High‐Throughput Alloy Development Using Advanced Characterization Techniques During Directed Energy Deposition Additive Manufacturing

In laser‐based direct energy deposition (DED‐LB) additive manufacturing (AM), wire or powder materials are melted by a high‐power laser beam. Process‐specific characteristics enable robust in situ fabrication of compositionally graded materials, e.g., through an adaption of powder mass flow from independent hoppers. Based on the high flexibility of this approach, pathways toward multimaterial AM have been unlocked. Obviously, such characteristics enable high‐throughput alloy development. However, rapid alloy development demands substantial characterization efforts to assess phase and microstructural evolution. So far, property analysis is considered as the limiting factor for these high‐throughput approaches. Herein, the use of high‐brilliance X‐Ray analysis and subsequent micropillar compression testing are introduced to tackle these challenges. As a proof of concept, their application to a compositionally graded material made from AISI 316L stainless steel and a CoCrMo alloy is presented. The results obtained reveal that X‐Ray analysis can be exploited to evaluate process robustness, chemical characteristics, and phase composition within the gradient regions. Moreover, the use of micropillar compression testing provides spatially resolved insights into the mechanical properties of the gradient regions. The combination of both characterization techniques eventually opens pathways toward a robust and time‐efficient alloy development using powder‐fed DED‐LB (DED‐LB/P).


Introduction
Additive manufacturing (AM) enables the rapid fabrication of parts with unprecedented geometrical freedom and complexity. [1] In the field of metallic materials, powder bed fusion (PBF) and direct energy deposition (DED) are the most prominent AM processes. [2] Since the layer-wise deposition of heterogeneous materials across a powder bed is associated with significant waste production, PBF is merely used for the processing of monolithic materials. So far, pathways toward multimaterial AM are the subject of current research efforts. [3,4] In contrast, powder-fed DED processes (DED/P) facilitate a localized adaption of powder mass flow to the interaction zone with a heat source, e.g., plasma, laser, or electron beam. Thereby, the in situ fabrication of compositionally graded and, thus, functionally graded material combinations is rendered possible. [5] As a result, in situ multimaterial AM is unlocked. [6,7] This enables resource-efficient use of materials being aligned with their envisaged application and may also be used to In laser-based direct energy deposition (DED-LB) additive manufacturing (AM), wire or powder materials are melted by a high-power laser beam. Process-specific characteristics enable robust in situ fabrication of compositionally graded materials, e.g., through an adaption of powder mass flow from independent hoppers. Based on the high flexibility of this approach, pathways toward multimaterial AM have been unlocked. Obviously, such characteristics enable high-throughput alloy development. However, rapid alloy development demands substantial characterization efforts to assess phase and microstructural evolution. So far, property analysis is considered as the limiting factor for these high-throughput approaches. Herein, the use of high-brilliance X-Ray analysis and subsequent micropillar compression testing are introduced to tackle these challenges. As a proof of concept, their application to a compositionally graded material made from AISI 316L stainless steel and a CoCrMo alloy is presented. The results obtained reveal that X-Ray analysis can be exploited to evaluate process robustness, chemical characteristics, and phase composition within the gradient regions. Moreover, the use of micropillar compression testing provides spatially resolved insights into the mechanical properties of the gradient regions. The combination of both characterization techniques eventually opens pathways toward a robust and time-efficient alloy development using powder-fed DED-LB (DED-LB/P). overcome metallurgical limitations when combining different materials for optimizing internal stresses and crack propagation resistance. [8] Moreover, such process characteristics yield the fundamentals for high-throughput alloy development, e.g., in the field of refractory complex concentrated alloys [9] or highentropy alloys. [10,11] While alloy development may also be accomplished using conventional casting techniques [12] or thin-film deposition, [13] the aforementioned methods are characterized by a multitude of inherent disadvantages such as element segregation, low throughput or discrepancy between film, and bulk properties. [14] In DED, rapid solidification and high cooling rates can hamper element segregation and promote the formation of refined microstructures. [15] As a well-defined chemical composition of any alloy system is crucial for its performance, experimental rapid alloy development approaches are frequently supplemented by preliminary, computationally assisted calculations of phase diagrams using the CALPHAD approach [16] or simulations aided by machine learning. [17,18] In any case, fabrication of a large number of samples with alternating chemical compositions requires efficient high-throughput characterization techniques with regard to compositional accuracy, phase evolution, and corresponding mechanical properties. [19] Usually, any kind of characterization is carried out postprocess. As a consequence, a characteristic, recursive approach is followed for rapid alloy development, schematically depicted in Figure 1a.
Clearly, the need for elaborate postprocess characterization limits high-throughput alloy development exploiting AM so far. The present study aims to accomplish preliminary steps to overcome these limitations and expand the generic methodology by exploiting in situ characterization techniques for analysis of process robustness, i.e., emergence of pores, process-induced cracking, and lack of fusion. Throughout this investigation, the term in situ shall thus describe spatially resolved observations. Furthermore, approaches are introduced allowing for rapid assessment of site-specific characteristics with respect to chemical composition, phase evolution, and mechanical properties. From the literature, it is known that high-energy radiation, e.g., provided by synchrotron radiation facilities, allows to study relevant elementary processes in operando, i.e., during the Figure 1. a) Generic methodology for rapid alloy development using laser-based AM, recompiled with permission [14] Copyright 2017, Elsevier; b) Methodology supplemented by in operando characterization techniques; and c) expanded decision tree imposed by in operando characterization yielding a bidirectional workflow.
manufacturing process of the sample. [20][21][22] However, rapid analysis of mechanical properties even here remains an open challenge.
In the present study, only laboratory equipment is used. Approaches considered were selected to unlock substantial time savings during the development of material libraries. Through utilization of gradient samples and their spatially resolved direct analysis, a bidirectional workflow is established between sample fabrication and in operando analysis (cf. Figure 1b). Based on this approach, a multistage decision tree reflecting respective outcomes of the in operando analysis is created so that timeconsuming postprocess analysis of unsuitable material combinations can be skipped and a swift optimization of the sample fabrication process is enabled (cf. Figure 1c). In the present study, in situ analysis is enabled through a highintensity, liquid-metal laboratory X-Ray source. It has been shown that the photon flux of liquid-metal X-ray sources lies close to those of bending-magnet synchrotron sources. [23] Postprocess analysis employing the combination of scanning electron microscopy (SEM) and in situ micropillar compression tests is used to effectively assess the mechanical properties of the established alloy with high spatial resolution and with respect to crystallographic or chemical conspicuous areas of the graded samples. This approach is considered to further reduce the number of samples for the otherwise time-consuming postprocess analysis.
In summary, the aim of the present study is to judge on the feasibility of single elementary techniques to be embedded in the proposed general approach allowing for in operando studies in the future. For this purpose, a compositionally graded sample made of a CoCrMo alloy and AISI 316L stainless steel, which was fabricated by DED-LB/P, is characterized postprocess, however exploiting the analysis techniques introduced before. Results obtained are discussed focusing on their reliability and time consumption. While the employed alloy systems contain a multitude of alloy elements which exacerbate the chemical and phase analysis in contrast to the use of elemental powders, [9] the material combination is a promising candidate for applications in the medical sector owing to the excellent biocompatibility of both alloys. [24,25]

Experimental Section
Thin-walled structures with approximate dimensions of 40 Â 40 Â 2 mm 3 were fabricated using process parameters elaborated in a previous study [5] on a modular DED-LB/P machine equipped with a 2 kW fiber laser (IPG YLS-2000-S2, IPG Photonics GmbH) being characterized by an effective beam diameter of 1.2 mm on the workpiece surface. A quasi-bidirectional scanning strategy and a two-channel powder feeder (GTV PF2/ 2, GTV Verschleißschutz GmbH) were applied. As powder materials, a CoCrMo alloy and AISI 316L stainless steel with nominal particle size ranges from 50 to 150 μm were used. Their chemical compositions are listed in Table 1. As substrate material, an AISI 304 sheet with dimensions of 50 Â 50 Â 10 mm 3 was employed. Oxidation of the powder material and substrate surface was prevented through the use of argon shielding gas (purity > 99.996%).
Through in situ adjustment of powder mass flow, the chemical composition of the fabricated sample was varied from a region consisting entirely of AISI 316L to a transition zone with 50% AISI 316L and 50% CoCrMo before manufacturing a top region of pure CoCrMo (cf. Figure 2). For additional information on the DED-LB/P process applied, the reader is referred to a previous publication. [5] CALPHAD was used to assess the miscibility of iron and cobalt employing Thermocalc (Version 2021b) with the SSOL7 material database. To investigate the phase formation within the transition zone, a Scheil solidification simulation with the arithmetic mean of both chemical compositions (cf. Table 1) was performed.
Radiography measurements were enabled through a liquid metal X-ray source (MetalJet E1 þ 160 kV, Excillum AB) operating at an acceleration voltage of 70 kV and a power level of 50 W with a Ga 68 In 22 Sn 10 alloy (in wt%). A fast optical detector with a CMOS camera and an attached scintillation-foil (TPX4CAM, Amsterdam Scientific Instruments B.V.) was employed for radiography imaging. Energy dispersive fluorescence and diffraction analysis were conducted using a tungsten X-ray tube (2θ = 25°) with an acceleration voltage of 60 kV, a current of 40 mA in an X-ray diffraction (XRD) system (HUBER Diffraktionstechnik GmbH & Co. KG) equipped with an analytical X-ray point detector (AXAS-M, KETEK GmbH).
To obtain further information on microstructural evolution, chemical composition, phase evolution and, thus, to validate Table 1. Chemical composition of the powder materials, in wt% as provided by material supplier.  www.advancedsciencenews.com www.aem-journal.com results obtained by X-ray analysis, a scanning electron microscope (SEM; Zeiss Cross Beam 550, Carl Zeiss Microscopy Deutschland GmbH) equipped with an electron backscatter diffraction unit (EBSD; Oxford Symmetry S2, Oxford Instruments plc) and an energy-dispersive X-ray spectroscopy unit (EDS, Oxford Instruments plc) was employed. Postprocessing of the acquired EBSD data was conducted using ATEX software (version 3.26). [26] Additionally, micropillars were milled from the upper and lower interfaces of the transition region using a focused ion beam (FIB) and SEM (Zeiss Cross Beam 550, Ga þ , Carl Zeiss Microscopy Deutschland GmbH) system. The micropillars had a top diameter of 5.1-5.9 μm and a height of 9-16.9 μm, leading to aspect ratios of 1.6-3.3 with a taper angle of 4°-6.5°. Postmortem, in situ mechanical testing was performed using micropillar compression tests (FT-NMT 03, FemtoTools AG) in an SEM (Zeiss Cross Beam 1540Esb, Carl Zeiss Microscopy Deutschland GmbH). The tests were performed in displacement control at a constant strain rate of 0.0025 s À1 . To account for the elastic sink-in, the elastic modulus of the substrate was measured by nanoindentation (G200, KLA Corporation) with a Berkovich indenter at a constant strain rate of 0.05 s À1 assuming a Poisson's ratio of 0.3. The tip area and frame stiffness were calibrated prior to the experiments by indentations on fused silica. A continuous stiffness measurement oscillation with a frequency of 41 Hz and an amplitude of 2 nm was used. The analysis was performed according to Oliver and Pharr. [27,28] 3. Results and Discussion

CALPHAD Simulations
The computation of the binary iron-cobalt phase diagram in Figure 3a reveals a comparatively large miscibility window without formation of intermetallic phases. In the range from 20 to 100 wt% iron, the governing phase at low temperatures is body-centered-cubic (bcc) A2. At lower iron contents (down to approximately 6 wt%), a phase mixture of bcc and face-centered-cubic (fcc) A1 prevails at low temperatures. This observation already indicates a relatively good processability of the transition composition through DED-LB/P. However, it has to be mentioned here that other alloying elements are excluded from these computations and may readily affect the phase formation behavior. Furthermore, the binary phase diagram only accounts for the phase formation behavior at very slow heating and cooling rates. Clearly, such rates are not characteristic for the applied DED-LB/P process. An approximation of the solidification behavior closer to the cooling rates of DED-LB/P processes is obtained using a Scheil simulation (cf. Figure 3b). Results obtained predict the formation of secondary and ternary phases, in particular carbides of type M 23 C 6 and molybdenum-bearing precipitates, embedded in an fcc-matrix.

X-Ray Analysis
To assess the fabrication robustness and exclude fundamental processing defects of the investigated alloy composition such as pores, cracks, or lack of fusion, which hinder the further implementation of the chemical composition, radiography images of the different regions within the structures were obtained. As can be derived from their visualization in Figure 4, all investigated regions are free of the aforementioned defects on a macroscopic scale.
Similar reports on the defect-free processability of CoCrMobimetallic structures with Ni-based Inconel 718 have recently been reported on by Groden et al. [29] However, it has to be noted that detector artifacts are present within all obtained images as highlighted by the white arrows in Figure 4. The identical location of the black-dot artifacts in all obtained images evidences dead pixels on the detector sensor. Furthermore, the elliptically shaped artifacts, specifically visible in the regions of 316L and CoCrMo, were found to be caused by residual marks on the scintillation-foil in front of the radiography detector. This observation is supported by the higher normalized intensity stemming www.advancedsciencenews.com www.aem-journal.com from these artifacts in contrast to the surrounding matrix of the sample. Due to the density differences of the respective regions, the appearance of the artifacts and their intensity signal differ slightly. As for future applications, such artifacts may be effectively omitted through reference measurements and subsequent image subtraction. Additionally, the comparison of alternating acquisition times reveals identical corresponding normalized intensity distributions for the regions of 316L and CoCrMo, respectively. With respect to reliability, it is evident that reduced acquisition times do not attenuate the clarity of the presented results, i.e., the detection of possible defects. Clearly, the photon flux of the employed, laboratory-scale liquid-metal X-ray source is sufficient to facilitate reduced acquisition and, thus, overall analysis time consumption. As the employed power level of 50 W represents merely 5% of the maximum output power level of 1000 W, a further increase of photon flux is to be anticipated with a further power increase. Consequently, a corresponding reduction of acquisition time during radiography imaging is very well viable so that pathways to high-speed in operando X-ray imaging during DED-LB AM [20] are opened on a laboratory scale.
Furthermore, the means of X-ray analysis facilitate further insights in both chemical composition and phase evolution of the structures in dependence of the build height, as highlighted in Figure 5 Similarly, Ni Kα and Kβ signals are only detectable in the AISI 316L region and transition zone. Nevertheless, it has to be noted that the Ni signal intensity in the transition zone is far lower than in the AISI 316L region and dissipates completely in the CoCrMo region. This variation of fluorescence signal intensity is in good agreement with the initial chemical composition of the two alloys and the composition variation within the transition zone following the sample fabrication process (cf. Table 2), which could be determined using optical emission spectroscopy elsewhere. [5] As Ni is virtually not present in CoCrMo per its alloying concept, the full disappearance of Kα and Kβ signals can be readily explained. These observations can also be supported from a quantitative viewpoint as the Ni Kβ intensity of the first measurement equals 3247 counts while it is reduced to 1750 counts in the transition region, which resembles a reduction of 46.1% and, thus, almost ideally the chemical gradation of 50% 316L and 50% CoCrMo in the transition zone. Analog quantitative analysis on the Co Kβ signal from the CoCrMo zone with 16 199 counts and the transition zone with 7600 counts reveals a similar increase which was to be expected based on the varying chemical composition of the sample. The observed signal intensity increase of %213% is in good agreement with the underlying chemical composition of the respective regions. However, it has to be noted that the aforementioned X-ray-induced fluorescence signals have not been subjected to means of quantification to account for variations in atomic mass. To obtain quantified results, a weighted consideration of the X-ray excitation spectra [30] would be necessary before isolating the fluorescence peaks from the diffraction spectra. Nevertheless, the depicted results www.advancedsciencenews.com www.aem-journal.com prove that relative changes to the chemical composition can be identified in situ. It should also be noted that evaporating elements may contribute to an increase in signal intensity if evaporation is to occur in the focal point of the X-ray source. However, given the lower density in the gaseous state, the contribution of evaporated elements to the overall signal intensity may be described as negligible and, thus, limit the associated measurement error. As with further regard to the reliability of the signals, the relative distance of the discussed fluorescence spectra to the diffraction spectra provides safety from secondary effects such as peak superimposition. This observation is further underlined by comparison of the Mo Kβ signals, which emerge in proximity to the diffraction peaks in Figure 5b. The signal intensity variation from 540 counts in the 316L region to 1304 counts in the transition zone ideally resembles the chemical gradation of the sample. Therefore, it can be stated that the use of X-rayinduced fluorescence signals can enable a robust insight on the relative changes in chemistry of compositionally graded AM samples. However, the current analysis time of 7200 s per spectrum clearly hinders a swift application of fluorescence analysis for an in operando setup. Additionally, the XRD analysis depicted in Figure 5b illustrates the phase evolution being directly affected by the variation of the chemical composition. As can be derived from the peak intensities along BD, the governing phases are Fe-γ and Co-γ with varying orientations. Manifold materials without inherent phase transformations upon laser-based AM, e.g., AISI 316L, tend to form large columnar grains and a pronounced <001> texture along BD upon fabrication, [31] i.e., most grains oriented parallel to BD. As such coarse-grained microstructures are to be expected from the exemplary sample at hand, [5] it is viable that only a very limited number of grains account for the identified diffraction peaks. However, the visibility of Fe-γ and Co-γ is in good agreement to the phase predictions using CALPHAD. In addition to that, it is evident that the emergence and disappearance of certain peaks, such as Co γ 220, are strongly influenced by the varying chemical composition along the BD. This indicates a change of grain orientation through an adaption of chemical composition and has been documented previously. [5] While the diffraction spectra were acquired simultaneously to the fluorescence spectra, the aforementioned acquisition time of 7200 s per spectrum is evidently too long to be integrated in operando setup. Yet, a substitution of the point-detector used in the investigation at hand with a high-resolution, flat-panel detector as in ref. [21] can effectively reduce the analysis time by multiple orders of magnitude. Consequently, this may also limit the necessary acquisition time for fluorescence analysis.

Validation by Electron Microscopy Analysis
The acquired information about process robustness, chemical composition, and phase formation can be validated using the SEM analysis, as depicted in Figure 6. As can be deduced from the inverse pole figure (IPF) mapping in Figure 6a, the interface between AISI 316L and the transition zone is characterized by grains with varying size, aspect ratio, and orientation. No defects such as cracks, pores, or lack of fusion can be identified within the analyzed specimen area. Additionally, the phase mapping in Table 2. Chemical composition of the investigated regions, adapted from graphical illustration under the terms of the CC-by-4.0 license [5]   Step size along structure in BD is 2 mm. See Figure 2 for measurement trajectory along BD.
www.advancedsciencenews.com www.aem-journal.com Figure 6b confirms an fcc matrix along the entire interface region without any bcc or CoMo 2 phases, although pronounced changes in chemistry are detectable through EDS (cf. Figure 6c,d). While these changes, particularly for Fe and Co, are to be anticipated based on the chemical gradient, the consideration of a similar carbon content in the two materials (cf. Table 1) pinpoints limited diffusion between the different regions. This assumption is further supported by recent reports of Shin et al. [32] on the fabrication of compositionally graded specimens of carbon steel and austenitic steel, who only observed minimal carbon diffusion between the graded layers. Moreover, an area of distinctively refined grains is formed beneath the interface of the two zones. This phenomenon may be attributed to the addition of dissimilar powder material and its role as quasiheterogeneous nucleation site. Similar heterogeneous nucleation agents were shown to be effective means for grain refinement during laser-based AM, as reported by Durga et al. [33] The further randomized orientations are in good agreement with the emergence of the Co γ 220-peak in the transition zone and extinction of the Fe γ 200-phase peak observed during diffraction analysis as the growth of <001>-oriented grains, i.e., parallel to BD, is halted beneath the interface line (cf. Figure 6a). www.advancedsciencenews.com www.aem-journal.com As for the interface of the transition zone and CoCrMo, the IPF mapping in Figure 7a illustrates grains with varying size and orientation, albeit a large number of grains within the CoCrMo region exhibit a <101> orientation, i.e., a 45°inclination with respect to BD. Similar to the previous interface, the chemical gradient of the interface is clearly distinguishable, as Figure 7b demonstrates. However, the phase mapping in Figure 7c reveals an fcc matrix throughout. Interestingly, the high-resolution IPF and phase mapping of the inset in Figure 7a show the formation of molybdenum-rich regions with approximate sizes of less than 10 μm surrounded by the fcc matrix, probably representing the formation of CoMo 2 precipitates. These precipitates are also well distinguishable through their chemical composition (cf. Figure 7e). The formation of such precipitates has been predicted by the CALPHAD analysis in the above subsection, albeit in a different chemical composition and stoichiometry. Obviously, the precipitation formation within the transition region, which resembles a complex alloy system of a multitude of constituents, is of complex nature and, thus, may not be completely predictable using CALPHAD. Similar discrepancies between the simulated and experimentally observed results have been documented by Preisler et al. [9] during DED-LB/P of elemental powders with regard to concentration profiles around particles. The prevailing differences were attributed to insufficient experimental data within the respective databases.
In summary, the SEM analysis of both interfacial regions is in excellent agreement with the results of radiography, fluorescence, and diffraction analysis presented above. This further supports the robustness and reliability of the aforementioned techniques for an envisaged application as means for in operando analysis and validates the reliable acquisition of intended information.

High-Throughput Mechanical Testing
To further demonstrate the potential of high-throughput alloy development with the specimen at hand, in situ micropillar compression tests at both interfacial regions were conducted, whose results are depicted in Figure 8. In these regions, the Young's modulus was measured by nanoindentation. In the 316L transition zone, a modulus of 169 GPa was evidenced and in the CoCrMo transition zone, 190 GPa was measured.
The compressive stress-strain diagrams shown in Figure 8a pinpoint the reproducible mechanical behavior of the pillars at the interface between 316L and the transition zone. They exhibit a yield strength (at 0.2% elongation) of %200-250 MPa and www.advancedsciencenews.com www.aem-journal.com similar apparent hardening behavior. The slight experimental scatter may be attributed to different local grain orientations and chemistry and, thus, is a direct consequence of the remarkable spatial resolution of the micropillar compression technique. Moreover, the postmortem SEM image of the pillar indicates the absence of cracks and reveals shear bands in 45°orientation to the compression direction (see arrows in inset image). The yield strength is in the range of macroscopic tensile tests on identical specimens, [5] which had been machined parallel to BD. In the former study, the samples broke exclusively in the 316L region, which was rationalized by the lower UTS of the steel compared to the CoCrMo alloy. The present study revealed a similar strength in the transition region of the 316L. This clearly indicates that the plastic deformation will always localize in the 316L transition zone. While the insufficient spatial resolution of tensile tests is unsatisfactory for the full characterization of this particular material combination due to fracture initiation in the weaker base material, the spatial resolution of pillar compression tests allows for reliable and robust mechanical characterization regardless of the location within the structure. As recent reports on the compression behavior of CoCrMo-Inconel 718 bimetallic structures also illustrate the deformation mechanism being controlled by the properties of the bulk material, [29] the limitations of these techniques to characterize the interfacial properties can be overcome using pillar compression testing. The measurements at the interface between the transition zone and CoCrMo in Figure 8b reveal a much higher strength. It has to be noted, however, that the experimental scatter is also substantially higher than at the 316L/transition zone interface, pointing at stronger heterogeneities in grain orientation and/or chemistry. Here, the observed molybdenum-rich precipitates and their local distribution may play a central role and shall be investigated further in future studies. Similar to the previous interface region, the postmortem SEM images reveal no cracks but a number of shear bands in 45°orientation to the compression direction. The substantially higher flow stress in the CoCrMo transition zone confirms that the plastic deformation of the tensile tests from the previous study [5] mainly had to take place in the 316L zone. This highlights the importance of applying specially resolved mechanical testing for the successful realization of the methodological approach introduced in the present study. However, the utilization of SEM-based methods is typically not considered as a high-throughput characterization approach. In the present study, milling of a single micropillar using FIB took approximately 14 min. In conjunction with a compression test duration of about 5 min (see Supporting Information for in situ video recordings of the latter), the presented approach delivers insight into the mechanical properties of the respective alloy composition in less than 30 min, clearly outmatching the conventional approach of sample extraction, e.g., by electric discharge machining, and subsequent tensile, or compression testing. Moreover, as evidenced using the exemplary sample at hand, the spatial resolution of conventional mechanical testing is insufficient for an in-depth analysis of the respective gradient regions. www.advancedsciencenews.com www.aem-journal.com In summary, the mechanical characterization using pillar compression tests enables access to the mechanical behavior of each respective region with high spatial resolution. Thus, it is excellently suited as a characterization tool to support the postprocess characterization of high-throughput alloy development using AM. Indeed, only a reduced number of samples is required for nanomechanical testing as well as SEM characterization.

Envisaged Setup for in Operando Analysis on Laboratory Scale
As could be evidenced by the SEM-validation of in situ X-ray radiography, fluorescence and diffraction measurements, the aforementioned techniques prove to enable robust and reliable techniques for future in operando characterization during DED-LB/P on a laboratory scale. Evidently, the time consumption of the shown X-ray analysis though is too high to facilitate in operando analysis. Based on the discussion of experimental durations and corresponding setups, the laboratory scale setup to be used for in operando X-ray analysis during DED-LB/P is presented in Figure 9a. The setup to be realized will allow to evaluate three distinctive case studies in operando and correlate radiography imaging, fluorescence, and diffraction analysis in the laser focus point (cf. schematic in Figure 9b), at a given point with horizontal offset to the laser focus (cf. Figure 9c) or vertical offset to the laser focus (cf. Figure 9d). To tackle the challenges of prolonged acquisition times during radiography imaging, only a high-intensity, liquid-metal X-ray source with superior photon flux can be applied as is evidenced by the results presented and discussed. Furthermore, the point-detector and the standard setup used for diffraction analysis need to be substituted adequately. Again, the superior flux of the liquid-metal X-ray source is to be exploited in combination with a fast flat-panel detector, so that a substantial decrease in measuring time during X-ray diffraction analysis is to be anticipated. The use of a pyrometer-array focused on different points along the laser trajectory and build height will allow for a correlation of the data obtained during the three distinctive case studies with corresponding time/ distance-temperature data.

Conclusion
The goal of this investigation was to expand the generic methodology of high-throughput alloy development through in operando characterization techniques, applicable on a laboratory scale. The reliability and time consumption of the envisaged analyses were assessed by exemplarily applying advanced in situ characterization methods postprocess to a compositionally graded sample fabricated by DED-LB/P. Results reveal that despite comparatively short acquisition times during radiography imaging, enabled through a liquid-metal laboratory X-ray source, a reliable analysis of fabrication robustness is rendered possible. X-ray fluorescence and diffraction analysis proved to be reliable means to investigate relative changes in chemistry as well as the phase evolution, which was validated using SEM analysis. However, acquisition times further need to be improved significantly to enable an in operando application on the laboratory scale. Therefore, pathways toward a setup for in operando analysis on a laboratory scale are introduced. The approach of postprocess, in situ micropillar compression tests yields an overview of the respective mechanical properties with high spatial resolution. In consideration of the time consumption, this approach clearly outmatches conventional sample extraction and testing, underlining the applicability of SEM-based analysis for www.advancedsciencenews.com www.aem-journal.com high-throughput characterization. As a result, it is anticipated that the presented methodology will eventually aid robust and rapid alloy development processes in the field of, e.g., compositionally sensitive high-entropy alloys as well as shape-memory alloys. Future studies will compare the characteristics of the three distinctive case studies (cf. Figure 9b,d) and, thus, enable spatially and temporally resolved insights during high-throughput alloy development using DED-LB/P on the laboratory scale.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.