Directed Energy Deposition of Low-Alloyed Steels: An Insight on Microstructural and Mechanical Properties

gears. Additive manufacturing technologies like directed energy deposition (DED-LB/M) allow for a fast and close-to-contour fabrication of sophisticated products without excessive waste of material. However, the DED-LB/M process cannot be considered as state-of-the-art for this group of materials. This study presents ﬁ ndings on the material properties of the additively manufactured low-alloyed steel Bainidur AM by means of DED-LB/M. This includes studies on the mechanical properties (hardness, compression strength) as well as the microstructural properties (scanning electron microscopy [SEM]). The microstructure in the as-built state appears like a bainitic – martensitic one with shares of retained austenite which is not fully transformed during cooling. As a differentiation is barely possible from the SEM images, a plethora of investigations is further used to assess the microstructure. As-built samples possess a good combination of ductility and hardness. Furthermore, the specimens are characterized by a good tempering stability up to 600 °C. This tempering stability is characterized by a homogeneous hardness of around 400 HV1 for all temperatures. In contrast, the conventionally hardened specimens show a drop-off in material hardness, further indicating the excellent material properties of additively manufactured Bainidur AM.

DOI: 10.1002/srin.202200925 Low-alloyed steels are used for a variety of different applications like bearings or gears. Additive manufacturing technologies like directed energy deposition (DED-LB/M) allow for a fast and close-to-contour fabrication of sophisticated products without excessive waste of material. However, the DED-LB/M process cannot be considered as state-of-the-art for this group of materials. This study presents findings on the material properties of the additively manufactured low-alloyed steel Bainidur AM by means of DED-LB/M. This includes studies on the mechanical properties (hardness, compression strength) as well as the microstructural properties (scanning electron microscopy [SEM]). The microstructure in the asbuilt state appears like a bainitic-martensitic one with shares of retained austenite which is not fully transformed during cooling. As a differentiation is barely possible from the SEM images, a plethora of investigations is further used to assess the microstructure. As-built samples possess a good combination of ductility and hardness. Furthermore, the specimens are characterized by a good tempering stability up to 600°C. This tempering stability is characterized by a homogeneous hardness of around 400 HV1 for all temperatures. In contrast, the conventionally hardened specimens show a drop-off in material hardness, further indicating the excellent material properties of additively manufactured Bainidur AM. microstructure formation and the material properties are missing within these works.
Further experiments were performed on the processing of this class of steels by means of PBF-LB/M in recent years. Schmitt et al., [18] Aumayr et al., [19] and Zumofen et al. [20] have shown that defect-free parts from this group of steels can be manufactured successfully using AM technologies. These low-alloyed steels possess a mixture of a bainitic-martensitic microstructure. This is further indicated by the material properties like hardness and strength which fall below the one of hardened reference specimens. [19,20] These investigations support the assumption that low-alloyed steels can be processed reasonably using laser-based technologies. In contrast to PBF-LB/M, the fabrication of low-alloyed steel components by means of DED-LB/M is barely studied. Ebrahimnia et al. [21] focused their work on the processability of 24CrNiMo. Up to two layers were manufactured additively. Single-layer specimens could be fabricated without significant crack or pore formation. Multilayer specimens, however, supported the crack formation. Wang et al. [22] investigated the laser metal deposition of a low-alloyed steel HSLA with a carbon content of approximately 0.07%. They have shown that the material can be processes successfully by means of laser metal deposition without any larger defects. Liu et al. [23] studied the laser metal deposition of the high-strength steel M300 which is characterized by a carbon content of around 0.40 wt%. The material could be fabricated without significant defects and possessed a martensitic-bainitic microstructure in the as-built state.
Goal of this work is to study the microstructure formation of the low-alloyed steel Bainidur AM when processing this material by means of DED-LB/M. Due to the high cooling rates and continuous intrinsic heat treatment, a martensitic-to-bainitic microstructure is assumed. [24] The microstructure formation will therefore be characterized by analyzing the material properties in the as-built and heat-treated state. Special importance is devoted to the tempering stability as bainite is known for a better tempering stability compared to martensite as shown in refs. [25,26].

Experimental Section
DED-LB/M experiments were performed on a 5-axis ERLAS 50237 machine (ERLAS GmbH, Germany) equipped with a 4 kW diode laser of type Laserline LDF 4000-4 (Laserline GmbH, Germany) with a characteristic wavelength between 940 and 963 nm. The laser spot size can be varied between 1 and 3 mm by using a zoom optic. As powder material, the case-hardening steel Bainidur AM (Deutsche Edelstahlwerke Speciality Steel GmbH, Germany) was used with a nominal particle size distribution between 45 and 90 μm. Figure 1 shows exemplary scanning electron microscopy (SEM) images of the Bainidur AM powder.
The powder material consists of mainly spherical and some nonspherical particles. Larger magnifications show that small satellites can be found on the surface of the particles, probably resulting from the atomizing process. The chemical composition of the powder material according to the supplier's certificate is listed in Table 1.
Experimentally determined d 10 , d 50 , and d 90 values are 58.8, 80.8, and 100.4 μm, respectively. Argon was used as both shielding and carrier gas. Laser-cut 16MnCr5 circular blanks with a nominal diameter of Ø 60 mm and a thickness of 3.5 mm were used as substrates for the DED-LB/M experiments. The blanks were inserted into a mounting fixture made from aluminum and were not clamped mechanically. All substrates were used in the as-delivered, hot-rolled condition.
Within this work, 12-layer cubic samples with an edge length of 12 mm in xand y-direction were manufactured additively. A meander-shape hatching strategy was used for generating the core of the specimens. This was followed by single contour weld tracks along the four edges of the specimen. After every layer, the fabrication sequence was shifted clockwise by 90°to homogenize the heat input into the specimen. The process parameters for both the hatching and the contour were maintained constant throughout the manufacturing process. Table 2 shows the investigated parameter combinations for the following experiments. While the feed rate and the spot size were maintained constant,  the laser power was varied to investigate the influence of varying energy inputs on the material properties. Modifying the feed rate would result in different build rates, which is undesirable when comparing the material properties and the influence of the intrinsic heat treatment. Powder mass flow, shielding gas, and carrier gas properties were kept constant at 2.60 AE 0.02 g min À1 , 20 L min À1 , and 4 L min À1 , respectively. The averaged weld track width was determined based on single weld tracks using the presented parameters. The overlap between adjacent weld tracks was set to 50% and maintained constant. Next, the additively manufactured samples were embedded in an epoxy resin, grinded, and polished for subsequent analysis of the material hardness. To analyze the microstructure formation, an additional etching step using a 3% Nital solution was performed.

Analysis of Microstructure Formation
In the first step, the relative density of the manufactured specimen was determined on nonetched cross sections. Therefore, images of the samples were made using an optical light microscope of type Olympus BX53M (EVIDENT Europe GmbH, Germany). The ratio of defect-free material and defects like pores or cracks was determined using binarization methods. This value resembles the relative density of the sample. Furthermore, magnified images of the etched cross sections were made to study the phase formation and orientation. SEM was used for analyzing the microstructure. A D8 Discover (Bruker 155 Corporation, Billerica, USA) system equipped with a Lynxeye 1D detector was used for X-ray diffraction (XRD) measurement of the retained austenite content.

Hardness Measurements
The hardness measurements are performed on nonetched cross sections. A KB30S (Hegewald & Peschke, Germany) was used for determining the hardness in the as-built state. The hardness of the heat-treated specimens was measured using a Qness indentation tester (ATM Qness GmbH, Germany). The hardness was determined in every layer, with a constant distance between two measurement points of the layer height tin z-direction and 2 mm y-direction. Eight measurements are performed along the y-direction for every specimen. An exemplary depiction of the measurement grid is provided in Figure 2.
For each specimen, three independent samples were manufactured and analyzed. The layer height is determined based on the height of the part and the number of manufactured layers.

Postprocess Heat Treatment
Additional postprocess heat treatment was performed on additively manufactured specimens to assess the tempering stability. Two different strategies were followed. On the one hand, the as-built specimens were exposed to a tempering heat treatment (AT) up to temperatures of 600°C. On the other hand, the additively manufactured samples were quenched and tempered (QT) at the same temperatures. The tempering parameters are shown in Table 3.
All austenitization heat treatments were performed in a furnace of type N 31/H (Nabertherm GmbH, Germany). Austenitization was performed at 920°C for 0.5 h. An N 30/85SHA (Nabertherm GmbH, Germany) oven was used for tempering the specimens. The specimens were tempered for 1 h. A nitrogen gas atmosphere was used for all heat treatment experiments.

Compression Testing
Furthermore, compression tests are performed on a universal testing machine of type QUASAR 100 kN (SCHÜTZ þ LICHT Prüftechnik GmbH, Germany) according to DIN 50106 using cylindrical samples. As the ratio of height-to-diameter shall be between 1 and 2, a specimen geometry with a diameter of 4 mm and a height of 6 mm was chosen. The specimens were machined from 12-layered specimens to allow a comparability with the hardness measurements, which were performed along build direction. The specimens were extracted from the center of the additively manufactured structure. After machining, the samples were polished. Three samples were machined from independent specimens for compression analysis.

Results and Discussion
The manufactured samples were analyzed regarding defect formation, phase formation, microstructural characteristics, and mechanical properties. Figure 3 shows the 12-layered specimen fabricated by means of DED-LB/M using the three different parameter combinations. All specimens could be manufactured without larger defects like cracks or pores, independent of the applied process parameter. The relative part density exceeds 99.9% in all cases. However, the specimens fabricated with a higher energy input are characterized both by a reduced dimensional accuracy and by a promoted etching indicated by the brown color in the top region. Both observations can be attributed to the heat agglomeration due to the layerwise process. For the lowest laser power, only the top three layers are characterized by a promoted brownish etching, whereas the lower layers are characterized by a bluish etching with darker regions at the bonding zone between two consecutive layers. The dark brown regions at the top indicate a fast cooling, which would correlate with a high material hardness. Increasing the laser power results in an enlarged size of the brownish etched zone, caused by the greater energy input during the manufacturing process. Furthermore, the brownish etched region appears brighter, appearing more like a ferritic structure. With an increasing size of this bright brown region, a reduced material hardness compared to the bluish parts of the specimens is assumed. Higher laser powers result in higher surface temperatures of the substrate. This effect can be linked very well with the etching response of the additively manufactured structure. The surface temperatures of the substrate indicate that the structure is continuously overheated. Correspondingly, the promoted heat accumulation of the substrate results in larger melt pools with increasing part height. This leads to an increased powder catchment within the enlarged melt pool, which explains the noticeable dimensional inaccuracies for higher energy inputs.

Hardness and Microstructure Formation
To further analyze and validate the material properties, the material hardness was measured for the additively manufactured Bainidur AM samples using the indentation tester. By determining the hardness in every layer, the influence of the build height and the corresponding intrinsic heat treatment on the mechanical properties is assessed. The corresponding hardness progressions are presented in Figure 4.
The maximum material hardness was measured in the top three layers of the specimens for laser powers of 600 W (455 HV1) and 700 W (440 HV1). For the highest laser power of 800 W, the highest hardness was determined close to the bonding zone of the substrate (third layer, 420 HV1). After that, a slight decrease in hardness was observed toward the top layer (390 HV1). This correlates well with the trends indicated by the etching colors in Figure 3. The differences in the maximum material hardness can be attributed to the different cooling rates and in situ preheating temperatures depending on the applied process parameters. Lower laser powers favor a faster cooling due to the smaller melt pool size, thus potentially resulting in a finer grain. [27] Furthermore, the decrease in material hardness along build direction for the highest laser power could be a cause of a sequential overheating. Due to the low thermal mass of the substrate, the excessive energy input will lead to higher in situ preheating temperatures. This assumption is supported by the finding that the specimens tended to bulge for higher energy inputs, without an increase in pore formation (see Figure 3). These promoted in situ temperatures will most likely result in a higher retained austenite content because the transformation to either martensite or bainite is not fully performed. This mechanism is temperature and time dependent and is therefore affected by the different processing times (due to the average weld track width) and the maximum temperatures (due to the energy input). Therefore, the obtained hardness values were correlated with the microstructure of the specimens. Figure 5 presents the SEM images of the cross sections for three different   The microstructures of the additively manufactured specimens appear similar at a low magnification of 3.0 K. Further zooming into the structures reveals differences for the different energy inputs. At the lowest energy input, a lath-like structure can be identified. The microstructure is characterized by precipitated carbides within the grain. Due to the fine dispersion of these carbides, probably cementite, the structure is either a martensitic or a lower bainitic one. Applying higher laser power (700 W) results in a minor change of the microstructure. The regions with the finely dispersed carbides within the lath are reduced. However, the mixture of the plate-and lath-like structure, which was also presented at a laser power of 600 W, is still evident. Contrary to these two lower laser powers, the highest applied laser power (800 W) results in an altered microstructure. The structure is characterized by an increased amount of the larger plates. These bright plates are separated by longish darker structures. This indicates that a partially (degenerated) upper bainitic structure might be present in the center of the specimen. [28] Moving toward the surface of the substrate, the hardness typically reaches its peak. This can be attributed to the faster cooling due to air quenching when finishing the build job. Figure 6 shows SEM images of the top region for the different laser powers.
Laser powers of 600 and 700 W result in a similar lath-like microstructure. The grains tend to precipitate very fine carbides. Furthermore, a globular-like structure can be identified within the grains for the lowest laser power. This might be some sort of globular bainite forming from inside these ferritic isles or a mixture of incomplete transformation products. [29] At a laser power of 700 W, a transformation toward a degenerated bainitic structure can be assumed. The granular structure can also be found within the ferritic cells and appears more pronounced compared to the lower laser power. Applying a laser power of 800 W results in a different microstructure compared to the two other energy inputs. At 800 W, a more isle-like structure could be identified. This structure appears like retained austenite which could not be transformed completely during cooling. One potential reason for this is the elevated temperature within the additively manufactured specimen. The higher preheating of the specimen most likely leads to an incomplete transformation of the austenite into bainite or martensite. The presumably increased retained austenite content helps in explaining the reduced material hardness in the top region of the specimen. Furthermore, the reduced hardness can be explained by the slightly coarser structure of the laths shown in Figure 6.

Analysis of Tempering Stability
Bainite is typically characterized by a higher tempering stability compared to a martensitic microstructure. [25] As the microstructure of the DED-LB/M specimens is difficult to assess due to its fine size, additional studies on the tempering stability of as-built (AT) and hardened (QT) samples were performed. This study was used as an additional analyzing method for differentiating the as-built microstructure from the hardened microstructure. Figure 7 presents the results of the experimentally determined hardness values and the measurement scheme for the different tempering temperatures.
The results show that the hardness in the as-built state remains constant for all tempering temperatures as high as 600°C. At room temperature, a hardness of around  Quenching and tempering the samples results in an increased material hardness for low tempering temperatures. A maximum hardness of 448 AE 1 HV was obtained in the as-quenched state. However, a continuous decrease in hardness can be observed for temperatures exceeding 150°C. The hardness (385 AE 4 HV1) falls below the one of the as-built specimens for the highest tempering temperature of 600°C. This trend is characteristic for martensitic microstructures due to the poorer tempering stability of martensite compared to, e.g., bainite. [26] The drop-off below the hardness of the as-built and tempered specimens makes it unlikely that the microstructure after DED-LB/M is a tempered martensitic one. In the next step, the microstructural characteristics were analyzed both for the as-built and tempered as well as the hardened and tempered state by means of SEM. Figure 8 shows SEM images of the microstructure in the different heat-treated states. The microstructure is characterized by finely precipitated carbides within the lath and remains mostly similar up to a tempering temperature of 400°C. These finely precipitated carbides could be an explanation for the excellent tempering stability of the material. Tempering at 600°C results in slight changes as the previously promoted isles of retained austenite tend to decompose. The retained austenite content was in the range of 10 AE 2% within the center of the specimen prior to tempering. After tempering at 600°C for 1 h, the retained austenite content fell below 3%. Additionally, the microstructural properties of the quenched and tempered specimens were analyzed by means of SEM. The microstructural characteristics are presented in Figure 9.
Quenching results in a destruction of the DED-specific microstructure. The quenched specimens possess a more promoted lath-like martensitic structure. Furthermore, the boundary of the lath-like martensitic structure is emphasized, which leads to a clear structure. Exposing the specimens to elevated temperatures results in a tempered martensitic microstructure. The lath boundaries are no longer as promoted as before as the carbon  diffuses toward to grain boundaries. This helps to explain the reduced material hardness. Comparing the hardness after tempering (AT) and after hardening and tempering (QT) shows that the underlying microstructure in the as-built state is neither a fully martensitic nor a tempered martensitic one. This is supported by the better tempering stability up to 600°C and the texture of the underlying microstructure. It is more probable that the marginal decomposition of the retained austenite and the beneficiary ferritic microstructure due to finely precipitated carbides [30] result in the homogeneous hardness along different tempering temperatures.

Compression Testing
Finally, the compression strength of the samples was analyzed. Compression testing can be seen as the inverse version of tensile testing and allows to assess the material properties for tooling applications. The maximum deformation rate for the abort criteria was set to 70%. All samples surpassed this deformation rate of 70% without breaking, which indicates the excellent ductility of additively manufactured Bainidur AM. The experimental results for the averaged values of the three different process parameters are presented in Figure 10.  The highest compression strength (3880 AE 82 N mm À2 ) was observed at the lowest laser power. Increasing the energy input resulted in a continuous decrease in compression strength, as the lowest strength (3755 AE 27 N mm À2 ) was determined at the highest laser power. The corresponding yield strengths were 1402 AE 9 MPa (600 W), 1390 AE 11 MPa (700 W), and 1352 AE 11 MPa (800 W) for the three different laser powers. Overall, the determined compression and yield strength are very similar for the three different parameters. This correlates well with the comparable hardness values for the laser powers. The material properties are similar, even though a slight decrease in performance can be assumed for higher laser powers, both regarding the yield strength and the material hardness. The obtained yield strengths range between the ones of the additively manufactured steel M300 (approximately 1200 MPa) [31] and H11 tool steel (1770 MPa). [32] As both these materials are commonly used for tooling, a high potential for Bainidur AM as a base material for tooling applications can be assumed. A potential process route could include the deposition of a wear-resistant coating already during DED-LB/M.

Conclusion
This work presents thorough investigations on the material properties of additively manufactured Bainidur AM by means of DED-LB/M. It was shown that the selected process parameters and the corresponding process-intrinsic heat treatment affect the microstructure formation along build direction due to altered cooling conditions. A change from a lath-like toward a plate-like microstructure was observed for increased laser powers. This correlates with the hardness measurements, as the maximum hardness was observed both in the top layers and for the lowest laser powers. The as-built microstructure of the core region is further characterized by an excellent tempering stability up to 600°C which indicates that the underlying microstructure should be at least partially bainitic. Bainidur AM also possesses an excellent yield strength for a low-alloyed steel around 1300-1400 MPa depending on the applied process parameters, linked with an excellent ductility. The high yield strength opens potentials in tooling, especially when additionally applying a wear-resistant coating during DED-LB/M. This shows the enormous potential of additively manufactured Bainidur AM as a base material due to its excellent properties, which are the consequence of the bainite-like microstructure. Future work will focus on the influence of different parameter combinations like higher or lower feed rates during DED-LB/M on the resulting microstructure. Furthermore, key properties like tensile strength will be determined to generate a database for the later use of this material.