyay, F.H. Froes, D.S. Gelles, Development of oxide dispersion strengthened ferritic steels for fusion, Journal of Nuclear Materials 258–263 (1998) 1209–1215. https://doi.org/10.1016/S0022-3115(98)00188-3. [4] M.B. Toloczko, D.S. Gelles, F.A. Garner, R.J. Kurtz, K. Abe, Irradiation creep and swelling from 400 to 600 °C of the oxide dispersion strengthened ferritic alloy MA957, Journal of Nuclear Materials 329–333 (2004) 352–355. https://doi.org/10.1016/j.jnucmat.2004.04.296. [5] S. Seils, A. Kauffmann, W. Delis, T. Boll, M. Heilmaier, Microstructure and mechanical properties of high-Mn-ODS steels, Materials Science and Engineering: A 825 (2021) 141859. https://doi.org/10.1016/j.msea.2021.141859. [6] S. Seils, A. Kauffmann, F. Hinrichs, D. Schliephake, T. Boll, M. Heilmaier, Temperature dependent strengthening contributions in austenitic and ferritic ODS steels, Materials Science and Engineering: A 786 (2020) 139452. https://doi.org/10.1016/j.msea.2020.139452. [7] D.S. Gelles, Research and Development of Iron-based Alloys for Nuclear Technology, ISIJ International 30 (1990) 905–916. https://doi.org/10.2355/isijinternational.30.905. [8] L.L. Hsiung, M.J. Fluss, S.J. Tumey, B.W. Choi, Y. Serruys, F. Willaime, A. Kimura, Formation mechanism and the role of nanoparticles in Fe-Cr ODS steels developed for radiation tolerance, Phys. Rev. B 82 (2010) 184103. https://doi.org/10.1103/PhysRevB.82.184103. [9] S. Ukai, M. Harada, H. Okada, M. Inoue, S. Nomura, S. Shikakura, T. Nishida, M. Fujiwara, K. Asabe, Tube manufacturing and mechanical properties of oxide dispersion strengthened ferritic steel, Journal of Nuclear Materials 204 (1993) 74–80. https://doi.org/10.1016/00223115(93)90201-9. [10] J.H. Schneibel, M. Heilmaier, W. Blum, G. Hasemann, T. Shanmugasundaram, Temperature dependence of the strength of fineand ultrafine-grained materials, Acta Materialia 59 (2011) 1300–1308. https://doi.org/10.1016/j.actamat.2010.10.062. [11] T.M. Smith, C.A. Kantzos, N.A. Zarkevich, B.J. Harder, M. Heczko, P.R. Gradl, A.C. Thompson, M.J. Mills, T.P. Gabb, J.W. Lawson, A 3D printable alloy designed for extreme environments, Nature 617 (2023) 513–518. https://doi.org/10.1038/s41586-023-05893-0. [12] J.H. Schneibel, C.T. Liu, M.K. Miller, M.J. Mills, P. Sarosi, M. Heilmaier, D. Sturm, Ultrafinegrained nanocluster-strengthened alloys with unusually high creep strength, Scripta Materialia 61 (2009) 793–796. https://doi.org/10.1016/j.scriptamat.2009.06.034. [13] A. Hirata, T. Fujita, Y.R. Wen, J.H. Schneibel, C.T. Liu, M.W. Chen, Atomic structure of nanoclusters in oxide-dispersion-strengthened steels, Nature Materials 10 (2011) 922–926. https://doi.org/10.1038/nmat3150. [14] C.A. Williams, E.A. Marquis, A. Cerezo, G.D.W. Smith, Nanoscale characterisation of ODS– Eurofer 97 steel: An atom-probe tomography study, Journal of Nuclear Materials 400 (2010) 37– 45. https://doi.org/10.1016/j.jnucmat.2010.02.007. [15] S. Ukai, M. Fujiwara, Perspective of ODS alloys application in nuclear environments, Journal of Nuclear Materials 307–311 (2002) 749–757. https://doi.org/10.1016/S0022-3115(02)01043-7. [16] M.J. Alinger, G.R. Odette, D.T. Hoelzer, The development and stability of Y–Ti–O nanoclusters in mechanically alloyed Fe–Cr based ferritic alloys, Journal of Nuclear Materials 329–333 (2004) 382–386. https://doi.org/10.1016/j.jnucmat.2004.04.042. [17] M.K. Miller, E.A. Kenik, K.F. Russell, L. Heatherly, D.T. Hoelzer, P.J. Maziasz, Atom probe tomography of nanoscale particles in ODS ferritic alloys, Materials Science and Engineering: A 353 (2003) 140–145. https://doi.org/10.1016/S0921-5093(02)00680-9. [18] Y. Li, J. Shen, F. Li, H. Yang, S. Kano, Y. Matsukawa, Y. Satoh, H. Fu, H. Abe, T. Muroga, Effects of fabrication processing on the microstructure and mechanical properties of oxide dispersion strengthening steels, Materials Science and Engineering: A 654 (2016) 203–212. https://doi.org/10.1016/j.msea.2015.12.032. 21 [19] K.G. Raghavendra, A. Dasgupta, N.S. Karthiselva, K. Jayasankar, S.R. Bakshi, Microstructural Characterization of Novel ZrO2 Dispersion-Strengthened 9Cr Steel by Spark Plasma Sintering, Journal of Materials Engineering and Performance 32 (2023) 9814–9824. https://doi.org/10.1007/s11665-023-08244-2. [20] Z. Dong, Z. Ma, L. Yu, Y. Liu, Achieving high strength and ductility in ODS-W alloy by employing oxide@W core-shell nanopowder as precursor, Nature Communications 12 (2021) 5052. https://doi.org/10.1038/s41467-021-25283-2. [21] H. Xu, Z. Lu, D. Wang, C. Liu, Microstructure Refinement and Strengthening Mechanisms of a 9Cr Oxide Dispersion Strengthened Steel by Zirconium Addition, Nuclear Engineering and Technology 49 (2017) 178–188. https://doi.org/10.1016/j.net.2017.01.002. [22] M. Nagini, R. Vijay, K.V. Rajulapati, A.V. Reddy, G. Sundararajan, Microstructure–mechanical property correlation in oxide dispersion strengthened 18Cr ferritic steel, Materials Science and Engineering: A 708 (2017) 451–459. https://doi.org/10.1016/j.msea.2017.10.023. [23] Cs. Balázsi, F. Gillemot, M. Horváth, F. Wéber, K. Balázsi, F.C. Sahin, Y. Onüralp, Á. Horváth, Preparation and structural investigation of nanostructured oxide dispersed strengthened steels, Journal of Materials Science 46 (2011) 4598–4605. https://doi.org/10.1007/s10853-011-5359-1. [24] A. Meza, E. Macía, A. García-Junceda, L.A. Díaz, P. Chekhonin, E. Altstadt, M. Serrano, M.E. Rabanal, M. Campos, Development of New 14 Cr ODS Steels by Using New Oxides Formers and B as an Inhibitor of the Grain Growth, Metals 10 (2020) 1344. https://doi.org/10.3390/met10101344. [25] C. Suryanarayana, E. Ivanov, V.V. Boldyrev, The science and technology of mechanical alloying, Materials Science and Engineering: A 304–306 (2001) 151–158. https://doi.org/10.1016/S09215093(00)01465-9. [26] T. Gräning, M. Rieth, J. Hoffmann, S. Seils, P.D. Edmondson, A. Möslang, Microstructural investigation of an extruded austenitic oxide dispersion strengthened steel containing a carboncontaining process control agent, Journal of Nuclear Materials 516 (2019) 335–346. https://doi.org/10.1016/j.jnucmat.2019.01.048. [27] S. Ukai, Alloying design of oxide dispersion strengthened ferritic steel for long life FBRs core materials, Journal of Nuclear Materials 204 (1993) 65–73. https://doi.org/10.1016/00223115(93)90200-I. [28] C. Heintze, M. Hernández-Mayoral, A. Ulbricht, F. Bergner, A. Shariq, T. Weissgärber, H. Frielinghaus, Nanoscale characterization of ODS Fe–9%Cr model alloys compacted by spark plasma sintering, Journal of Nuclear Materials 428 (2012) 139–146. https://doi.org/10.1016/j.jnucmat.2011.08.053. [29] U. Martin, M. Heilmaier, Novel Dispersion Strengthened Metals by Mechanical Alloying, Adv Eng Mater 6 (2004) 515–520. https://doi.org/10.1002/adem.200400410. [30] B.S. Murty, S. Ranganathan, Novel materials synthesis by mechanical alloying/milling, International Materials Reviews 43 (1998) 101–141. [31] M. Heilmaier, H. Saage, J. Eckert, Formation of ODS L12–(Al,Cr)3Ti by mechanical alloying, Materials Science and Engineering: A 239–240 (1997) 652–657. https://doi.org/10.1016/S09215093(97)00644-8. [32] Y. Wu, Q. Huang, L. Zhang, Y. Jiang, G. Zhu, J. Shen, Microstructure and Mechanical Properties of Y4Zr3O12-Added Fe–13.5Cr–2W Oxide-Dispersion-Strengthened Steels, Containing High Contents of C and N, Prepared by Mechanical Alloying and Two-Step Spark Plasma Sintering, Materials 16 (2023) 2433. https://doi.org/10.3390/ma16062433. [33] Z. Oksiuta, N. Baluc, Optimization of the chemical composition and manufacturing route for ODS RAF steels for fusion reactor application, Nuclear Fusion 49 (2009) 055003. https://doi.org/10.1088/0029-5515/49/5/055003. [34] Y. Zhang, Effect of Oxygen Content on Microstructure and Tensile Properties of a 22Cr-5Al ODS Steel, Materials 14 (2021). [35] N. Oono, S. Ukai, Precipitation of Oxide Particles in Oxide Dispersion Strengthened (ODS) Ferritic Steels, Materials Transactions 59 (2018) 1651–1658. https://doi.org/10.2320/matertrans.M2018110. [36] K. Nowik, Formation and Microstructural Evolution of Ferritic ODS Steel Powders during Mechanical Alloying, Materials 16 (2023) 765. https://doi.org/10.3390/ma16020765. 22 [37] A. Chauhan, F. Bergner, A. Etienne, J. Aktaa, Y. De Carlan, C. Heintze, D. Litvinov, M. Hernandez-Mayoral, E. Oñorbe, B. Radiguet, A. Ulbricht, Microstructure characterization and strengthening mechanisms of oxide dispersion strengthened (ODS) Fe-9%Cr and Fe-14%Cr extruded bars, Journal of Nuclear Materials 495 (2017) 6–19. https://doi.org/10.1016/j.jnucmat.2017.07.060. [38] M. Serrano, M. Hernández-Mayoral, A. García-Junceda, Microstructural anisotropy effect on the mechanical properties of a 14Cr ODS steel, Journal of Nuclear Materials 428 (2012) 103–109. https://doi.org/10.1016/j.jnucmat.2011.08.016. [39] J.R. Davis, ed., Metals Handbook Desk Edition, ASM International, Materials Park, 1998. https://doi.org/10.31399/asm.hb.mhde2.9781627081993. [40] Z. Liu, Z. Jin, C. Xia, 873 K Isothermal section of phase diagram for YFeTi ternary system, Scripta Materialia 37 (1997) 1129–1134. https://doi.org/10.1016/S1359-6462(97)00234-0. [41] O. Moze, L. Pareti, M. Solzi, W.IF. David, Neutron diffraction and magnetic anisotropy study of Y-Fe-Ti intermetallic compounds, Solid State Communications 66 (1988) 465–469. https://doi.org/10.1016/0038-1098(88)90962-3. [42] D.B. De Mooij, K.H.J. Buschow, Some novel ternary ThMn12-type compounds, Journal of the Less Common Metals 136 (1988) 207–215. https://doi.org/10.1016/0022-5088(88)90424-9. [43] P. Qian, N.-X. Chen, J. Shen, Atomistic simulation for the phase stability, site preference and thermal expansion of YFe12−xTx (T=Ti, V, Cr, Mn, Zr, Nb, Mo, W), Solid State Communications 134 (2005) 771–776. https://doi.org/10.1016/j.ssc.2005.01.048. [44] N. García-Rodríguez, Capability of mechanical alloying and SPS technique to develop nanostructured high Cr, Al alloyed ODS steels, Materials Science and Technology 30 (2014) 1676–1684. https://doi.org/10.1179/1743284714Y.0000000595. [45] R. Gerling, R. Leitgeb, F.P. Schimansky, Porosity and argon concentration in gas atomized g-TiAl powder and hot isostatically pressed compacts, Materials Science and Engineering: A 252 (1998) 239–247. [46] C. Cui, F. Stern, N. Ellendt, V. Uhlenwinkel, M. Steinbacher, J. Tenkamp, F. Walther, R. FechteHeinen, Gas Atomization of Duplex Stainless Steel Powder for Laser Powder Bed Fusion, Materials 16 (2023) 435. https://doi.org/10.3390/ma16010435. [47] S. Shah, On the Use of a Cluster Identification Method and a Statistical Approach for Analyzing Atom Probe Tomography Data for GP Zones in Al–Zn–Mg(–Cu) Alloys, Microscopy and Microanalysis 30 (2024) 1–13. https://doi.org/10.1093/micmic/ozad133. [48] T. Ogura, Atom probe tomography of nanoscale microstructures within precipitate free zones in Al–Zn–Mg(–Ag) alloys, Acta Materialia 58 (2010) 5714–5723. https://doi.org/10.1016/j.actamat.2010.06.046. [49] J.R. Davis, Metals handbook desk edition, in: ASM international, 1998. [50] T.M. Abu-Lebdeh, G.P. León, S.A. Hamoush, R.D. Seals, V.E. Lamberti, Gas Atomization of Molten Metal: Part II. Applications, American Journal of Engineering and Applied Sciences 9 (2016) 334–349. https://doi.org/10.3844/ajeassp.2016.334.349. [51] A. Jacob, Revised thermodynamic description of the Fe-Cr system based on an improved sublattice model of the σ phase, Calphad 60 (2018) 16–28. https://doi.org/10.1016/j.calphad.2017.10.002. [52] A. Meyer, Iron self diffusion in liquid pure iron and iron-carbon alloys, Journal of Physics: Condensed Matter 31 (2019) 395401. https://doi.org/10.1088/1361-648X/ab2855. [53] D. Murali, Diffusion of Y and Ti/Zr in bcc iron: A first principles study, Journal of Nuclear Materials 219 (2011) 208–212. https://doi.org/10.1016/j.jnucmat.2011.05.018. [54] B. Bałasz, M. Bielecki, W. Gulbiński, Ł. Słoboda, Comparison of ultrasonic and other atomization methods in metal powder production, Journal of Achievements in Materials and Manufacturing Engineering 116 (2023) 11–24. https://doi.org/10.5604/01.3001.0016.3393. Acknowledgements We gratefully acknowledge financial support from the Deutsche Forschungsgemeinschaft (DFG), no. KA 4631/2-1 and HE 1872/41-1. LY is financially supported by China Scholarship Council (CSC) with No. 202207000023. This work was partly carried out with the support of the Karlsruhe Nano Micro Facility, proposal no. 2024-032-032040 (KNMFi, www.knmf.kit.edu), a Helmholtz Research 23 Infrastructure at Karlsruhe Institute of Technology (KIT, www.kit.edu). We acknowledge the chemical analysis by HCGE at the Institute for Applied Materials (IAM-AWP) by Dr. Bergfeldt, Karlsruhe Institute of Technology (KIT). Conflict of Interest The authors declare no conflict of interest. Data Availability Statement The data presented in this study are available on zenodo at https://doi.org/ 10.5281/zenodo.17508782 under CC BY-SA 4.0 license. Further information is available upon request with [email protected]. Author Contributions Sandipan Sen: Investigation, Formal analysis, Data Curation, Visualization, Project administration, Writing - Original Draft, Writing - Review & Editing Shubhashis Dixit: Data Curation, Formal analysis, Investigation, Writing - Original Draft, Writing - Review & Editing Marcel Muench: Investigation, Formal analysis, Writing - Review & Editing Liu Yang: Investigation, Formal analysis, Writing - Review & Editing Christoph Somsen: Investigation, Formal analysis, Writing - Review & Editing Sascha Seils: Investigation, Formal analysis, Writing - Review & Editing Daniel Schliephake: Investigation, Formal analysis, Writing - Review & Editing Alexander Kauffmann: Formal analysis, Data Curation, Visualization, Supervision, Writing - Original Draft, Writing - Review & Editing Martin Heilmaier: Supervision, Resources, Funding acquisition, Writing - Review & Editing