Selective Laser Melting (SLM) is one of the very advantageous AM processes for metallic materials in comparison to subtractive manufacturing due to its ability to create highly complex geometries, low amount of waste material, reduced lead time from design to testing, simplified supply chains, decreased number of joining processes possibly leading to an improved part life, etc. Moreover, unlike conventional manufacturing, the SLM process does not necessitate design of special tools / molds and directly starts from the CAD file of the part to be produced as shown in Fig. 1. SLM has recently been increasingly adopted for manufacturing functional end-parts in diverse demanding industries such as aerospace, defence and biomedical. However, to explore the full potential of SLM, some barriers still need to be overcome. One of the very important limitations is the high number of process parameters that affect the part performance as well as the low build rate. The process parameters do not only have direct influence on the part performance but also their interactions due to the physical phenomena occurring during the SLM process play a significant role on the outcome. Although the machine vendors generally provide or sell an initial set of process parameters optimized for each material, these may be far from optimum when multiple criteria are taken into account. Generally, the given process parameters are optimized taking density into consideration since mechanical properties severely deteriorate if excessive porosity is present. However, due to inherent nature of the SLM process, other criteria become dominant on the part performance as well. For example, due to the high cooling rates encountered in the process, residual stresses leading to part deformations or even cracks during the process become one of the problems. Moreover, especially for internal features, it is generally difficult to improve the surface quality after the process is completed. Without reaching the desired levels of surface roughness, the advantage of creating very complex geometries stays limited. Thus, as-built surface roughness is also critical. Moreover, productivity needs to be taken into account as one of the bottlenecks of SLM. The productivity, which can be expressed in terms of build rate, depends on many factors such as material properties, machine/laser configuration (multiple lasers, bi-directional powder coating, etc.), part orientation, scanning parameters and nesting. Additionally, the selected layer thickness is very critical in terms of the build rate due to its direct effect on the number of layers.
In the last decade, the SLM studies on the 17 − 4 PH stainless steel (SS) in the literature have been increased. Murr et al. has presented one of the most comprehensive studies on the SLM of this material studying martensitic or mostly austenitic phase powders which were produced by atomization in argon or nitrogen, respectively. Those powders were than used in the SLM process using Ar or N2 atmospheres. It was concluded that the phase in the final part is the same as the powder phase (austenitic or martensitic) when N2 is utilized. The final parts exhibit a martensitic structure with either an austenitic or martensitic pre-alloyed 17 − 4 PH SS powder provided that argon is used in the SLM [1]. Auguste et al. have investigated powder batches from different suppliers and applied various heat treatments. They concluded that the chemical composition of different powder batches leads to different phases in the end, one being mostly ferritic while the other being mostly martensitic leading to different mechanical properties [2]. On the other hand, the studies involving optimizing process parameters for the SLM of 17 − 4 PH SS are quite limited. Mahmoudi et al. have investigated the mechanical properties and microstructural characterization to understand the effect process parameters including build orientation as well as thermal history and applied heat treatment. However, the process parameters were only studied as “default and optimized” sets without giving any further information [3]. In another study by Hu et al., the effect of scan speed, hatch distance and layer thickness was studied with single factor experiments on the density and microhardness [4].
Although N2 is often used in the gas-atomization of stainless steels, it is not inert. In addition to its role to stabilize the FCC (face-centered cubic) austenite phase, it substitutes for C in various carbide phases leading to the formation of other carbide/nitride phases, similar to C [5]. Moreover, during the SLM process, the lower conductivity of argon in comparison to nitrogen leads to martensitic products from either austenitic or martensitic 17 − 4 PH SS powder [3]. In addition to the processing gas, the material composition of the powder has a significant impact on the final parts’ microstructure and thus mechanical properties for 17 − 4 PH [2, 6, 7]. The Creq/Nieq value calculated by WRC-1992 equation varies based on the volume fraction of residual delta-ferrite. The study by Vunnam et al. has concluded that a lower Creq/Nieq value results in martensite formation and a less retained delta-ferrite after the SLM process [7].
The difference in mechanical properties obtained with SLM in comparison to conventional manufacturing is generally attributed to the layerwise nature of the process leading to finer and elongated grains as well as high cooling rates. However, the difference in the 17 − 4 PH SS cannot be only explained with grain refinement due to high cooling rates. Some researchers point out the presence of a high volume fraction of retained austenite within the martensitic alloy leading to significant difference in resulting mechanical properties [9]. In the literature, the range of retained austenite severely changes from 5–95% although using very similar process parameters generally optimized for maximum density. The factors influencing the amount of retained austenite have only partially been addressed in the literature. AlMangour and Yang has concluded that fine grains led by the rapid cooling rates of the SLM process reduces the austenite to martensite formation (Ms) temperature [10]. Additionally, a high nitrogen content in the material composition coming from the gas atomization can further reduce this temperature [11]. This leads to a significant volume fraction of retained austenite in the as-built material at room temperature. During deformation, due to its being a metasable phase, the retained austenite transforms to martensite bringing an additional complication with respect to measurement of retained austenite fractions. This is due to the fact that some sample preparation methods may result in local deformations and consequent martensite formation at the surface [12]. By applying a heat treatment above austenite transition temperature, the volume of retained austenite can be reduced by transforming into martensite [13]. It is not easy to control the cooling rate in SLM by simply changing the process parameters. This is also demonstrated in the study by Gu et al., which addressed the change of the amount of retained austenite by varying laser power and speed [13]. It shall also be noted that XRD data may not be sufficient to differentiate the martensite having a body centered tetragonal (BCT) structure and BCC ferrite phase. This is mainly due to very low magnitude of the lattice distortion in the martensite in stainless steels having a C level smaller than 0,02 wt.% [7]. This necessitates the use of EBSD for phase differentiation.
Table 1 summarizes the reported work for the obtained microstructure/mechanical properties of the as-built specimens from 17 − 4 PH SS. The variability in the reported microstructures and obtained mechanical properties is much higher when compared to other materials used in SLM. The hardness values range from approximately 150 to 350 HV while reported tensile properties exhibit a high scatter as shown in Fig. 2 mainly depending on their microstructure. UTS/YS ratios change from 1 to 2 in different studies.
Table 1
Comparison of various studies on SLM of 17 − 4 PH Stainless Steel (approximate values are given for some data)
Ref. | Atomization Gas | Processing Gas | Mat. Comp. Ratio (Cr)eq/(Ni)eq | Hardness [HV] | Microstructure | Tens. Prop. UTS/YS Elongation |
[1] | Argon | Argon | Not given | 30 HRC | Completely martensitic | Not given |
[1] | Nitrogen | Argon | Not given | 32 HRC | Completely martensitic | Not given |
[1] | Nitrogen | Nitrogen | Not given | 22 HRC | Austenitic components with roughly 15% martensite | Not given |
[2] | Not given | Argon | SLM 2,04 | 330 HV | Mainly martensitic | 1100/500 |
[2] | Not given | Argon | TLS 2,41 | 300 HV | Ferritic with martensitic structure and retained austenite | 900/850 |
[3] | Not given | Argon | 3D Systems SLM powder | 300–350 HV | Depending on the dwell time, the volume fraction of austenite 13–26%. | 950/650 1050/625 |
[4] | Not given | Argon | Hengji (10–74 µm) | 260 ± 20 HV | Alpha and gamma phases present | 1100/640 |
[5] | Nitrogen | Nitrogen | EOS StainlessSteel GP1 | Not given | A significant fraction of retained austenite in the as-built condition, > 90 % | Not given |
[6] | Not given | Argon | Powder1 (d90 < 16µm) | 278 ± 57 HV | Alpha phase 38% vol. Gamma phase 62% vol. | 880/614 |
[6] | Not given | Argon | Powder2 (d90 < 25µm) | 226 ± 69 HV | Alpha phase 6% vol. Gamma phase 94% vol. | Not given |
[7] | Argon | | Powder A 2,76 | 277.3 ± 10 HV | Columnar grains, small martensitic laths | 763/607 22.3 ± 0.7 % |
[7] | Argon | | Powder B 2,65 | 330.7 ± 9 HV | | 871/699 19.4 ± 1.0 % |
[7] | Argon | | Powder C 2,43 | 333.0 ± 5 HV | A fine grain microstructure | 917/723 10.9 ± 0.9 % |
[8] | Argon | Not given | Powder A 3,88 | Not given | Ferritic grains and visible martensitic grain structures | 900 MPa XY 800 MPa Z |
[8] | Argon | Not given | Powder B Ar 3,89 | Not given | Ferritic grains | 900 MPa XY 800 MPa Z |
[8] | Nitrogen | Not given | Powder B N2 3,41 | Not given | Large ferrite grains accompanied by a fine and equiaxed austenite grains | 1050 |
[8] | Argon | Not given | Powder C- 3,31 | Not given | Contains largely martensite | 1050 |
[12] | Argon | Argon | 2.77 | 380 HV | 72% austenite, 28% martensite | 1300/600 |
[13] | Nitrogen | Argon | EOS | Not given | More than 96% austenitic | Not given |
[13] | Nitrogen | Nitrogen | EOS | Not given | Completely austenite | Not given |
[13] | Argon | Argon | LPW | Not given | Mostly martensite | Not given |
[13] | Argon | Nitrogen | LPW | Not given | Mostly martensite | Not given |
[15] | Nitrogen | Nitrogen | | 310 HV | Mostly martensite | Not given |