3.1. Geometric characterization
The geometrical characteristics of the straight walls were influenced by the ED and the PDD, as shown in Fig. 4. The figure displays the variation of each geometrical variable with respect to ED and PDD on multi-axis graphs. The data points were fitted with a surface using Matlab software.
The central height of the straight walls increased with both ED and PDD, as shown in Fig. 4a and Fig. 4b. The increase in central height was due to higher ED providing more energy to melt the powder particles, and higher PDD delivering more powder particles to the melt pool region, resulting in more material being deposited. As a result, more material was deposited and the central height of the walls increased. The central height of the straight walls has experienced a significant increase at ED of 54 J/mm2 and a PDD of 0.06 g/mm2, the height was measured at 6.4 mm. However, with an increased ED of 71 J/mm2 (1.3 times higher) and a higher PDD of 0.1 g/mm2 (1.7 times higher), the height doubled to 12 mm. Similarly, when the ED rose by 1.5 times to 86 J/mm2 and the PDD increased by 1.2 times to 0.11 g/mm2, the wall height grew from 8.5 mm to 13.2 mm. In this case, it can be observed that the increase in ED was more pronounced than the increase in PDD. Consequently, the growth rate of the wall height was reduced from 2 times to 1.3 times. This analysis demonstrates that the effect of PDD on wall height is more dominant compared to ED. While the increase in ED has progressed from 1.3 times to 1.5 times, the rise in PDD has decreased from 1.7 times to 1.3 times. Consequently, the overall impact on wall height has diminished from doubling to 1.3 times. In summary, these findings highlight the significant influence of PDD in determining wall height, surpassing the effects of ED. However, there was a limit to how much PDD could increase the height, as too high PDD could lead to unmelted powder particles and lower material utilization. This could affect the buy-to-fly ratio, which is important for aerospace applications [46]. The central height ranges from 6.56 mm to 11.96 mm. Figure 4a shows also some jumps in the height at 71.42 J/mm2 as PDD increased. Similar jumps were observed at other ED levels as well. Figure 4b shows that the increase in height was more pronounced at higher ED levels when PDD increased.
The central width of the straight walls also increased with both ED and PDD, as shown in Fig. 4c and Fig. 4d. This was because higher ED increased the energy per unit area and thus enlarged the melt pool width [47] which determined the width of the deposit. Higher PDD also contributed to a wider deposit by delivering more material to the melt pool region. The maximum width, as depicted in Fig. 4d, was observed at the combination of the highest values of ED and PDD. Specifically, the central width of the straight wall increased from 1 mm at 35 J/mm2 ED and 0.1 g/mm2 PDD to 3 mm at 100 J/mm2 ED and 0.3 g/mm2 PDD. The increase in central width was more pronounced at higher ED levels when PDD increased. The laser spot diameter might also influence the width of the single-track multilayer deposit. The average width ranged from 3.97 mm to 5.00 mm.
Figure 4e and Fig. 4f show the variation of the melt pool penetration depth with ED and PDD. The melt pool depth ranged from 77.53 µm to 307.5 µm. Higher ED increased the substrate temperature and thus increased the melt pool penetration depth below the substrate. Higher PDD reduced the ED incident on the substrate by increasing the laser-powder interaction in the air. This resulted in a lower melt pool penetration depth below the substrate. The minimum melt pool penetration depth was observed at low ED and high PDD, as shown in Fig. 4f. The maximum melt pool penetration depth was observed at high ED and low PDD, as shown in Fig. 4f. The maximum depth of the straight wall exhibited a decrease from 0.3 mm to 0.2 mm when comparing different combinations of ED and PDD values. Specifically, at an ED of 54 J/mm2 and a PDD of 0.008 g/mm2, the maximum depth was measured at 0.3 mm. However, with an increased ED of 61 J/mm2 and a higher PDD of 0.1 g/mm2, the maximum depth decreased to 0.2 mm. This trend suggests that as the PDD parameter increased beyond certain thresholds, there was a notable decrease in the maximum depth of the straight wall. In other words, higher PDD values had a diminishing effect on the ED. These finding highlight the inverse relationship between PDD and the ED. Increasing PDD beyond certain thresholds led to a decrease in the maximum depth, indicating that the impact of PDD on the depth is significant.
3.2. Porosity Analysis
Porosity is a common defect observed in additively manufactured parts, particularly in powder-fed processes. The presence of porosities in the deposited material can be attributed to unmelted powder particles or entrapped gas within the melt pool region. Porosity originating from the feedstock powder can be caused by the presence of satellites and irregularly shaped powder particles. Unmelted powder particles result in lack of fusion (LOF) defects, occurring when there is insufficient energy to fully melt the powder particles [48, 49]. LOF porosity is typically found at the interfacial boundaries of each track or layer [2]. These porosities exhibit sharp edges and irregular shapes [50], leading to the development of high local stresses [2]. Proper adjustment of energy density (ED) and overlap percentage can help minimize LOF porosity.
In the case of the Laser Metal Deposition (LMD) process, a carrier gas, commonly Argon, is utilized to deliver the powder to the melt region. Additionally, a shielding gas, also typically Argon, is employed to protect the melt pool region from oxidation. During the process, gas bubbles can become trapped within the melt, resulting in gas porosity. Gas porosities are spherical in shape, occur within layers, and do not form at fixed locations [2]. Porosity originating from the feedstock is reported as a potential source of porosity within the layers during the LMD process [51, 52]. The presence of porosity is undesirable as it reduces material properties [53]. Pores can lead to the formation of small cracks, diminishing the material's ability to conduct heat effectively [54], while also decreasing the fatigue life due to acting as stress concentration sites [55].
Figures 5a and 5b illustrate the fluctuation of cross-sectional porosity levels (%) in relation to ED and Powder Deposition Density (PDD). Figure 5a demonstrates that cross-sectional porosity is influenced by both ED and PDD. Notably, Fig. 5b reveals that the highest cross-sectional porosity occurs at high ED and low PDD. This behavior can be attributed to the presence of convective fluid flow within the melt pool, known as the Marangoni flow. The Marangoni flow arises from surface tension gradients, driving surface flows from regions of low surface tension to those of high surface tension under the thermocapillary force [56]. The dynamics of the flow in the melt pool region intensify with increasing ED. As a result, the Marangoni flow tends to retain more entrapped gas bubbles. The flow drags these gas bubbles towards the bottom of the melt pool, where they become trapped upon solidification, leading to increased gas porosity. Additionally, the evaporation of material at high ED also contributes to pore formation [57]. The rotational flow facilitates the collision and coalescence of gas bubbles, resulting in the formation of larger pores [58]. It is worth noting that gas porosity can also be observed at low ED, where the high solidification rates entrap gas bubbles before they can escape, thus causing porosity [48, 59].
Figures 5c and 5d illustrate the variations in cross-sectional maximum pore diameter in relation to ED and PDD. Figure 5c shows that the lowest maximum pore diameter, measuring 15.61 µm, occurs at low ED/PDD values (48.98 J/mm², 0.0087 gram/mm²). Conversely, the highest maximum pore diameter is observed at high ED and PDD (85.71 J/mm², 0.0035 gram/mm²). The pore diameter ranges from 15.61 µm to 54.19 µm in all deposited samples. The spherical gas porosities can be visualized in Fig. 6 and Fig. 7a in the subsequent sections.
Figures 5e and 5f depict the changes in cross-sectional maximum lack of fusion (LOF) gap length in relation to ED and PDD. The maximum LOF gap length is observed at low ED and PDD values (53.57 J/mm², 0.0020 gram/mm²) as shown in Fig. 5f. Figure 5f further illustrates that, at a constant PDD, increasing ED leads to a decrease in the LOF gap length. The lowest LOF gap length is observed at the highest ED and PDD values (85.71 J/mm², 0.0035 gram/mm²). The LOF gap lengths range from 28.27 µm to 174.34 µm across all deposited samples.
3.3. Microstructural characterization
In metal additive manufacturing processes, solidification commonly occurs at a high rate and occurs in the direction of heat transfer, with molten metal crystals growing in the same direction. Conduction and convection are the primary modes of heat transfer in the additive manufacturing process. Conduction, which occurs within the melt pool during deposition, is the most influential mode of heat transfer in LMD. Heat dissipates from the hot region to the cold region during solidification.
The macro-structural morphology of the laser-deposited PH 13 − 8 Mo is determined by grains with columnar, cellular, and fine dendritic solidification patterns, as depicted in Fig. 6. At low energy density and high powder deposition density, the grains were fine and equiaxed, indicating a rapid solidification and nucleation process. At high energy density and low powder deposition density, the grains were coarse and columnar, indicating a slower solidification and growth process. Following inoculation using Ni-based heterogeneous nuclei, the formation of columnar grains is effectively hindered, thus triggering a transition from columnar to equiaxed microstructure as shown in Fig. 6 region C. In this transformed microstructure, equiaxed crystals prevail, and their size undergoes a gradual coarsening along the solidification direction. Equiaxed crystals exhibit a greater degree of diversity compared to columnar grains and lack discernible directional alignment at the central region. According to the Hall-Petch theory, the presence of a refined equiaxed grain structure exerts a pronounced influence on enhancing the yield strength of the alloy. Moreover, the abundance of grain boundaries in this microstructure facilitates material deformation processes [60].
Columnar grains spanning multiple layers are started to form due to epitaxial grain growth at bottom as shown Fig. 6A and B regions. The solidification process begins at the solid-liquid interface, such as the interface between the melt pool and the substrate or between the melt pool and the previously deposited layers [61]. Crystals develop at these interfaces, acting as nucleation sites. As solidification progresses, the crystals nucleated on the substrate or previously deposited layers grow in size. In the initial stage of solidification, competitive growth of primary δ-ferrite grains occurs for the first layer. The growth direction of these grains < 100 > matches the heat dissipation direction, with grains aligned in this direction having a higher chance of growth compared to those with less favorable crystallographic orientation. With subsequent layers, the crystals follow the crystallographic orientation of the previous layer, resulting in epitaxial growth. This growth mechanism leads to a reduction in the number of grains and an increase in their size, resulting in coarse columnar grains [62]. The microstructure of the PH 13 − 8 Mo at bottom where near the substrate consists of columnar dendrites growing epitaxially from the substrate and primary austenite cells containing intercellular ferrite. Additionally, martensite laths nearly parallel to the retained austenite can be observed. The dominant microstructure of additively manufactured PH 13 − 8 Mo at room temperature is characterized by δ-ferrite embedded in a low carbon martensite matrix.
The phase transformation sequence for martensitic PH stainless steel is as follows:
liquid (L) → L + δ-ferrite → δ-ferrite → δ-ferrite + γ → δ-ferrite + γ + M → δ-ferrite + M (4)
Two solid-state transformations are involved in the cooling process of the martensitic PH SS. The cooling process of martensitic PH stainless steel involves two solid-state transformations: shear transformation (austenite to martensite) [63] and diffusion-controlled transformation (δ-ferrite to austenite) [64]. The fast cooling rate and on-equilibrium nature of additive manufacturing restrain the transformation of δ-ferrite to austenite. Consequently, a considerable amount of untransformed δ-ferrite phase remains in the martensitic matrix at room temperature [65]. Consequently, a considerable amount of untransformed δ-ferrite phase remains in the martensitic matrix at room temperature [20].
High cooling rates can result in a larger volume fraction of δ-ferrite at room temperature because of the restricted diffusional transformation of δ-ferrite to austenite. For low cooling rates, there is sufficient time for the diffusional transformation of δ-ferrite to austenite, decreasing the amount of δ-ferrite at room temperature [63].
Figure 7 shows selected optical microscope images of the microstructure of PH 13 − 8 Mo stainless steel specimens exposed to the lowest and highest energy densities (ED) during the LMD process. These images reveal the differences in microstructure at different ED levels.
NiAl precipitates, which can be up to 7 nm in diameter, may exist in the as-built conditions but are not detectable through light microscopy. This can be because the precipitates are too small and in low volumetric amounts or dissolved in austenite during the process and did not have sufficient time to precipitate because of the high cooling rates inherent to LMD. NiAl precipitates can be up to 7 nm in diameter [23]. It should be noted that the formation of copper precipitates along the grain boundaries during the deposition of as-built 17 − 4 PH stainless steel via LMD has been reported [66]. Additionally, beyond the dendritic solidified region, the laser-deposited structure exhibits a fine cellular grain structure, as shown in Fig. 8. This transformation is attributed to the temperature gradient and solidification rate values. A higher temperature gradient favors the formation of cellular grains, while higher solidification rates promote dendritic morphology.
During the LMD process, heat dissipates primarily through the substrate or previously deposited layers, with some heat dissipating through the neighboring solidified layer. The cooling rate and solidification velocity at the bottom layer are very high, resulting in a progressive decrease in the cooling rate as the height of the deposit increases. This decrease in cooling rate leads to an increase in the secondary dendritic arm spacing (SDAS). Fig. 9 in section shows the relationship between the secondary dendrite arm spacing (SDAS) and the energy density for the first pass and last pass of the additively manufactured parts. The graph shows that the SDAS increases with increasing energy density, indicating a coarser and more columnar grain structure at higher energy densities. This is consistent with the optical microscopy analysis presented in section 3.3, which showed that the grain size and morphology varied with the energy density and powder deposition density
The graph also shows that the SDAS is higher for the last pass than for the first pass, indicating a more columnar and directional growth of the grains in the last pass. This is because the last pass experiences the high thermal The relationship between the SDAS and the energy density is important for understanding the microstructure and properties of the additively manufactured parts. The SDAS can affect the anisotropy, texture, and mechanical properties of the parts, as discussed in section 3.4. Therefore, it is important to carefully select and optimize the energy density and other process parameters to achieve the desired SDAS and microstructure of the parts. gradients and cooling rates, which can promote the formation of columnar grains.
3.4. Mechanical properties evaluation
The variation in hardness values with the layer number, wall number, energy density, and powder deposition density is important for understanding the microstructure and properties of the additively manufactured parts. The hardness values can affect the wear resistance, fatigue resistance, and other mechanical properties of the parts, as well as their suitability for different applications. Therefore, it is important to carefully select and optimize the energy density, powder deposition density, and other process parameters to achieve the desired hardness values and microstructure of the parts. The results of micro hardness test showed that the hardness of the parts varied with the energy density and powder deposition density, with the highest hardness occurring at the highest energy density and lowest powder deposition density. The hardness also varied with the distance from the build plate, with the highest hardness occurring at the top surface and decreasing towards the bottom. The microhardness results were consistent with the grain structure and porosity results presented in section 3.3. Table 3. shows the average Vickers hardness in different layers for the straight walls. The hardness values range from 311.59 ± 13.10 HV in layer 7 to 393.30 ± 29.33 HV in layer 2. The highest hardness value is almost 27% higher than the lowest hardness value, indicating a significant variation in hardness values with the layer number.The average hardness for the walls ranges from 331 HV to 354 HV for a wide variation in ED/PDD. It is slightly higher than the as-built PH 13 − 8 Mo part (336 ± 8 HV) [18] fabricated via wire arc additive manufacturing and the as-built additively manufactured stainless steel CX part (332 ± 10 HV) [67] manufactured via laser powder bed fusion. The hardness of the PH 13–8 Mo stainless steel can be increased by subjecting them to solutionizing and ageing heat treatment due to the precipitation of NiAl precipitates [68].
Table 3
The average Vickers hardness in different layers of the walls.
Layer Number
|
Average Hardness (HV)
|
1
|
390.89 ± 14.61
|
2
|
393.30 ± 29.33
|
3
|
345.44 ± 18.36
|
4
|
324.47 ± 13.45
|
5
|
320.13 ± 18.32
|
6
|
312.53 ± 9.21
|
7
|
311.59 ± 13.10
|
8
|
327.19 ± 17.83
|
9
|
320.36 ± 10.15
|
Figure 10 illustrates how the average Vickers hardness changes in each layer for each of the straight walls. The graph shows that the hardness values vary with the energy density and powder deposition density. For example, at an ED of 71.4 J/mm3 and a PDD of 0.0103 g/mm3, the hardness values range from 372.53 ± 9.21 HV in layer 2 of wall 4 to 428.30 ± 29.33 HV in layer 2 of wall 9 at an ED of 64.3 J/mm3 and a PD 0.0077 g/mm3. These results indicate that the energy density and powder deposition density can have a significant impact on the hardness values and microstructure of the parts. Furthermore, it was observed that there was a significant 35% increase in hardness between the second layer and the fifth layer in wall five, despite having the same ED and PDD ratios. This high value in hardness in the initial layers can be attributed to rapid cooling caused by the high thermal gradient. As the number of layers increases, the thermal gradient becomes less pronounced, resulting in a decrease in the cooling rate and subsequently a decrease in hardness. This phenomenon highlights the influence of layer height and cooling dynamics on the material properties during the additive manufacturing process.
Table 3 and Fig. 10 show a slight increase in the average Vickers hardness in the initial layers of the straight walls. This can be due to an increased cooling rate due to proximity to the substrate.
Table 4
As built horizontal tensile sample results extracted at different locations.
Specimen Number
|
Yield strength
(MPa)
|
Ultimate tensile strength(MPa)
|
Elongation (%)
|
A
|
744.79
|
1129.67
|
6.94
|
B
|
787.29
|
1182.12
|
7.28
|
C
|
641.00
|
1133.78
|
5.17
|
Average
|
724.3 ± 75
|
1148.5 ± 29
|
6.4 ± 1.1
|
The tensile test was conducted on the sample produced with an ED value is 53.57 J/mm3 and PDD ratio is 0.0067 g/mm3, and three tensile samples were extracted from different locations. The resulting test data is presented in the table below. For better visualization, Fig. 11 shows the stress elongation curve.
Minor variations are observed in the yield strength, ultimate tensile strength and % elongation for the test coupons extracted at different locations. This can be attributed to the thermal gradient occurred at different rates in different regions at different locations. The cooling rate varies because the substrate acts like a heat sink, interlayer dwell time and heat accumulation in the deposit as successive layers are deposited [69]. While the yield and fracture properties of these three different samples were very close to each other, the elongation amount of sample C was slightly lower, and it was assumed that porosity and other impurities might cause this. The average yield strength, determined to be 724.3 MPa, showcases the ability of the samples to resist deformation before experiencing plastic flow. Additionally, the average ultimate tensile strength of 1148.5 MPa reflects the maximum stress these samples can withstand before failure. Lastly, the average elongation of 6.4% demonstrates the extent to which the samples can stretch or deform before breaking. These average values provide valuable insights into the mechanical behavior of the samples, highlighting their strength and ductility characteristics. For PH 13 − 8 Mo stainless steel horizontal as-built samples fabricated via wire arc additive manufacturing technique, [69] reported average ultimate tensile strength (UTS) of (1115 ± 17 MPa) and percentage elongation of (11.4 ± 1.8%). In comparison, laser deposited horizontal as-built PH 13 − 8 Mo has a slightly higher average UTS, but the average % elongation is lower, as seen in Table 4. For stainless steel CX horizontal as-built sample fabricated via laser powder bed fusion, the authors reported yield strength of (1036 MPa), UTS of (1113 MPa) and % elongation of 21.7% [68]. In comparison, laser deposited horizontal as-built PH 13 − 8 Mo has higher UTS with lower yield strength and % elongation (see Table 4). For 17 − 4 PH horizontal as-built sample fabricated via laser metal deposition, Yu et al. reported yield strength of (758 ± 19 MPa), UTS of (1129 ± 5 MPa) and % elongation of (14.1 ± 0.3%) [27]. In comparison, laser deposited as-built PH 13 − 8 Mo has higher UTS with lower yield strength and % elongation (see Table 4).