1 Introduction

The development of additive manufacturing technologies has allowed the industry to overcome constraints associated with conventional techniques [1]. As a single-step process, the primary advantages of AM include freedom of design, energy and material savings, and shortened design-to-manufacture time [2]. One of the most promising metal additive manufacturing processes is SLM, a laser powder-bed fusion process (LPBF) that uses a scanning laser beam to build up parts by selectively melting certain portions of the metal powder-bed, layer by layer, based on computer-aided design (CAD) data. SLM is increasingly popular because of its ability to process a wide variety of materials, including those that are expensive or difficult to produce, using traditional techniques such as steel, aluminum alloys, titanium alloys, and nickel alloys [3]. In addition, this technology provides the opportunity to enhance functionality and customize the microstructure and subsequent properties of the manufactured part [4].

Despite the numerous benefits of SLM, the industry still faces challenges in fully accepting it. The presence of defects such as porosity, lack of fusion, cracks, anisotropy, inhomogeneous microstructure, and high residual stress makes it a complicated process. The complex thermal history and rapid cooling rate are the main reasons behind these issues [5, 6].

15-5PH stainless steel is a martensitic precipitation hardening stainless steel with high corrosion resistance, good toughness, and good mechanical properties at temperatures up to 300 °C [7, 8]. This makes it suitable for various aerospace, chemical, petrochemical, and food processing applications [9]. Due to the low carbon content and the addition of copper as an alloying element, its body-centered cubic (BCC) martensite matrix can be strengthened through the formation of nanometric-sized Cu precipitates [10, 11]. The conventional heat treatment for this type of alloy consists of two steps: (1) solution treatment in the austenite phase field, followed by air or water quenching; (2) aging in a heat treatment furnace at 480–620 °C for a few minutes to several hours [12]. Aging starts with forming small and coherent Cu clusters with a BCC structure. With increasing aging temperature and/or aging time, the precipitate size and shape evolve until they transform into incoherent face-centered cubic (FCC) clusters [13, 14]. By selecting proper post-process parameters, the characteristics of this precipitation in terms of volume fraction and size can be tuned to obtain different application-oriented mechanical properties [15, 16]. This combination of characteristics and high-value applications makes the 15-5PH stainless steel an excellent candidate for additive manufacturing. Recently, SLM has shown promising results for fabricating precipitation hardening stainless steels [17, 18]. Several studies have focused on comparing the performance of SLM-processed 15-5PH and traditionally manufactured 15-5PH stainless steel. Coffy et al. [19] reported a smaller grain size with a layered microstructure and a significant volume fraction of retained austenite phase in SLM-processed 15-5PH compared to the conventional alloy. Their study focused on martensite/austenite phase fractions and did not address mechanical properties. Roberts et al. [20] compared the microstructure, microhardness, and high-temperature mechanical behavior of AM and wrought 15-5PH steel in the as-built condition. According to their preliminary results, additively manufactured 15-5PH showed better performance except for ductility. However, further testing and investigations are needed. Some researchers investigated the effect of processing parameters such as energy density, scanning strategy, building direction, and powder characteristics on the microstructure and mechanical properties of selective laser-melted 15-5PH to obtain high-quality final parts [21,22,23,24]. However, this approach cannot effectively control anisotropy and heterogeneous microstructure caused by the complex metallurgical interactions during the SLM process, especially for precipitation hardening grade. Recent studies focused on post-process heat treatment to address this challenge to achieve a better tradeoff between microstructure and mechanical properties for 15-5PH stainless steel [9]. Nong et al. [25] examined the effect of the standard precipitation hardening heat treatment H900 (i.e., solution heat treatment at 1040 °C for 30 min, followed by water quenching and subsequent aging at 482 °C for 1 h, followed by air-cooling) on the microstructural and mechanical characteristics of SLM-fabricated 15-5PH parts. They reported that the heat-treated SLM sample demonstrated greater strength and higher hardness than the aged wrought 15-5PH sample and comparable ductility. The finer grain structure, the high concentration of dislocations around grain boundaries, and the retained austenite explained this. Mechanical properties, microstructural evolution, and corrosion resistance were evaluated in several conditions (solution annealed, aged at different temperatures and times) [26,27,28,29]. Experimental results showed that different aging processes affect the Cu-rich precipitate characteristics and the amount of retained austenite, resulting in different mechanical behavior and corrosion resistance. These studies focused on standard heat treatments mainly designed for conventionally manufactured PH SS. A large dispersion of the mechanical property values is observed due to the high variability in phase composition and phase fractions of SLM-processed PH SSs. Thus, specific optimized heat treatment should be developed. Alafaghani et al. [9] investigated the influence of different solution annealing treatments on the microstructure and mechanical properties of SLM 15-5PH steel. They found that extending the solution treatment time improved microstructure isotropy and tensile strength at the expense of ductility, while increasing the solution treatment temperature adversely affected SLM parts. However, simple correlations drawn from experimental results without systematic analysis would not be effective for heat treatment optimization. Moreover, other parameters such as cooling rate, aging temperature, and aging time, outside the standard range were not yet investigated.

Therefore, it is necessary to conduct a comprehensive study on the effect of heat treatment parameters on the microstructure of SLM 15-5PH. The wide application of 15-5PH in different industries highlights the need for more research. To obtain the desired mechanical properties, it is essential to systematically identify optimal heat treatment parameters using efficient experimental design and analysis methods. The Taguchi method is a statistical experiment that optimizes designs for performance, quality, and cost [30].

Therefore, the main contribution of this research is to provide detailed insights into the heat treatment of 15-5PH fabricated through SLM and improve its microstructural and mechanical properties. This study aims to investigate the influence of precipitation hardening heat treatment parameters on the microstructure and microhardness of 15-5PH parts processed by SLM based on Taguchi's design. A statistical approach based on ANOVA, RSM, and linear regression has been utilized to identify the significant heat treatment parameters and predict their effects on 15-5PH microhardness. Additionally, microstructure characterization has been performed and correlated to the obtained hardness.

The findings for this work are expected to develop the application range of SLM 15-5PH in different industries where the properties enhancement of material is necessary.

2 Material and methods

2.1 Design of experiment

The Taguchi method is a simple and powerful optimization tool to evaluate the significance of different process parameters for a target response. It allows reducing the number and cost of experimental tests without affecting accuracy, thanks to the orthogonal array design [31]. The procedure of Taguchi design includes the following steps: (1) identification of the quality characteristics and selection of the design parameters; (2) selection of the appropriate orthogonal array according to the number of parameters and number of levels; (3) conduction of the experiments and statistical analysis of the results; and (4) identification of the optimal levels of design parameters [30].

The microhardness of precipitation hardening stainless steel relies on a complex microstructure developed during a quenching sequence after austenitization, followed by aging heat treatment [10]. Aging temperature and aging time play a prominent role in controlling the Cu precipitate characteristics and, thus, the alloy’s strength. The homogenization step and the subsequent quench are also crucial in the control of austenite/martensite proportions. Accordingly, quenching solution (A), aging temperature (B), and aging time (C) are selected as design parameters for the Taguchi plan. Three levels for each parameter are set, as shown in Table 1.

Table 1 Heat treatment parameters and their levels
Full size table

For conventional 15-5PH, a solution heat treatment is typically applied in the austenitic domain (∼ 1050 °C), followed by quenching, to obtain a fully martensitic structure. A subsequent aging treatment in the range of 480–620 °C is conducted for 1–4 h, which results in precipitation hardening by copper precipitates. Precipitation begins with small and coherent spherical Cu clusters with a BCC structure, which transform into an elliptical incoherent FCC structure, called ε-Cu, upon over-aging [13].

Aging parameters were chosen to cover the typical under-aged, peak-aged, and over-aged conditions known in standard precipitation hardening treatments while exploring a lower range of temperature and time to reduce post-processing cost and time. For quenching, three different aqueous solutions were used in the experiments: distilled water (S1), NaCl solution at 6% of mass concentration (S2), and 12 mass% NaCl solution (S3), allowing three different quenching rates as reported in [32]. Based on the Taguchi method, the appropriate experimental design for this study is an L9 orthogonal array. This array has eight degrees of freedom and, thus, is suitable for three-level design parameters. Therefore, nine unique heat treatment combinations are available, as shown in Table 2. Overall, parameters’ levels are determined to balance the need for precision and the practical limitations of the experiment, taking into account prior knowledge and experience.

Table 2 Experimental layout using L9 orthogonal array
Full size table

2.2 Specimen fabrication and preparation

The samples were produced on the EOS M 290 system using the commercialized PowderRange 155 stainless steel (15-5PH) powder from Carpenter Additive. The particle size distribution ranges from 15 to 45 µm. Table 3 lists the chemical composition of the 15-5PH powder used.

Table 3 Chemical composition of 15-5PH stainless steel powder (wt%)
Full size table

The SLM machine has a Yb-fiber laser with a beam diameter of 0.1 mm. To avoid oxidation, printing was done in an argon atmosphere. Specimens were produced using optimized processing parameters recommended by the manufacturer, which are listed in Table 4. Figure 1 depicts the dimensions and building orientation of the specimen. Laser scan paths were rotated at a 45° angle between two adjacent layers. It is worth noting that both the substrate and tensile samples were made of 15-5PH. After the builds were completed, tensile samples were separated from the substrate using a wire EDM machine and then sandblasted to homogenize surfaces.

Table 4 SLM process parameters for 15-5PH stainless steel
Full size table
Fig. 1
figure 1

Dimensions (in mm) and building direction of samples

Full size image

For comparison, we kept two samples in their as-built condition. In the first step, the remaining samples were solution-treated at 1020 °C for 15 min. Each group of three is then cooled in a different quenching solution (1, 2, and 3). After quenching, samples are individually aged in the furnace at different temperatures and holding durations according to the orthogonal array (see Table 2) and cooled in air, giving rise to precipitation hardening. Two other validation samples were later solution-treated at 1020 °C for 15 min, water-quenched, and aged in a furnace according to the optimal set of parameters.

2.3 Microstructure, microhardness, and tensile tests

To evaluate the microhardness and microstructure of SLM-fabricated 15-5PH steel samples at their cross-section parallel to the building direction, samples were mounted in epoxy and polished with diamond suspensions with average diamond particle diameters of 9 μm, 6 μm, 3 μm, and 1 μm. The samples were then chemically etched with Fry’s reagent (5 g CuCl2 + 40 mL HCl + 30 mL water + 25 mL ethanol) to reveal grain structure. Microstructure characterization was performed using LEXT OLS4100 laser confocal microscope, Hitachi TM3000 scanning electron microscope (SEM), and energy-dispersive spectroscopy (EDS). Vickers microhardness measurements were taken with a testing load of 300 gf and 10 s of dwell time using a Clemex machine, conforming to ASTM E384 standard. Values are automatically converted to Rockwell C by Clemex software based on ISO 18265 standard. For each sample, the hardness value is determined by averaging 20 indentations spaced at 200 µm and performed along the surface’s diagonal.

After collecting experimental data, ANOVA and RSM were performed using Minitab 19 statistical analysis software to determine significant heat treatment parameters and identify optimal parameters combination, respectively.

Room-temperature tensile tests were carried out on rectangular cross-section specimens (6 × 3 mm) with a 25mm gauge length using an MTS-810 tensile testing machine according to ASTM-E8 standards. An extensometer measured the elongation of the tensile specimens during testing. Two samples for each of the as-built and optimal heat-treated conditions are tested.

3 Results and discussion

3.1 Microstructure analysis

The complex thermal cycle in the SLM process and subsequent heat treatments induce a complex microstructural evolution in 15-5PH stainless steel that, in turn, affects its mechanical properties. To investigate the structure–property relationship, the as-built sample and aged samples S1-600 °C-180 min, S2-400 °C-120 min, and S3-500 °C-60 min from the Taguchi array were subjected to microstructure examination. The specimens are selected according to the microhardness results reported in Table 5 as they have the lowest, intermediate, and highest hardness, respectively. The as-built sample in Fig. 2a reveals a typical layered microstructure with well-overlapped molten pools formed by the laser beam. An alternating scan strategy, in which the scanning direction is rotated by 45° between two successive layers, is evident in scan track shapes. The high-magnification SEM image of the as-built sample in Fig. 2b displays equiaxed and columnar sub-grain structures grown epitaxially. The boundaries of the melt pools are apparent by dark lines, and refined equiaxed grains are formed around melt pool boundaries.

Table 5 Microhardness test results
Full size table
Fig. 2
figure 2

Microstructure features of as-built 15-5PH stainless steel: a optical microscopy and b SEM. The black arrows indicate epitaxial growth (color figure online)

Full size image

In contrast, columnar grains arose parallel to the building direction toward the center of the melt pool. As shown in OM images in Fig. 3, scan track boundaries are eliminated after heat treatment. The heat-treated 15-5PH stainless steel samples represent a more homogenized microstructure. Some defects and porosities can be seen in Figs. 2 and 3. The presence of these flaws is commonly attributed to the inadequate melting of the powder and/or the entrapment of gas bubbles [33]. After heat treatment, defects seem to be reduced but not eliminated compared to the as-built specimen. Examining SEM micrographs in Fig. 4 reveals a large proportion of martensite and ferrite phases with retained austenite. The effect of heat treatment on the microstructure of other materials fabricated by SLM has been also reported in previous studies [34, 35].

Fig. 3
figure 3

Optical micrographs of SLM 15-5PH samples after different heat treatments: a S1-600 °C-180 min, b S2-400 °C-120 min, and c S3-500 °C-60 min. The red circles indicate porosities (color figure online)

Full size image
Fig. 4
figure 4

SEM micrographs of SLM 15-5PH samples after different heat treatments: a S1-600 °C-180 min, b S2-400 °C-120 min, and c S3-500 °C-60 min. Black arrows indicate Cu precipitates (color figure online)

Full size image

Although conventional 15-5PH is a martensitic alloy, ferrite and retained austenite in the SLM-processed 15-5PH have been widely reported in the literature [36, 37]. The high cooling rate of the SLM process leads to fine grain size with large strain at the grain boundaries; hence, the austenite retention becomes favorable [38]. The presence of nitrogen in the initial powder may also promote austenite formation. The build chamber temperature during the SLM process may also be greater than the martensite finish temperature, leading to incomplete martensitic transformation [19]. Nong et al. [39] declared that the matrix of the as-built 15-5PH is BCC (martensite and ferrite) with an approximate 10.8% volume fraction of FCC (austenite), which permitted more plastic strain to be accommodated before crack initiation and slowed crack propagation but softened the material.

Since the SLM process has several common points with welding, basically the fast cooling rate, the phase composition of SLM-fabricated 15-5PH stainless steel can be effectively predicted using the Schaeffler diagram, commonly used for welded stainless steel [40]. The effect of austenite and martensite/ferrite stabilizing elements on the microstructure is quantified using the following equations [36]:

$${\text{Ni}}_{{{\text{eq}}}} \left( {{\text{wt}}\% } \right) = \% {\text{Ni}} + 30 \times \% {\text{C}} + 30 \times \% {\text{N}} + 0.5 \times \% {\text{Mn}}$$
(1)
$${\text{Cr}}_{{{\text{eq}}}} \left( {{\text{wt}}\% } \right) = \% {\text{Cr}} + \% {\text{Mo}} + 1.5 \times \% {\text{Si}} + 0.5 \times \% {\text{Nb}}$$
(2)

Based on the chemical composition in Table 3, the calculated \({\text{Ni}}_{\text{eq}}\) an \({\text{Cr}}_{\text{eq}}\) are 15.85% and 8.85%, respectively. The point of intersection is used to predict phase constituents of the 15-5PH microstructure as illustrated in Fig. 5. This point, marked with an orange dot, is worth noting that both the substrate and tensile samples were made of 15-5PH from the substrate using a wire EDM machine, without subsequent heat treatment [7].

Fig. 5
figure 5

Schaeffler diagram [7]. The orange dot represents the phase composition of our as-built sample

Full size image

After solution treatment of the SLM-fabricated sample, the microstructure transformed into a mostly martensitic matrix with a small residual austenite [41]. During the subsequent quench, martensite laths grow inside the former austenite grains [10]. Upon aging heat treatment, the second-phase elements, such as copper, are regrouped into clusters by the diffusion process and constitute precipitates.

Figure 4a shows the microstructure of Sample S1-600 °C-80 min, which was aged at 600 °C for 180 min following water quenching. It indicates a heavily tempered martensitic structure and abundant Cu-rich particles with different sizes around the grain boundaries. Figure 4b depicts the microstructure of Sample S2-400 °C-120 min, which was aged at 400 °C for 120 min after being solution-treated and quenched in 6 mass% NaCl solution. The microstructure reveals a higher percentage of austenite than the as-quenched sample (not presented here). This may be explained by the formation of reverted austenite during the aging treatment, as reported by Sarkar et al. [38]. It is also noted that the volume fraction of Cu precipitates is less than that of Sample S1-600 °C-180 min due to the lower aging temperature and aging time. Figure 6 displays the EDS mapping analysis of the second-phase precipitates in Sample S2-400 °C-120 min. It confirms that these were Cu-enriched particles with varying sizes (> 100 nm). Pasebeni reported similar large Cu precipitates non-uniformly distributed in gas-atomized 17-4PH powder produced by SLM following solutionizing and aging at 482 °C. Figure 4c shows the microstructure of Sample S3-500 °C-60 min heat-treated at 500 °C for 60 min after quenching in 12 mass% NaCl solution. It reveals less retained austenite and finer Cu precipitates evenly distributed compared to the two previous samples, which was consistent with the precipitates in traditional 15-5PH stainless steel after similar aging conditions [42]. Some studies have also reported the formation of carbides and nanometric oxide inclusions during the aging of conventional [10, 43] and additively manufactured 15-5PH stainless steel [7, 44] that acted as strengthening factors.

Fig. 6
figure 6

EDS mapping of Cu precipitates in Sample S2-400 °C-120 min

Full size image

3.2 Microhardness results

Microhardness is a significant mechanical property for metallic components, especially when used in applications that include friction and contact. Microhardness measurements resulting from the various heat treatment combinations are listed in Table 5.

The hardness values range between 36.3 HRC and 46.5 HRC with 39.5 HRC for the as-built specimen. Considering the orthogonal array combinations, Sample S3-500 °C-60 min exhibited the highest hardness, whereas Sample S1-600 °C-180 min exhibited the lowest hardness. The as-quenched samples had the lowest values among all the samples. These findings show good agreement with the literature [7, 38, 45], since our test Sample S3-500 °C-60 min is comparable to the H900 standard aging treatment, which demonstrated the maximum strength and hardness. After being reduced by solution treatment, aging at 400 °C brings back hardness to the as-built value. Further, a rise in temperature to 500 °C induces an increase in hardness by ~ 15% compared to the as-built sample and ~ 24% compared to the as-quenched samples. In contrast, after aging at 600 °C, the hardness values decline again. This variation is correlated to the complex microstructural evolution. The presence of retained and reverted austenite softens the material, while fine Cu-rich precipitates harden it. Primig et al. [12] proved that a significant evolution of 15-5PH chemical composition occurs when aged between 450 and 530 °C, leading to a pronounced hardness peak. Over-aging occurs when the temperature is raised to 600 °C due to the coarsening of copper precipitates [12], and the amount of reverted austenite increases as the heating temperature and duration increase. Consequently, hardness decreases.

3.3 Analysis of variance

ANOVA is a statistical tool that uses the F test to investigate which design parameters significantly affect the target response. In this study, ANOVA was performed to assess the significance of each factor in the quench and aging heat treatment process, namely, quenching solution (A), aging temperature (B), and aging time (C), on the hardness of 15-5PH stainless steel parts processed by SLM. The analysis presented in Table 6 is ensured using Minitab 19 statistical analysis software according to the general stepwise method so that only significant factors are kept in the final model. The initial model included independent variables (A, B, C), quadratic variables (A2, B2, C2), and two-way interactions (A × B, A × C, B × C). Statistically significant parameters are determined using the p value. Comparing the p value with the fixed significance level (α) determines whether the null hypothesis can be rejected. In this case, the null hypothesis states that the effect of studied parameters is not significant on the hardness of 15-5PH stainless steel produced by SLM. We reject the null hypothesis when the computed p value is less than the designated significance level. In this study, we used a significance level of α = 0.05. The percentage contribution is also a rough but effective guide to the relative importance of each model term [46]. An empirical model for hardness as a function of selected variables was developed by applying linear regression analysis to the experimental data. The general quadratic equation model is stated by [46]:

$$y = \beta_{0} + \mathop \sum \limits_{i = 1}^{3} \beta_{i} x_{i} + \mathop \sum \limits_{i = 1}^{3} \beta_{ii} x_{ii}^{2} + \mathop \sum \limits_{i = 1}^{2} \mathop \sum \limits_{j > 1}^{3} \beta_{ij} x_{i} x_{j} + \varepsilon$$
(3)

where \(y\) represents the hardness, \({x}_{i}\) represent the heat treatment parameters, the β’s are regression coefficients, and \(\epsilon\) is the residual error term. ANOVA results (Table 6) show that hardness is mainly influenced by B2, B, and C variables. It can be observed that aging temperature (B) is the most influential parameter, accounting for 90.1% of the total variability for the quadratic term B2 and 5.82% for the linear term B. Aging time (C) is less significant, with a contribution of 2.75%. The quenching solution (A) and the remaining quadratic and interaction terms were omitted from the final model as their p values were much higher than the significance level of α = 0.05, meaning that they do not significantly affect the hardness.

Table 6 ANOVA for hardness
Full size table

To have a better visual perspective of the impact of each heat treatment parameter on the hardness of SLM-fabricated parts, the main effect plots are shown in Fig. 7; aging time hurts hardness. The hardness decrease with prolonged aging time may be ascribed to the coarsening of Cu precipitates and the formation of reverted austenite [47, 48]. The steep variation of mean hardness concerning aging temperature confirms that this factor is highly significant. It has shown a positive effect between 400 and 500 °C and a negative effect between 500 and 600 °C. This variation is in agreement with the chemical and microstructural evolution of 15-5PH stainless steel processed by SLM, which is explained in the microstructure analysis section. The peak hardness was achieved at 500 °C aging temperature due to the uniformly dispersed precipitation of fine Cu precipitates. Lower aging temperature (400 °C) resulted in under-aging characterized by a small volume fraction of Cu precipitates. In comparison, higher aging temperatures led to over-aging, featured by the coarsening of Cu-rich particles and the formation of an extensively reverted austenite phase. Both conditions result in a lower hardness compared to the peak value.

Fig. 7
figure 7

Main effect plots

Full size image

To estimate the hardness of heat-treated SLM parts as a function of the significant factors, a multiple linear regression model was developed using Minitab. The obtained model for predicting hardness is a quadratic model given by Eq. 4.

$$H = - 129.2 - 0.01222C + 0.716B - 0.000727 B^{2}$$
(4)

The coefficient of determination R2 that measures the goodness-of-fit of the regression model is 98.6%. In other words, the regression model explains 98.6% of the total variation in hardness. The adjusted R2 (97.88%) and the predicted R2 (95.72%) are in reasonable agreement, confirming that the predicted model for hardness can be employed. The standard deviation S of the data values around the fitted values is 0,55, meaning that our model describes the hardness response well.

3.4 Response surface model

The ANOVA results led us to conclude that aging temperature and aging time are the most influential factors in the heat treatment process that affect the hardness of SLM-fabricated 15-5PH components. Further investigation into this relationship allows for estimating the optimal set of parameters that provide the highest hardness. For this purpose, the response surface method (RSM), a statistical technique for modeling and analysis of such optimization problems, is used. The optimum can be localized with reasonable precision by generating contour plots for response surface analysis. Figures 8 and 9 show the 2D contour plot and the 3D response surface for hardness versus aging time and aging temperature. The variation of hardness as a function of the quenching solution is not considered in the plots since ANOVA results show it to be insignificant. Thus, it is taken as a constant parameter at level 1 since it showed slightly higher hardness according to the main effect plot. It can be inferred from these plots that hardness increases with decreasing holding time. The quadratic relationship between aging temperature and hardness is visible through the parabolic curves of the contour plot. The hardness reaches its maximum when the precipitation hardening temperature ranges between 470 and 515 °C approximately and the aging time ranges between 60 and 80 min. Thus, we can consider 490 °C and 70 min as the best set of parameters for precipitation hardening heat treatment.

Fig. 8
figure 8

Two-dimensional contour plot of hardness versus aging temperature and aging time. The black dot represents the optimum point (color figure online)

Full size image
Fig. 9
figure 9

Three-dimensional response surface plot of hardness versus aging temperature and aging time

Full size image

3.5 Tensile properties

Both hardness and tensile strength are essential properties of metallic components that are linearly correlated [49]. Achieving maximal hardness implies having the highest tensile strength as well. Therefore, to validate the fitted model and assess the mechanical properties of heat-treated 15-5PH stainless steel produced by SLM, tensile tests were conducted on peak-aged and as-built specimens using MTS-810 tensile testing machine. Precipitation hardening was performed at 490 °C for 70 min after solution treatment and water quenching to obtain the maximum hardness determined by RSM analysis. Two samples for each testing condition are tested, and the mean values of their tensile properties are summarized in Table 7. Figure 10 illustrates the stress–strain curves. The results of tensile tests indicate that precipitation hardening heat treatment has enhanced the mechanical properties of SLM-fabricated 15-5PH stainless steel. The yield strength of the heat-treated specimen has increased by approximately 32%, ultimate tensile strength by 26%, and ductility by 89% compared to the as-built specimen. The selected set of parameters enabled the optimal combination of strength, hardness, and ductility to be achieved. This outstanding combination of strength and ductility can be ascribed to (1) grain refinement, (2) high dislocation density, (3) high volume fraction of fine Cu-rich precipitates, and (4) transformation-induced plasticity during tensile deformation due to the presence of retained austenite [44, 47, 50].

Table 7 Average tensile properties of 15-5PH stainless steel under as-built and optimal heat-treated conditions
Full size table
Fig. 10
figure 10

Stress–strain curves of as-built and peak-aged 15-5PH stainless steel

Full size image

4 Conclusion

This work investigates the effect of heat treatment parameters on the hardness and microstructure of selective laser-melted 15-5PH stainless steel. The observation of 15-5PH SS microstructure before and after heat treatment revealed the presence of ferrite, retained/reverted austenite, and copper precipitates. The experimental approach using Taguchi design and analysis of variance (ANOVA) allowed a deep understanding of the influence of quenching solution, aging temperature, and aging time on the hardness of samples. The findings reveal that hardness is mainly affected by aging temperature and less by aging time. Predictive regression and RSM models for hardness were developed using the aforementioned heat treatment parameters as independent variables. The optimal set of parameters was then selected and validated by tensile tests. The validation samples were solution-treated at 1020 °C for 15 min, water-quenched, and aged at 490 °C for 70 min. The resulting mechanical properties are very promising, with higher hardness (up to 46 HRC), higher mechanical strength (ultimate tensile strength up to 1692 MPa), and improved ductility (elongation at break up to 8%) compared to the as-built samples. This can be attributed to the high volume fraction of fine Cu-rich precipitates and transformation-induced plasticity during tensile deformation due to retained austenite. The results of the present study can be used to tailor the mechanical properties of the SLM-fabricated 15-5PH stainless steel by tuning post-process heat treatment parameters. In future works, validation hardness tests and microstructure analysis of the optimal sample obtained from RSM should be given priority. Performing additional tests outside the variation range of parameters presented in this study is needed to improve and assess the accuracy of the predictive model. More in-depth analyses, using more advanced techniques such as high-resolution transmission electron microscopy or atom probe tomography, are also necessary to properly characterize the unique microstructure of SLM-processed 15-5PH stainless steel and investigate the process-structure–property relationship deeply.