DOI: https://doi.org/10.21203/rs.3.rs-428820/v1
Although Selective Laser Melting has become attractive in industrial applications seeking a high geometrical complexity with short lead times and customization, some bottlenecks still exist for wider adoption. Build rate is one of them while the high number of process parameters and their interactions easily exceeding hundreds which affects the part performance is the second. The machine manufacturers supply parameter sets generally optimized for maximum density leading to good mechanical properties. However, other factors need to be considered in process development. This study aims at increasing the build rate of at least 2 times for 17-4 PH stainless steel without any significant effect on the density, surface quality, material composition, mechanical properties and residual stresses. The results show an excessive ultimate tensile strength to yield strength ratio in comparison to reported literature which is attributed to the double yield phenomenon mainly attributed to the phases present in the microstructure as a result of powder chemical composition and processing gas. Thus, it is concluded that powder chemical composition and processing gas are much more effective on the outcome while the process parameters with an increased build rate do not significantly change the results provided that almost full density is reached.
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.
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 |
The used powder material was obtained from SLM Solutions having almost spherical particles with some minor satellites. For density evaluations, cubic specimens having dimensions of 15x15x15 mm were built on an SLM Solutions SLM 280 machine under nitrogen as protective atmosphere using a scan strategy of stripe scanning. The layer thickness was doubled in comparison to the initial set of parameters and kept constant at 60 µm while the effects of the laser power (200-275-350 W), scan speed (600-800-1000-1200 mm/s) and hatch distance (0.09-0.12-0.15 mm) were tested with a full factorial test strategy. After the production was complete, the specimens’ densities were measured by the Archimedes’ method taking the reference density as 7.8 g/cm3 while the surface quality was assessed by Mitutoyo SJ-310 surface profilometer. Some parameters sets from the first batch leading to the lowest porosity values yielding a good level of build rates were identified and new specimens were produced in the second batch to test the hardness, tensile properties and deformations due to residual stresses as well as to investigate the material composition and microstructural features in comparison to initial parameter set with 30 µm of layer thickness (see Table 2). The tensile properties were tested per EN 2002-001: 2005 while the bridge curvature method (BCM) was used to quantitatively compare the deformations led by the residual stresses after the parts were removed from the base plate [18]. The microhardness tests were accomplished on with a loading of 1 kgf. Optical Microscopy (OM) and Scanning Electron Microscopy (SEM) were used to reveal the microstructural defects and features. The specimens were firstly etched with Vilella’s reagent for microstructural investigation and then electro-etched to clearly observe melt pool boundaries. The chemical compositions were obtained by spectral analysis on Spectromax and compared to the values specified in ASTM A564 while the EBSD analysis was carried out on a QUANTA 400F Field Emission SEM.
Par. Set # | Layer Thickness [µm] | Scan Speed [mm/s] | Laser Power [W] | Hatch Distance [µm] |
---|---|---|---|---|
B2-St-30 | 30 | 800 | 200 | 120 |
B2-1-30 | 30 | 1000 | 350 | 150 |
B2-2-60 | 60 | 1000 | 350 | 150 |
B2-3-60 | 60 | 1000 | 350 | 120 |
B2-4-60 | 60 | 1200 | 275 | 90 |
The relative density results of the first batch specimens range from 86–99.5% for parameters using a layer thickness of 60 µm whereas the initial parameter set for a layer thickness of 30 µm leads to a density result of 99.8%. The variation of the relative density with respect to the scan speed at different laser power and hatch distance values are depicted in Fig. 3 with corresponding trend lines and R2 values. In Fig. 3a where a hatch distance of 90 µm is used meaning that consecutive scan lines are closely placed, very low porosity values are obtained. All parameter sets with a hatch distance of 90 µm led to a porosity of less than 1% except the one with the lowest energy density (200 W laser power, 1200 mm/s scan speed). In Fig. 3b where a hatch distance of 120 µm is used, it is seen that when the linear energy density (the ratio between the laser power and scan speed) is sufficiently high, good densities above 99% can be obtained. However, as the linear energy density is lowered then the porosity values start to increase as demonstrated with a laser power of 200 W giving the lowest energy density in the tested range for Fig. 3b. The effect of lowering the laser power is more pronounced when the hatch distance is set to 150 µm meaning that the consecutive laser lines are sparsely placed in comparison to other values as demonstrated in Fig. 3c.
Since the aim of this study is to increase the build rate by doubling the layer thickness, the relative densities with respect to the indicative build rates, calculated as the multiplication of the scan speed, hatch distance and layer thickness, are presented in Fig. 4. The initial parameter set with a layer thickness of 30 µm gives the highest density at a rather low build rate, i.e. 4.3 mm3/s, as shown. Most of the parameter sets utilizing a high laser power (275–350 W) give a density above 99% up to an indicative build rate of 8.6 mm3/s. However, reducing the laser power to 200 W almost totally eliminates high densities above 99% with high build rates.
As a result of the first batch aiming to identify the suitable parameter sets to provide increased build rates without significantly changing the density, the parameter sets with a layer thickness of 60 µm given in Table 2 are selected for further testing as well as two sets of reference parameters with a layer thickness of 30 µm. The measured microhardness values with these parameter sets are presented in Table 3. It can be concluded that the variation of the process parameters does not significantly alter the obtained microhardness results. Moreover, they are very high in comparison to other reported hardness test results (see Table 1).
Parameter set # |
Layer Thickness [µm] |
Scan Speed [mm/s] |
Laser Power [W] |
Hatch Distance [µm] |
Relative Density [%] |
Indicative Build Rate [mm3/s] |
Microhardness [HV] |
BCM Deviation [°] |
---|---|---|---|---|---|---|---|---|
B2-St-30 |
30 |
800 |
200 |
120 |
99,81% |
2,9 |
355,7 ± 6,7 |
0,16 |
B2-1-30 |
30 |
1000 |
350 |
150 |
99,74% |
4,5 |
362,7 ± 8,7 |
0,15 |
B2-2-60 |
60 |
1000 |
350 |
150 |
99,65% |
9,0 |
344,0 ± 6,9 |
0,14 |
B2-3-60 |
60 |
1000 |
350 |
120 |
99,49% |
7,2 |
368,0 ± 2,7 |
0,28 |
B2-4-60 |
60 |
1200 |
275 |
90 |
99,48% |
6,5 |
362,7 ± 5,7 |
0,21 |
It is well known that different heat inputs may lead to different levels of residual stresses in the SLMed parts. As a fast and easy method enabling qualitative comparisons, BCM is utilized in this study. When the bridge geometry is produced on the base plate, it is cut off by wire electro discharge machining (WEDM), it is curled up because of the relaxation of the residual stresses. By measuring this angle of curling using the planes at the bottom of the bridge, a quantitative indicator is obtained. The curling angles with different process parameter sets are shown in Table 3 in the last column. Although some parameters lead to a smaller angle, all measured angles are very low in comparison to other curling angles obtained in other studies in the literature [15, 17]. For example, the standard parameter set in [18] led to a curling angle of 1,3° for AISI 316L stainless steel whereas the maximum curling angle in this study is only limited to 0,28°. Thus, it is concluded that changing the layer thickness and increasing the productivity does not significantly change the residual stresses in the part within the tested range.
It is observed that the change of the process parameters does not influence the chemical composition of the produced specimens according to the spectral analysis results per ASTM A564 as presented in Table 4. Regarding the elements of C, S, O and N, LECO analytical elemental analysis was performed per ASTM E 1019. Each specimen was measured three times and the average of these measurements were taken to check its conformity per specifications. It is seen that none of the tested parameter sets lead to non-conformance as shown in Table 5.
B2-St-30 |
B2-1-30 |
B2-2-60 |
B2-3-60 |
B2-4-60 |
||
---|---|---|---|---|---|---|
C |
Max 0,07% |
0,04% |
0,04% |
0,04% |
0,04% |
0,04% |
Si |
Max 1% |
0,70% |
0,69% |
0,68% |
0,68% |
0,70% |
Mn |
Max 1% |
0,37% |
0,37% |
0,39% |
0,40% |
0,39% |
P |
Max 0,040% |
0,01% |
0,01% |
0,01% |
0,01% |
0,01% |
S |
Max 0,030% |
0,00% |
0,00% |
0,01% |
0,00% |
0,01% |
Cr |
15–17,5% |
15,51% |
15,54% |
15,56% |
15,55% |
15,52% |
N |
Max 0,10% |
0,08% |
0,07% |
0,07% |
0,07% |
0,07% |
Fe |
Balanced |
74,60% |
74,50% |
74,40% |
74,40% |
74,50% |
Ni |
3–5% |
4,26% |
4,24% |
4,21% |
4,24% |
4,22% |
Cu |
3–5% |
4,09% |
4,13% |
4,21% |
4,20% |
4,18% |
Nb |
0,15 − 0,45% |
0,17% |
0,16% |
0,16% |
0,16% |
0,17% |
Ta |
< 0,02% |
< 0,02% |
< 0,02% |
< 0,02% |
< 0,02% |
The microstructures of the specimens produced within the second batch are investigated. As shown in Fig. 5, the microstructures with 30 µm layer thickness do not contain any residual pores. The etched micrographs indicate that the melt pool depths are much higher than the selected layer thickness shown with yellow lines. A higher laser power with B2-1-30 having the same layer thickness in contrast to B2-St-30 leads to much deeper and narrower melt pools.
In Fig. 5, the optical micrographs with 60 µm of layer thickness are also demonstrated with different parameter sets. The ratio of the melt pool depth to the layer thickness in these parts is smaller compared to parts made with a layer thickness of 30 µm. However, as shown in the micrographs, there is a good fusion between the consecutive tracks and layers leading to a very small amount of porosities. However, with B2-4-60, some big and irregular pores due to lack of fusion are visible. The lack of fusion porosity is evident from irregular shape and pores filled with unmelted powder particles.
Parameter set # |
C [wt.%] |
S [wt.%] |
O [wt.%] |
N [wt.%] |
H [wt.%] |
---|---|---|---|---|---|
B2-St-30 |
0,0525 |
0,0039 |
0,0501 |
0,0811 |
0,0006 |
B2-1-30 |
0,0444 |
0,0042 |
0,0448 |
0,0769 |
0,0004 |
B2-2-60 |
0,0523 |
0,0045 |
0,0714 |
0,0815 |
0,0004 |
B2-3-60 |
0,0457 |
0,0039 |
0,0557 |
0,0803 |
0,0004 |
B2-4-60 |
0,0446 |
0,0041 |
0,0431 |
0,0772 |
0,0004 |
Spec. per ASTM A 564/A 564M[20] |
Max. 0,07 |
Max. 0,03 |
- |
- |
- |
Spec. per SLM Solutions for powder [22] |
0,07 |
0,015 |
0,04 |
0,10 |
- |
One of the other remarkable features of the optical microscopy images given in Fig. 5 is the darker areas at the deep end of the melt pools close to the boundary. Although these dark areas can be confused with porosity in the etched versions, a closer look into these areas reveal that these regions have very dense cellular structures (see Fig. 6a-b). The cell size is very small being below 1 µm and the cells mostly have a homogenous size distribution as depicted in Fig. 6c. The SEM images shown in Fig. 6 demonstrate some white areas. In order to understand the source of these white areas, EDS analysis was performed on B2-2-60 and B2-St-30 specimens. The results show that these are mainly iron carbides as depicted in Fig. 7. This is consistent with the results of some other studies finding non-metallic inclusions in the SLM 17 − 4 PH SS [19, 20].
The results of the EBSD analysis of B2-St-30 and B2-2-60 are given in Fig. 8. The average fraction of ferrite is 39% and the one of austenite is 8,5% while the one of martensite is 52,5%. This is a different microstructure than wrought 17–4 PH steels which often have a fully martensitic microstructure. However, it is not uncommon to find some retained austenite in the microstructure due to the presence of N2 in both gas atomization and SLM processing as an austenite stabilizer. Moreover, the martensite finish temperature is slightly above the room temperature. Thus, very low variations in the material composition, even staying in the 17-4PH SS specification, can have a significant change on the microstructure regarding the phase composition. Inverse Pole Figures (IPF) given in Fig. 9 show that the specimens exhibits mixed grain structure having a columnar BCC grains parallel to the build direction with finer equiaxed FCC grains at melt pool boundaries. The IPF’s for both of the specimens reveal similar structures. However, the main difference is much more elongated grains along the build direction in B2-2-60 specimen, mainly due to doubled layer thickness and much higher laser power.
Schaeffler diagram can be used to understand the phase evolution in 17 − 4 PH stainless steel [23]. In this study, Creq is equal to 16,65 − 16,66 where as Nieq is calculated as 5,61 − 5,64. When these values are approximately marked on the diagram, the intersection point lies in the region of martensite + ferrite. However, there is some level of retained austenite 5–10% in the obtained microstructures. This may be due to the high cooling rates encountered in the SLM process not allowing fully austenite transformation.
The results of the tensile tests are presented in Fig. 10 with 95% confidence intervals from 3 repetitions. It is evident from the figures that the average ultimate tensile strength (UTS) and yield strength (YS) values are a bit higher, namely 4% and 15% respectively, when a layer thickness of 30 µm is used rather than 60 µm. Regarding elongation at break, it can also be concluded that the confidence intervals indicate a smaller value when a thinner layer thickness is preferred. An interesting observation of the tensile results is the high discrepancy between UTS and YS values. According to SLM Solutions Material Datasheet [22], the tensile strength for a layer thickness of 30 µm shall be 931 ± 45 MPa while the YS shall be 506 ± 25 MPa as minimum. While the UTS values obtained in this study significantly exceeds the given threshold significantly, YS values are far below the given limit. A representative stress-strain curve from B2-2-60 specimen is shown in Fig. 11. The tensile test results show that there is a large discrepancy between YS and UTS. The obtained YS is much lower than expected whereas the UTS is higher than most of the reported values in the literature. This conclusion is valid for all tested parameter sets and independent from the layer thickness. One of the reasons of having a low YS values may be attributed to the selected strain rate. The strain rate, which was taken as mm/min appropriately per tensile testing standard, has shown to have a significant effect on the yield stress by Wang et al. [28]. Moreover, it is also reported that the phase transformation can be observed in martensitic PH stainless steels during deformation. This can lead to an increased UTS as a result of the transformation induced plasticity (TRIP) effect. As shown in Fig. 11, the phenomenon of having “double yield” in the stress-strain curve is thus attributed to the formation of strain-induced martensite and further strengthening during deformation. The number of studies showing this effect is very limited for additive manufacturing [29]. Moreover, the highest UTS/YS ratios found in literature for this material after SLM are about 2 whereas it is equal to almost 4 for this study emphasizing the importance of double yield phenomenon. A post-heat treatment to be applied after the SLM process before aging is thus recommended to eliminate the metasable austenite and to increase the yield strength.
SEM images of the tensile fracture surface of various specimens in the as-built condition are shown in Fig. 12. The shape of the fracture dimples in terms of size, depth and quantity, relies on the ductility of materials as well as the the second phase particles [23]. The fracture surfaces of specimens having a layer thickness of 30 µm reveal a cup-and-cone fracture. At high magnifications, micro-void coalescence fracture, which is also known as dimpled rupture, is observed. It is evident from the figures that porosity and unmelted powder particles, especially observed in B2-4-60, can act as nucleation sites for cracks under tensile loading. Additionally, during loading in tensile testing, void nucleation can stem from cracking of particle–matrix interfaces for both secondary phases and unmelted particles. The co-existing deep holes, which are present in all samples can be attributed to molten liquid shrinkage or vaporization. The deep holes resemble holes observed with other materials in AM which reveal brittle rod-like intergranular fracture of dendrites [27].
This study aims at identifying a new process parameter set for an increased productivity by doubling the layer thickness while not affecting the residual stresses, material composition and mechanical properties. Starting from an initial set of parameters optimized for a layer thickness of 30 µm for 17 − 4 PH stainless steel, a wide range of various scan speed, laser power and scan spacing was tested for a layer thickness of 60 µm. Some of the tested parameters giving a high density above 99,5% and leading to a build rate increase of 2–3 times in comparison to initial set of parameters were tested in terms of hardness, material composition, tensile properties, residual stresses and microstructure.
The results within the tested ranges show that
The variation of the process parameters does not significantly alter the obtained microhardness results.
Using the bridge curvature method, changing the layer thickness and increasing the productivity does not significantly change the residual stresses in the part.
None of the tested parameter sets lead to non-conformance in terms of material composition although a very high laser power is utilized for thick layers.
A hierarchical microstructure is obtained with a cellular structure having a cell size below 1 µm.
Although wrought 17–4 PH steels which usually have a fully martensitic (BCC) microstructure, it is observed that 5–10% retained austenite is present in the microstructure while the ferritic and martensitic phases are found in fractions of 41% and 51%, respectively. This can be considered as independent from the utilized process parameters.
Tensile test results reveal a very high ultimate tensile strength and low yield strength in comparison to reported values for SLM 17 − 4 PH stainless steel with a specific “double yield” behaviour. This can be attributed to the strain induced hardening during tensile testing due to reversion of retained austenite to martensite. This is a very important outcome from this study as an important point to clarify the wide scatter of mechanical properties from this alloy processed by SLM. The phases and their fractions highly depend on the powder chemical composition (Creq/Nieq ratio) and processing gases used in atomization and SLM process. Thus, a solution annealing is recommended to manage the microstructure before aging for this material as a post-processing step after SLM.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Conflict of Interest / Competing Interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Availability of data and material
All data generated or analysed during this study are included in this published article.
Code availability
Code sharing not applicable to this article as no codes were generated during the current study.
Authors' contributions
All authors contributed to the study conception and design. Production, material preparation and characterization as well as the analysis of results were performed by Evren Yasa, İlker Atik and İpek Kandemir. The first draft of the manuscript was written Evren Yasa and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Ethics approval
This paper/study does not contain any studies with human participants or animals performed by any of the authors.
Consent to participate
Not applicable since this paper/study does not contain any studies with human participants performed by any of the authors.
Consent for publication
Not applicable since this paper/study does not contain any studies with human participants performed by any of the authors.