Effect of Process Variables on Microstructure, Flow, and Tensile Properties of Nickel‐Based Superalloys with Designed Porous Structures Manufactured by Laser‐Based Powder Bed Fusion

The nickel‐based superalloys Haynes 282 and Inconel 625 are processed using laser‐based powder bed fusion of metals (PBF‐LB/M) to manufacture designed materials (DMs) with configurable open‐porous morphology. Such structures can be utilized in large gas turbines, for example, as high‐frequency dampers, heat exchangers, and for transpiration cooling features. This research investigates the effects of the applied volume energy input and laser deflection angle (as calculated by position) on the macro‐ and microstructure and tensile and flow properties of DMs made from Haynes 282. Furthermore, a novel combination of DMs and cubic lattice structures made from Inconel 625 has been investigated regarding potential improvements. The substantial interface between the two features, DM and lattice structure, is visually analyzed using microcomputed tomography. The analysis of the hierarchical DMs reveals that the combination of lattice structures and DMs is manufacturable and has benefits in terms of mechanical strength while having less significant effects on flow properties.

also investigated. [19,26,33] It was demonstrated for DMs that depowdering of the smallest cavities is possible up to certain relative densities. [15,16] However, tolerances in morphological properties must be considered. [15,16] Several types of influences were investigated for Haynes 282 DMs: first, the effect of building position on morphological properties was evaluated. The relative density (RD) decreases as the laser deflection angle (LDA) increases. [15,17] Furthermore, the morphology of DMs was customized by varying the process parameters. [17] In addition to correlations between individual process parameters and morphology, the volume energy (VE) as a combination of these parameters is a factor that contributes to morphological changes. [17] Increasing the VE increases the RD while decreasing the mean pore diameter and surface ratio. [17] A number of studies also looked into the effect of process parameters on the mechanical and functional properties of DMs. [10,[18][19][20][21]24,[29][30][31]33] It was demonstrated that a higher energy input results in increased mechanical strength of the openporous material. [18,20,21,24,30] Furthermore, the energy introduced during the manufacturing process influences the permeability of DMs, so increasing energy results in a less permeable structure. [10,19,20,29,31,33] It is reasonable to assume that the position-and processparameter-dependent morphology of DMs influences both mechanical and functional properties. The question of whether positioning and process parameter effects on these characteristics can be statistically determined has yet to be answered. To the best of our knowledge, the effects of process parameters and positioning on the mechanical and functional properties of DMs in Haynes 282 have not been investigated up to now. In the current investigation, tensile and flow samples using various settings have been manufactured via PBF-LB/M. Mechanical and functional properties have been tested, and the resulting data were linked to the manufacturing settings to understand how they interact. It has been showed that the positioning of the sample in build chamber and the process parameters have an impact on both mechanical and functional properties. Furthermore, it has been demonstrated to what extent a heat treatment developed for Haynes 282 bulk material affects the microstructure of the DMs. Influences on grain size were clearly visible when comparing samples with and without heat treatment, but not when comparing samples of different positioning or process parameter settings. Finally, the investigation of hybrid materials, by combining lattice structures and DM, has been investigated for Inconel 625. It has been tested if this novel arrangement results in increased mechanical strength and substantial changes in the discharge coefficient (a measure of pressure loss). The combination of lattice structure and DM yielded promising tensile and flow results.

Materials
TruForm powder (Praxair Surface Technologies, Inc., Indianapolis, USA) was used in the manufacture of Haynes 282 DMs. Powder sieve measurements in accordance with ASTM B214 showed that all powder particles were smaller than 53 μm. Laser diffraction analysis was used to determine the fraction of particle sizes in the powder in accordance with ASTM B822. The D10 was 22.2 μm, D50 was 31.0 μm, and D90 was 45.6 μm for the resulting distribution.
EOS NickelAlloy powder (Electro Optical Systems Finland Oy, Turku, Finland) atomized from Inconel 625 was used. The powder particles were sieved according to ASTM B214 measurement to ensure that their size was less than 63 μm. The particle size distribution of the powder was measured by laser diffraction in accordance with ISO 13320, with D10 being 20.0 μm, D50 being 34.3 μm, and D90 being 54.2 μm.

PBF-LB/M Process
Haynes 282 samples were manufactured using four lasers on the PBF-LB/M machine EOS M 400-4 (EOS GmbH, Krailling/ Munich, Germany). Overlapping areas or the use of multiple lasers on a single sample were not considered. All components were exposed against the shielding gas flow. Layer-wise recoating was accomplished using the EOS soft recoating system for the M400 series with natural rubber lip (EOS GmbH, Krailling/ Munich, Germany).
In PBF-LB/M, the LDA proved to be a good measure of positional dependency. [15,17] The sample-specific LDA is calculated using Equation (1). The Z-distance in the PBF-LB/M machine EOS M 400-4 from the laser deflection mirrors vertically to the build platform is 490 mm, and the X-and Y-distances are measured for each sample center point from the undeflected laser.
The specimens used to test the effect of position on tensile or flow properties were manufactured in various positions across the build platform: for the tensile samples, an LDA range of 0.73°to 14.52°was determined by applying Equation (1) to the specimen's center point. The LDA values for the flow samples ranged from 0.35°to 13.81°. Please see our recent publications for more information on the utilized PBF-LB/M machine EOS M 400-4. [15][16][17] The hierarchical DM specimens from Inconel 625 were manufactured using a single-laser PBF-LB/M machine EOS M 290 (EOS GmbH, Krailling/Munich, Germany). For part exposure order, the setting against the laminar shielding gas flow was applied. Powder was distributed on the build platform using the EOS soft recoating system with carbon fiber brush (EOS GmbH, Krailling/Munich, Germany).
Based on the results of a previous survey's buildability study, the parameters for Haynes 282 were chosen. [15] The process variables for the laser power L P , scan speed L S , and hatch distance H D were changed to examine whether the mechanical and functional properties change as the VE is altered. [17] An overview of the parameters used is provided in Table 1 where the underlined values represent the experiment's center point settings. All process parameters for Haynes 282 and Inconel 625 DM manufacturing are standardized to a parameter set used in bulk material production. The hatch rotation was kept constant at 67°.
To ensure complete removal of powder possibly trapped in pores, both Haynes 282 and Inconel 625 specimens were depowdered in two steps. As in previous studies, an SFM AT800 depowdering machine (Solukon Maschinenbau GmbH, Augsburg, Germany) was used after thorough vacuum cleaning. [15][16][17]

Heat Treatment
Heat treatment (HT) for both base materials was carried out in a Schmetz Type E horizontal high-temperature vacuum furnace (IVA Schmetz GmbH, Menden, Germany) at a pressure of < 10 À4 mbar with argon cooling. Both HTs are standard procedures for the corresponding bulk materials. For Haynes 282, a three-step HT with one solution annealing and two aging steps was performed. [34,35] The Inconel 625 specimens were solution annealed in a single step. [36] 2.3. Sample Design Figure 1 depicts a DM flow sample in horizontal (0°) orientation with a quarter section, as well as one half of a DM and a hierarchical DM flow sample in vertical (90°) orientation. All sample types contain a DM (2) or hierarchical DM cylinder (3) with a thickness of 12 mm and a diameter of 80 mm. A 5 mm-thick bulk ring (1) was used for sealing.
Blank 0°specimens were built directly on the build platform with a height of 22 mm. Wire electrical discharge machining (WEDM) was used to cut a thin slice from the top of the specimen, resulting in the creation of a reference plane (5). The WEDM cutting plane was then adjusted to reduce the specimen thickness to 12 mm (6). The force-free WEDM ensured that the DM surface was not smeared, which avoided influencing the test results.
After being cutoff the build platform, vertical (90°) flow samples were tested as-printed. To achieve a defined sealing surface, the bulk ring was designed with a 0.5 mm offset on both specimen sides. Figure 2 shows horizontally (0°) and vertically (90°) oriented flat tensile specimens. Their shape is in accordance with DIN 50125 (shape E, cross section 5 mm by 10 mm). To ensure secure clamping during the tensile test, all specimens have dense clamping ends (1).
The 0°specimens were built horizontally on the build platform with additional stock material so that five specimens could be fabricated from each block. To provide a reference plane (5) for the thickness setting, a thin slice was cutoff the top of the specimens using WEDM. Iterative equidistant cuts (6) were then used to prepare specimens with a thickness of 5 mm.
Vertical (90°) flat tensile specimens were tested without any further machining, except for the removal of the supports (4), which were designed to be broken off by hand. As a result, all dimensions and side face finish were as-printed.
The design of 0°and 90°flow and tensile samples for DMs and hierarchical DMs is based on the expected surface conditions in the intended LGT application.
The hierarchical DM specimens have a cubic lattice structure imprinted into the DM with a strut thickness of 0.75 mm and a cell size of 3.5 mm. This specimen type was only made from  Inconel 625. According to a previously established procedure, the density of the DM section was 4.15 g cm À3 : [15] Geometrical analysis revealed that the density of hierarchical DMs increased to 4.56 g cm À3 . This translates to a 10% increase in RD for hierarchical DMs over DM samples without an imprinted lattice structure. Figure 3 shows the layout of the flow test rig where the flow samples were tested. The flow testing device was connected to a 7 bar compressed air source (1). The mass flow rate was set using a downstream mass flow controller and (2) F 206BI FAD 00 V (Bronkhorst Deutschland Nord GmbH, Kamen, Germany). A manually operated valve (3) was installed in case the rig needed to be turned off in an emergency. A mass flow meter (4) was passed before the air entered the actual flow test section. An orifice plate Deltatop V06-1436-02 (EndressþHauser GmbH þ Co. KG, Reinach, Switzerland) with an inner diameter of 13.6 mm was mounted in the mass flow meter. A differential pressure gauge Deltabar S (Endress þ Hauser GmbH þ Co. KG, Reinach, Switzerland) was used to measure the pressure difference before and after the orifice plate, and the absolute pressure was measured with a pressure transducer Cerabar S (Endress þ Hauser GmbH þ Co. KG, Reinach, Switzerland). The flow was introduced into the upstream pipe after passing through this combined measuring arrangement (5).

Flow Testing
The length of the test tube was approximately 40 times the inner tube diameter of 76.6 mm. To ensure laminar and swirlfree flow, these values were chosen in accordance with DIN EN ISO 51672. A mounting flange (7) was attached to the pipe's upstream end; this component both secured and centered the specimens (9). Two flat ethylene propylene diene monomer rubber rings (8) were placed between the specimens' sealing surfaces and the flange. A downstream pipe (6) 2.5 times the inner pipe diameter was installed on the other side of the mounting flange. Following that, the air was released into the environment. A differential pressure gauge (10) Rackmount Intelligent Pressure Scanner 9816 (Measurement Specialties, Inc., Hampton, USA) was used to measure the pressure difference caused by the specimen at a distance of 76.6 mm before and after the sample. All samples were tested at a differential pressure of p 2 p À1 1 ¼ 0.97. Furthermore, the temperature of the air was measured at position (10) using a PT100 resistance thermometer in accordance with DIN EN 60751.
Using Equation (2), the collected data were employed to calculate the discharge coefficient C D . [19,37] The discharge coefficient for the circular tube is defined by the ratio of the measured mass flow m : to the ideal compressible the upstream and downstream pressures p 1 and p 2 , and the pipe diameter D were employed in the calculation. Furthermore, values for the specific gas constant R ¼ 287 m 2 s À2 K À1 and the isentropic coefficient κ ¼ 1.4 were used (both selected for dry air at 20°C and 1 atm). [38] By testing 11 flow samples in 0°and 24 flow samples in 90°p rint orientation, the effect of varying VE on the discharge coefficient was investigated (all from Haynes 282). In total, 16 samples were printed with the same VE but in different positions among the 90°specimens. On those samples, the effect of LDA was investigated. Flow testing was also performed on two Inconel 625 90°DM samples. Their results were compared to the flow test results of two Inconel 625 90°hierarchical DM specimens.

Tensile Testing
Tensile tests were carried out using a refurbished RetroLine testControl II AllroundLine 1475 materials tester (ZwickRoell GmbH & Co. KG, Ulm, Germany). An Inspekt Retrofit 1475 100 kN (Hegewald & Peschke GmbH, Nossen, Germany) was used to replace the control section. The built-in load cell had an accuracy of 0.5% over a measurement range of up to 100 kN: To obtain more detailed data in the linear-elastic tensile curve section, an extensometer type WAD 07.03.01 (Dr.-Ing. Georg Wazau Mess-und Prüfsysteme GmbH, Berlin, Germany) with a maximum measuring range of 1 mm and a measurement accuracy of 2 μm was used. Strains greater than 0.6% were calculated from the distance traveled by the crosshead with a measurement accuracy of less than 10 μm: The test velocity was kept constant at 0.5 mm min À1 : The software LabMaster 2.7 was used to process the raw data (Hegewald & Peschke GmbH, Nossen, Germany). Tensile data were used to determine ultimate tensile strength, Young's modulus, and elongation at fracture.
Testing 40 tensile samples in a 0°print orientation and 52 tensile samples in a 90°print orientation allowed us to examine how changing VE affected mechanical properties (from Haynes 282). Totally, 28 samples were printed with the same VE but at various places among the 90°specimens. On those samples, LDA's impact was examined. Two Inconel 625 90°DM samples were also tensile tested. Their results were contrasted with those of two Inconel 625 90°hierarchical DM specimens that also underwent tensile testing.

Normalization of Flow and Tensile Properties
The functional and mechanical data were normalized to improve comparability. For the Haynes 282 samples, a ratio was formed by dividing each value by the overall mean of the respective property (calculated from 0°and 90°data). To normalize the data of the Inconel 625 specimens, the mean of each tested property was calculated using all data from DM and hierarchical DM specimens. As described in Chapter 2.3, the density of hierarchical DM is 10% higher than that of DM alone. To compensate for the increased density, the normalized properties of the hierarchical DMs were corrected by À5% and those of the DMs by þ5%.

EBSD Analysis
Electron backscatter diffraction (EBSD) measurements were performed on DM tensile samples from Haynes 282 with and without HT. Different VEs and build positions were investigated.
All samples were sectioned and hot mounted in Duroplast conductive resin (ATM Qness GmbH, Mammelzen, Germany). Metallographic preparation was performed by grinding with silicon carbide papers up to P1200, followed by polishing with diamond suspension sizes of 6, 3, and 1 μm.
A Zeiss Sigma scanning electron microscope (SEM) (Carl Zeiss Microscopy Deutschland GmbH, Oberkochen, Germany), equipped with a Quantax e-flash (Bruker Corporation, Billerica, USA) EBSD-detector, was used at a magnification of 50 times and an accelerating voltage of 20 kV. The data were processed with the software Esprit 2.2 (Bruker Corporation, Billerica, USA). Figure 4 shows the locations of the sections for the EBSD study. In the X-Y plane, the horizontal sections H1 to H3 of 90°samples were studied for the microstructural HT effect. The same is true for the vertical section in the Y-Z plane (V1 to V3). Sections H1 and V1 were investigated for VE and LDA effects on the DM microstructure for both 0°and 90°tensile samples.
After depowdering, one Haynes 282 tensile sample with process parameter center point settings was removed from the build www.advancedsciencenews.com www.aem-journal.com platform. It had been cut in half. One side received an HT, while the other did not. The resulting two specimens were used to examine DM's microstructure with and without HT. All sections V1 to V3 and H1 to H3 were investigated here. Other microstructure studies used heat-treated samples in a variety of settings. We examined two 0°tensile samples at the most extreme VE settings. Two specimens at the most extreme LDA and two specimens at the most extreme VE were examined for the 90°samples. Only sections H1 and V1 were analyzed in these studies.

Fractography
With a Tescan Vega 3 SEM (TAZ GmbH, Aichach, Germany) in the secondary electron mode and an accelerating voltage of 20 kV, fracture surfaces were investigated on one 0°and one 90°tensile specimens with center point settings for Haynes 282. The fracture surfaces of hierarchical DMs were studied on a single Inconel 625 tensile specimen. Here, a lattice strut and a DM strut were in focus.

μCT
Microcomputed tomography (μCT) scanning was performed using a Yxlon FF85CT inspection system (Comet Technologies USA, Inc., Hudson, USA) to visualize the morphology of hierarchical DMs. In each of the 2160 projections, microfocus was used with a X-ray tube current of 380 μA and an X-ray tube voltage of 220 kV, as in two previous studies. [16,17] A prefilter made of 1 mm thick copper was also employed. [16,17] Data segmentation was carried out using VGStudio Max 3.3 (Volume Graphics GmbH, Heidelberg, Germany). The surface was defined using the ISO50 gray value threshold method. [39] The sphere method of VGStudio Max 3.3 was used to visualize the thickness of strut diameters, as described in two previous publications. [16,17,40] μCT was used to scan one hierarchical DM tensile sample from Inconel 625.

Haynes 282 DMs
3.1.1. Influence of the Heat Treatment Figure 5 depicts the grains of various material sections in the X-Y (H1 to H3) and Y-Z planes (V1 to V3), as well as the average grain size (GS). Without HT, DM (H1 and V1), bulk-DM-transition (H2 and V2), and bulk (H3 and V3) have comparable GS. This is also supported by the D10, D50, and D90 values shown in Figure S1-S3, Supporting Information. The grains of the nonheat-treated material sections have a rather equiaxed shape in X-Y plane (H1 to H3). Grains in the Y-Z plane (V1 to V3) without HT have a columnar morphology.
Following HT, the average GS increases for all analyzed material sections but in a nonuniform manner as depicted in Figure 5 and S1-S3, Supporting Information. The bulk section (H3 and V3) shows the biggest increase in GS and the DM sections (H1 and V1) the smallest one. Reviewing the average sizes of all three material sections, the grains in Y-Z (V1 to V3) are larger than in X-Y plane (H1 to H3). Despite the size, Figure 5 shows that the grain shape remained equiaxed in the X-Y plane (H1 to H3) and changed to equiaxed in Y-Z plane (V1 to V3). It is also obvious www.advancedsciencenews.com www.aem-journal.com that the GS in the bulk (H3 and V3) material after HT is larger than the actual strut size in the analyzed DM area (H1 and V1). It is clear from both analyzed planes that the grains are not distributed homogeneously in terms of size within the DM section. The EBSD results further show a morphological difference in the DM between the horizontal section (H1) and the vertical section (V1). The DM struts in the X-Y plane (H1) are more elongated. In contrast, the morphology in the Y-Z plane (V1) appears more random, with roundish strut sections.

Discharge Coefficient
The impact of VE (at 0°and 90°) and LDA (at 90°) on the discharge coefficient of DMs made from Haynes 282 is shown www.advancedsciencenews.com www.aem-journal.com in Figure 6. The average discharge coefficient is 1.56% in the 0°o rientation and 1.21% in the 90°orientation. The C D data are normalized with a 1.32% global mean.
The repeated samples (marked red in Figure 6b) that were investigated for the LDA effect indicate scatter. The LDA analysis revealed a significant correlation between LDA and C D with p < 0.05. With r ¼ 0.674, it was inferred that increasing LDA affects the C D positively. Figure 6 shows a significant and strong correlation of VE and discharge coefficient C D with p < 0.05 and r < À0.7 for both sample orientations. As a result, increasing VE is accompanied by decreasing C D . The average of the 90°oriented samples is marginally lower than that of the 0°oriented specimens. The scatter of the discharge coefficient in the samples with the same VE (highlighted in red in Figure 6b) ranges between þ20% and À28% from their mean. These highlighted samples were also used in the LDA effect analysis, which revealed a significant correlation between LDA and C D with p < 0.05. It was concluded with r ¼ 0.674 that increasing LDA has a positive effect on C D . Figure 7 depicts the effects of VE (in 0°and 90°) and LDA (in 90°) on the ultimate tensile strength σ u , Young's modulus E, and elongation at fracture ε A of DMs made from Haynes 282. The σ u data were normalized with an overall mean value of 93.8 MPa, while the E data were normalized with 21.5 GPa. In addition, for normalization, we used the ε A global mean value of 1.5%.

Tensile Properties
All three investigated tensile properties, σ u , E, and ε A , show strong positive correlations with VE in 0°orientation, as indicated by p < 0.05 and r > 0.7. As a result, as VE increases, so do σ u , E, and ε A . The 90°tensile specimens exhibited similar behavior. In this orientation, only ε A showed no significant correlation with VE. When comparing 0°and 90°oriented specimens, the 90°samples have slightly lower values for σ u , E, and ε A .  www.advancedsciencenews.com www.aem-journal.com LDA and the tensile properties σ u and E have statistically significant correlations with a probability of p < 0.05. The strength values r ¼ À0.484 (for σ u ) and r ¼ À0.524 (for E) indicate that both properties decrease slightly with increasing LDA. Changing LDAs had no effect on ε A . This property exhibited high scatter. Figure 8 shows the EBSD results for four samples manufactured at different VE levels in both the horizontal and vertical directions. Per each orientation, 0°and 90°, one specimen with low VE and one specimen with high VE (consult Figure 7) were investigated in vertical (V1) and horizontal (H1) sections. All samples were heat treated.

Macro-Versus Microstructure for Different Settings
The dominant difference in the sectional RD is first apparent when comparing the low VE samples to the high VE samples. Another noticeable contrasting aspect of the macrostructure is the reduced connectivity of the struts in the low VE samples compared to the high VE samples. Besides, more highly elongated struts are visible in the X-Y plane (H1) for high VE samples than roundish ones in the Y-Z plane (V1). For the low VE samples, the morphological differences in the X-Y section (H1) and the Y-Z section (V1) are less prominent.
The average GS for samples built with higher VEs is larger compared to low VE samples, with the D50 and D90 values (shown in Figure S4 and S5, Supporting Information) being primarily responsible for the increases. Furthermore, the average GS values in the X-Y sections (H1) are slightly lower than in the Y-Z sections (V1) for both VE levels. The grains are not distributed homogeneously in size, as evidenced by all analyzed samples and sections. Figure 9 depicts the EBSD results of two tensile samples under extreme LDA conditions high and low VE (compare Figure 7).
The sectional RD decreases in both orientations for the higher LDA sample when compared to the lower LDA sample. Furthermore, the morphology in the X-Y plane (H1) is dominated by elongated struts. The strut morphology appears more random and roundish in the Y-Z plane (V1). There are no significant differences in strut connectivity between high and low LDA settings. In the Y-Z section (V1), the average GS decreases at higher LDA when compared to lower LDA. When comparing low and high LDA samples, the value increases for the average GS in the X-Y plane (H1). The distributions shown in Figure S6, Supporting Information, support these inconclusive GS results. However, as evidenced by all analyzed samples and sections, the grains are not distributed homogeneously in size. www.advancedsciencenews.com www.aem-journal.com Figure 10 depicts the fracture surfaces of a DM specimen built at 90°(A1 to A3) and one built at 0°(B1 to B3). The fracture can be seen in images A1 and B1 at many of the DM struts. There is no evidence of strut necking. According to SEM images A2 and B2, some powder particles were not completely melted during the PBF-LB/M process and were thus sintered to the inner DM surfaces. The highest resolution images (A3 and B3) show the fracture surfaces with numerous cup and cone-shaped dimples. As a result, the material can be concluded to exhibit ductile fracture behavior in both orientations. Figure 11 depicts the strut diameters obtained by a μCT analysis of a hierarchical DM tensile specimen made of Inconel 625 using the sphere calculation method. Red in the color coding indicates material sections with larger dimensions. Medium strut sizes are indicated in green, while smaller sizes approaching 0 mm are indicated in blue. To inspect the internal structure, digital sectioning was used. Following the strut diameter analysis method, the regular cubic lattice structure with its thicker struts is indicated in red. Because the DM struts are smaller, they are mostly  www.advancedsciencenews.com www.aem-journal.com green with a few blue portions. When the connection between the lattice struts and the DM is examined, both sections are well connected in all directions.

Flow and Tensile Properties
The top row of Figure 12 shows the normalized mechanical and functional properties of hierarchical DM and DM specimens (1a to 4a). The average values of hierarchical DM and DM were calculated for normalization (59.9 MPa for σ u , 22.4 GPa for E, 3.5% for ε A , and 1.4% for C D ). Furthermore, Figure 12 shows the results where the higher density of the hierarchical DM was equalized in the bottom row (1b to 4b).
Regardless of whether the density effect was equalized or not, the hierarchical DM yielded significantly higher values for all three mechanical properties (σ u , E, and ε A ). Normalizing the results and compensating for the RD increase for hierarchical DMs revealed that σ u and ε A can be improved by 42%. E achieves Figure 11. Reconstructed volume data of a μCT scan of a hierarchical DM tensile sample in two different sections with indicated strut diameters. www.advancedsciencenews.com www.aem-journal.com an even higher value of 62%. Thus, by increasing the density of hierarchical DMs by 10%, mechanical behavior can be improved by at least 42%. However, this improvement in mechanical performance must be viewed in the context of possible changes in functional properties. When compared to DMs, hierarchical DMs have a 16% lower discharge coefficient C D . When the results are density normalized, this effect increases to 24%. Nonetheless, by introducing a cubic lattice structure, the flowable area decreases in a slightly smaller fashion compared to increases in σ u , E, and ε A .

Fracture Surface
The fracture surface of a 90°hierarchical DM sample is shown in Figure 13. A lattice strut fracture is shown in A1 to A3, and a DM fracture is shown in B1 to B3. A1 and B1 show that the lattice strut fracture has larger dimensions than the DM strut fracture. These two images lack necking, both for the lattice and the DM strut. However, the images A1, A2, B1, and B2 show attached powder particles on the inner surface of the hierarchical DM. A3 and B3 SEM images reveal small dimples on the fracture surface. This indicates that the Inconel 625 hierarchical DM exhibited ductile fracture.

Haynes 282 DMs
The microstructural, mechanical, and functional properties of DMs made from Haynes 282 with various settings were investigated. The impact of the applied HT on the microstructure will be reviewed first. Afterward, the effects of the applied process parameters and build position on functional and mechanical properties, as well as possible interactions of the obtained characteristics, will be discussed. Our nonheat-treated specimen has the typical size and columnar shape of nickel-based superalloys manufactured by PBF-LB/M in as-build condition. [35,41,42] Grain growth was greater in our bulk specimens than in previous studies that applied a similar HT. [35,43] This variation could be due to different measurement techniques or slightly higher temperatures in the first two cycles of the employed HT. [35,43] The microstructure in bulk and openporous sections after HT differs significantly. On the one hand, smaller grains are noticed, yet they are not evenly dispersed. The second aspect is of particular interest, as it appears that the microstructure in the DM struts did not fully recrystallize. We have no evident explanation for this behavior at this time, but one reason could be the complex morphology of DM, where a large number of grains form simultaneously during HT and block each other's growth. Even the larger grains in the DM are smaller than those in the bulk section. The morphology of DM provides one possible explanation, as grain growth may be limited by the DM strut diameters, as seen in another study analyzing nickel foams. [44] The fact that GS increased for denser DMs highlights this, as grains are given more volume to grow in. Nonetheless, the applied HT aligned the differences in grain shape between the X-Y and Y-Z planes in bulk and DM in a similar manner. [35] The fact that the HT has no effect on the macroscopic shape of DMs is a morphology-related finding.
Our finding that VE has an effect on C D confirms the findings of other studies that used different materials and machines. [19,20,29,33] These results are most likely related to altered morphological characteristics associated with varying VE. [17] This possible explanation is supported by the sectional RDs resulting from various VE settings. For example, increasing the RD of a DM by increasing the applied VE results in a lower C D because more material is blocking the flow. Here, the results of increasing C D and www.advancedsciencenews.com www.aem-journal.com decreasing RDs with higher LDAs provide another hint to this correlation. [15] The correlation between RD and C D indicates that previously developed compensation approaches for limiting the effect of LDA on the RD could also aid in reducing C D scattering when components are distributed across the build platform. [17] However, nonadjustable factors like plume production or shielding gas flow may affect the properties of parts made using lower energy parameter sets. [45][46][47][48] Nonetheless, C D values in 0°samples were slightly higher than in 90°samples. This could be related to the lower sectional RDs in the X-Y plane; only one sample exhibited the opposite pattern. Furthermore, the different strut morphology and connectivity in the X-Y plane versus the Y-Z plane can influence the flow properties. HT as an influence factor in flow can be excluded as it has no morphological effect. Tensile behavior for material strength and Young's modulus under varying VE is consistent with previous research. [20,21,24,30] According to one of these studies, increasing porosity is associated with decreased tensile strength. [20] Decreasing VEs were also discovered to be responsible for lower sectional RD and worsening tensile properties in our study. Furthermore, as LDAs increased, so the sectional RD decreased, such as material strength and Young's modulus. With the exception of one sample, previously developed prediction models fit the presented results when the average sectional RDs of both directions per sample are compared to the calculated volumetric RD that takes into account both introduced VE and LDA. [17] As a result, it stands to reason that the tensile results of this study could be related to previously observed changes in morphological characteristics for DMs made from Haynes 282. [15,17] As the link of tensile properties and the decrease in RD is clear for increasing LDAs, this effect could be mitigated by the compensation approaches presented in a previous study. [17] Nonadjustable factors such as plume production or shielding gas flow may also have an impact on the repeatability of parts manufactured with lower energy parameter values. [45][46][47][48] However, not only does the RD differ due to modifying VE and LDA, the GS and morphology are also changing. Higher tensile properties have been associated with rising GSs in our results. This contradicts the Hall-Petch strengthening, in which smaller grains are associated with increased tensile properties. [44] We have seen, however, that increasing GS correlates with increasing RD. As a result, we cannot separate these effects, but morphological differences such as changing stress concentration factors, or strut connectivity, are more likely to be the main contributor. Therefore, larger GSs are most likely not the cause for improved mechanical properties, but they correlate with DM's morphology. Although it is more likely that the macroscopic effects dominate the tensile results, it is difficult to rule out microstructural changes for different VEs and LDAs based on the results presented. Also, the delta between 0°and 90°tensile results can be explained by differences in morphology seen in the EBSD results. Because of the more random structure, with fewer elongated struts and connectivity, 90°t ensile properties may be lower than comparable results from 0°s pecimens. Additionally, morphological differences may be the reason for the high scatter in the ε A results in the 90°test direction. Because GS is comparable in both 0°and 90°direction, it can be ruled out as a cause of the deviation between 0°and 90°orientation. However, it cannot be excluded that the differences in 0°and 90°orientations are due to various outer surface finishes for 0°( WEDM) and 90°(as-printed) samples. But it is likely that the large inner surface that remains as-printed is dominating during tensile testing. However, higher ultimate tensile strengths in horizontal (0°) orientation are a known behavior in AM bulk material and were also presented for Haynes 282. [43] Future research should concentrate on generating more comprehensive data on morphological, microstructural, and mechanical properties for single samples in order to eliminate single factors and improve explanation of behaviors.
When the fracture surfaces of the 90°and 0°bulk samples are compared to the outcomes of a prior study using a comparable HT for Haynes 282, it is clear that ductile fracture was present in both instances. [43] The size of the DM's and bulk material's dimples are comparable. [43] But DMs did not exhibit the necking that was evident in the bulk material tensile samples. [43]

Inconel 625 Hierarchical DMs
For the first time, the combination of DMs and lattice structures in hierarchical DMs was evaluated. Our study revealed beneficial results for such a hierarchical combination. Because low mechanical properties were observed for Haynes 282 DMs, further research with different cell types, strut dimensions, or even hollow struts could aid in the development of DMs from such materials. The same ductile fracture mode was observed in both lattice and DM strut as in a previous study using the same HT for Inconel 625 bulk materials. [36]

Conclusion
This study offers several new perspectives on the characteristics of DMs and hierarchical DMs: 1) the study of flow characteristics in two orientations (0°and 90°) revealed that the applied VE has a significant impact on the discharge coefficient of DMs. The VE effect on the vertical and horizontal tensile properties of DMs was also investigated. It has been demonstrated that σ u and E significantly and strongly correlate with the change in VE. The degree of correlation was same for ε A in the 0°direction. There was no significant correlation for ε A in the vertical (90°) orientation. The discharge coefficient decreased, and the mechanical strength increased with increasing VE. Because of this, there is a conflict between DMs' functional tasks and mechanical integrity, which must be resolved in the application design. 2) Along with declining LDA, the discharge coefficient also decreased. The tensile properties E and σ u also showed a definite sign of an LDA effect. Thus, mechanical strength and flow property showed significant correlations with the change in LDA. This, in combination with the known decrease in RD at increasing LDAs, underlines the relation of morphological, mechanical, and functional properties. [15] However, ε A did not correlate with changes in LDA due to high degree of scatter.
3) In comparison to 90°specimens, 0°flow samples had higher discharge coefficients. In addition, compared to 90°specimens, 0°samples showed higher mechanical values. For more sensitive applications, the slight variations in specimen processing might need to be considered. 4) It was discovered that the fracture mode of DMs made from Haynes 282 was ductile, similar to that of bulk material. 5) The applied HT increased the GS and changed its shape in the 0°and 90°orientations for all analyzed sections.
For the bulk material, the growth effect was more pronounced compared to DM. 6) Additionally, for denser DMs, the GS increases. The impact of changing micro-or macrostructure on the tensile behavior cannot therefore be determined. 7) The mechanical and functional performance of DMs with imprinted cubic lattices was examined. A promising materially interlocking connection between the DM and lattice structure was revealed by μCT scans. Furthermore, in the hierarchical DMs made from Inconel 625, ductile fracture mode was present for both DM and lattice struts. Hierarchical DMs exhibited substantial mechanical improvements compared to DMs with the same settings. Concurrent with the mechanical advances, lattice-structure-reinforced DMs provided slightly lower discharge coefficients compared to DMs with the same settings. 8) The potential of adjustable open-porous structures produced via PBF-LB/M for the application of DMs and hierarchical DMs in high-temperature applications such as LGTs was demonstrated. A wide range of products can benefit from customizable functional properties. For example, integral parts for transpiration cooling with open-porous and bulk sections can be created in a single process. However, the evaluated mechanical properties are poor, but they can be improved by incorporating lattice structures. This makes the use of the investigated structures appealing for higher loaded components.

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