Additive manufacturing (AM) is an emerging production technology with enormous potential to replace and supplement conventional manufacturing processes. Especially in the field of metal AM, recent developments in equipment as well as improved part quality have allowed advancements from rapid prototyping of single pieces to final part production. Among the various powder-based metal AM techniques, laser powder bed fusion (LPBF) is currently the most widely used method, as it allows for higher geometrical flexibility than laser metal deposition (LMD) and higher resolution compared to electron beam melting (EBM) [1
]. Geometrical freedom, reduced material waste, energy usage, and high degrees of automation are additional advantages of LPBF that contribute to meeting global challenges such as increased individualization, environmental friendliness, and digitalization. Furthermore, elemental segregation can be strongly reduced due to high cooling rates, which makes a large spectrum of materials processable [1
Whereas the various AM techniques enable high degrees of freedom in geometrical design, the methods are rather inflexible with respect to material inputs. So far, mostly pre-alloyed powders or powder blends consisting of two similar materials, e.g., Ti and TiC or TiB, have been utilized to guarantee process stability and chemical and microstructural homogeneity in reproducible properties. Additive manufactured TiC/TiB-reinforced Ti matrix nanocomposites have been used in biomedical applications [6
]. The use of diversified powder blends would open up new degrees of freedom for powder-based AM, especially in alloy design. Utilizing mixtures of multiple powders would allow for fast and simple variations in chemical composition, which would enable the rapid design and screening of new alloys. Therefore, powder-based AM techniques in combination with powder blends might be a new solution for rapidly designing and screening chemically complex materials such as multi-principal element alloys (MPEAs), also known as high-entropy alloys (HEAs) [9
MPEAs are a relatively new class of alloys, which instead of relying on one base element contain at least three principal elements with fractions of 5–35 at% each [9
]. Hence, a vast space for chemical compositions and properties is possible, which has stimulated intensive research in this field. With this concept, mechanical and functional property combinations that cannot be found in conventional alloys may be achievable. These property combinations can be varied within a wide range by varying the concentration of one or more of the elements [9
Earlier studies carried out by Haase et al. [14
] introduced a robust methodology utilizing thermodynamic modeling and rapid screening using LMD for new MPEAs. The methodology is based on using powder blends as an input material to enable rapid and resource-efficient screening of MPEAs. However, LMD is very limited with respect to producing parts with complex geometry and is prone to defects, which motivates the application of LPBF instead [15
]. The possibility of using powder blends as an input material for LPBF has been successful in various previous studies [16
]. For example, the mechanical properties of high-manganese steels were adjusted by adding elementary Al powder to pre-alloyed steel powder [16
]. However, this approach was not tested with powder blends consisting of more than two powders or different powder morphologies from their production processes (e.g., inert gas atomization, water atomization, grinding).
The goal of the present study is the application of the rapid alloy development methodology introduced by Haase et al. [14
] for MPEAs in the LPBF process. First, the Al-C-Co-Fe-Mn-Ni system was exemplarily chosen as the considered material in the present study. This chemically complex system demonstrated an extreme case for powder blends in LPBF, since a variety of different elemental powders, both metallic and nonmetallic, with differing morphologies were used. Secondly, a base composition without C was qualified for the LPBF process to produce fully dense parts. Different energy densities were selected to investigate the homogeneity of the produced samples. Thirdly, the fabricated alloys were evaluated regarding their processability, microstructure, and mechanical properties by using optical microscopy, electron backscatter diffraction (EBSD), energy-dispersive X-ray spectroscopy (EDS), and tensile testing. Finally, parts with complex geometry, i.e. lattice structures, were produced to demonstrate the applicability of powder blends in LPBF. The lattice structures were subsequently compression-tested to evaluate their energy absorption capability. Based on the findings, the correlation between process, microstructure, and mechanical properties is discussed, and the application of elemental powder blends in LPBF is critically evaluated.
In the following, the adaption of the rapid alloy development methodology from LMD to LPBF as well as the processability of powder blends and the influence of process parameters on the mechanical properties are discussed.
After process parameter optimization, several parameter sets with relative densities above 99.5% and a homogenous elemental distribution were determined for BASE and BASE + 0.6C (Figure 4
, Figure 5
and Figure 6
). These results showed the successful transition of the rapid alloy development methodology first applied in LMD [14
] to LPBF. With LPBF, less elemental segregation takes place during solidification due to the higher cooling rates compared to LMD [18
]. Therefore, the possibility of processable materials is larger than in LMD, and better mechanical properties can be achieved because of grain refinement [28
]. However, the powder consumption in LPBF is higher compared to LMD due to the necessity of filling the building chamber. With respect to flexible alloy modification, LMD enables in-situ mixing of various powders, chemical gradients within specimens/parts, and efficient powder consumption [1
]. Overall, LPBF is more suitable for alloy design due to higher reproducibility and lower defect density compared to LMD.
The powder blends utilized in the process contained up to six different elemental powders with varying morphology. The Ni, Al, and Fe powders were inert gas-atomized, which is the most used production technology for the synthesis of powders used in LPBF. The characteristics of inert gas-atomized powders are a spherical shape, high flowability, and high purity [32
]. Mn and C were produced by grinding, resulting in flake-shaped particles with poor flowability and thus poor recoating behavior during LPBF [32
]. The Co powder was water-atomized and revealed an irregular splash-shaped morphology. Nevertheless, the Co powder still retained reasonable flowability. The bulk density of the powders was in the range between 0.544 g ml−1
(C) and 4.545 g ml−1
(Ni). The particle size distribution (PSD) for all powders was in the range of 10 and 45 µm, except for Mn and C, where particles with a diameter below 10 µm were present, which reduced the flowability and thus influenced powder recoating behavior. The flowability (defined by the average avalanche angle) was between 48° for Co and 62° for C. As evidenced by the high relative densities achieved (Figure 4
), the powder blends revealed a suitable recoating behavior. Therefore, the poor flowability of powders (C, Mn) could be compensated for by blending them with powders characterized by high flowability (Fe, Ni, and Al).
The absorption tested for the used wavelength of 1064 nm was determined to be 48% for Al and up to 90% for C. Therefore, the energy input into the powder varied significantly between the different powders. This could cause overheating of powders with higher absorption as well as incomplete melting of powders with low absorption. However, once the melt pool was fully developed, the absorption remained constant and the influence of the different absorption coefficients vanished due to a preflowing melt pool [33
]. The use of elemental powders led to higher evaporation, since the elements with lower melting temperatures melted first and could overheat, whereas the elements with higher melting temperatures showed delayed transformation into the liquid phase [34
] (Table 3
). Consequently, Mn and Co had the highest and lowest evaporation rates due to their evaporation temperatures, respectively. This behavior was validated by EDS mapping, as shown in Figure 6
. The used elemental manganese powder showed high oxidation and therefore a high oxygen content due to the production process. The oxygen introduced by the manganese powder resulted in the formation of Al-oxides due to the high affinity of Al to oxygen [35
]. A low fraction of Al-oxides could be observed in the EDS element mapping, as shown in Figure 6
. The Al-oxide clusters are shown as small white dots in the aluminum mapping. Nevertheless, an overall homogenous element distribution could be observed in the LPBF-produced samples using powder blends. This proved in general the applicability of powder blends in LPBF, even for complex powder blends consisting of up to six different powders with varying morphologies and PSDs.
The results showed that dense parts could be produced using a broad spectrum of process parameters. This allowed for the application of a wide processing window, with EV
in the range between 60 and 240 J mm−3
for the investigated alloying system. A higher energy input resulted in higher homogeneity of the elemental distribution due to enlargement of the melt pool size by means of width and depth, which resulted in a higher number of powder particles in the melt pool [36
]. The calculated particle amount contained in the melt pools shown in Figure 5
was 15 for Figure 5
b, 57 for Figure 5
d, and 386 for Figure 5
f. This enlarged melt pool enabled the enhanced mixing of elements and at the same time facilitated the remelting of previously solidified layers. Hence, the higher the energy density was, the higher the homogeneity of the elemental distribution due to complete melting and mixing of the various elemental powder particles was. In Figure 6
a, partially molten particles and heterogeneously composed areas can be found. Furthermore, the increased energy input also promoted higher homogeneity due to a higher melt pool temperature [36
]. As a consequence, less nonfused material was observed (Figure 5
and Figure 6
Homogeneity also affected the mechanical properties. Whereas specimens produced with a higher EV
resulted in increased elongation, samples manufactured with a lower EV
revealed reduced fracture strain (Figure 7
). Presumably, the loss of ductility was associated with locally heterogeneous deformation behavior, as well as notch effects in the vicinity of unmolten particles [37
]. Since the process development allowed for the production of samples with similarly high relative densities, the detrimental effects of the remaining porosity could be ruled out as a cause of reduced ductility. However, it must also be noted that very high energy densities may result in pronounced evaporation/gas entrapment and thus more gas pores [38
]. These pores usually contribute to deteriorated mechanical properties [39
]. Hence, an optimum process window avoids the effects associated with very low and very high input energies.
Furthermore, the rapid alloy development methodology introduced was used for material screening. As an example, C was added to the BASE alloy to tailor the mechanical properties. C was chosen because it is known as a very efficient solid solution strengthening element in fcc alloys. By adding 0.6 wt% C, the mechanical properties of the BASE alloy were improved with respect to both strength and ductility (Figure 7
). To exclude further microstructural modifications, the process parameters were kept constant for both BASE and BASE + 0.6C. A usable parameter set was determined within the process parameter development (Figure 3
and Figure 4
). The development showed similar results for relative density for both alloys using the same process parameter sets. Nevertheless, a negligible influence of the process parameters on the relative density was determined. Surprisingly though, the application of powder blends containing elemental C powder with poor flowability and flake-shaped morphology did not deteriorate processability. Furthermore, pure C particles and large carbides were not observed, which suggested the complete dissolution and homogeneous distribution of interstitial C atoms. Therefore, the usage of complex powder blends to tailor material properties to design materials by and for AM can be confirmed.
To demonstrate the potential to produce geometrically complex structures with powder blends, f2
lattice structures were synthesized. These delicate structures are a suitable measure for the reliability of the material produced from powder blends due to the sensitivity of the thin struts to material homogeneity. For instance, most common strut sizes for LPBF-manufactured lattice structures are 500 µm in diameter [21
], which means that a laser beam melts approximately 350 particles on the top layer per strut. An energy input of 173 J mm−3
, considered to be the optimal value for the BASE + 0.6C alloy, was chosen, resulting in ductile and homogeneous deformation behavior (Figure 8
). Compared to the results of Kies et al. [18
], where high-manganese transformation-induced plasticity (TRIP) and twinning-induced plasticity (TWIP) steels were tested, a similar specific energy absorption was achieved. Furthermore, compared to the benchmark material 316L, the specific energy absorption of the tested MPEAs was increased by approximately 75% [18
]. Therefore, the usability of the rapid alloying methodology was additionally validated due to successfully manufacturing lattice structure specimens with reliable deformation behavior.