3.2. Effect of SLM Process Parameters on Microstructure of 24CrNiMo Alloy Steel
Figure 5 shows the microstructure of the XOZ cross section of the sample under different laser scanning speeds and the same laser power (300 W). As shown in
Figure 5a, when the scanning speed was 500 mm/s, the melt channel existed in the form of droplets. As the scanning speed increased, the width and depth of the weld increased slightly, and the cladding layer tended to be flat (
Figure 5d,e), but when the scanning speed reached 1100 mm/s, the molten pool became larger and the cladding layer was uneven (
Figure 5f).
It can also be seen in
Figure 5 that there were different degrees of holes (black spots in the figure) in SLM-formed alloy steel, and the size of the holes ranged from a few microns to dozens of microns. The large-size pores were irregular, which were mainly formed at the bottom of the overlap zone between the molten channels, and there were spherical particles in some pores, indicating that these pores were mainly unfused pores. The formation of pores was the direct reason that the density of alloy steel was affected.
Figure 6 shows the microstructure of the XOZ section of the sample at different laser power and the same scanning speed (600 mm/s). The effect of laser power on the morphology of cladding layer was similar to that of scanning speed. Increasing laser power widened and deepened the weld. Under the condition of 350 W laser power (
Figure 6d), the cladding layer was flatter than other samples. Similarly, varying degrees of pores (black spots in
Figure 6) were also found in SLM-formed alloy steel in
Figure 6. When the laser power was less than 200 W, the sample forming showed a certain fluctuation, and the independent melting channel could be observed. With the increase of laser power, the fusion between the fusion channels was better, and the surface forming gradually became smoother and smoother. At the same time, the porosity also decreased significantly at medium scanning speed, indicating that the stability of the forming process was improved.
In the first orthogonal test, the samples under different laser processes showed defects such as spheroidization, pores and cracks, as shown in
Figure 7. The main reasons for the formation of cracks were analyzed: at medium scanning speed and low laser power, the amount of powder melted by laser energy was insufficient, resulting in poor stability of the molten pool, and the unstable molten pool was entrained in the surrounding metal powder. In the final solidification area, it was easy to form enriched oxides or inclusions, and finally cracks were formed under the action of stress. In the first orthogonal test, because the sample still had spherical pores and other defects, the optimization test was used to optimize the process parameters again, and the macroscopic morphology of the sample is shown in
Figure 8a.
Figure 8b shows the results of the sample density test. Obviously, the density of No. 9 sample was the highest, reaching 99.78%. After comprehensively considering other indicators such as microstructure and density, the optimal process parameters were determined as follows: laser power 300 W, scanning speed 530 mm/s, scanning spacing d = 0.11 mm, fixed layer thickness h = 50 μm.
Under the optimized process parameters, 24CrNiMo alloy steel samples were prepared by SLM technology. The surface morphology and low-power optical microstructure of the samples are shown in
Figure 9. The tensile strength and yield strength of the deposited sample reached 1199 MPa and 1053 MPa, respectively; the elongation after fracture was 10.8%, and the hardness was 365~388 HV
0.2. Compared with the test results of other groups, the surface of the SLM formed sample under this process parameter was smooth and well formed, the microstructure was basically free of defects, such as pores, and the density was higher under this process parameter. Therefore, the optimized process parameters obtained in the experiment were reasonable.
As shown in
Figure 10a, due to the forming characteristics of SLM, the martensite structure was easily formed in the alloy steel in the molten pool area. Under the reheating action of the subsequent cladding layer, the martensite was tempered to a certain extent. Due to the technological characteristics of SLM layer-by-layer stacking, the formed martensite in the molten pool changed differently with the distance from the new molten pool during the laser scanning in the next layer. The nearest structure was remelted to become part of the new molten pool. The area where the phase transition occurred when it was not melted and heated above the austenitizing temperature was the heat-affected zone, which became the dividing line between the new molten pool and the old molten pool. The structure of the old molten pool was tempered. The closer the structure was to the new molten pool, the higher the tempering temperature and the longer the tempering time and forming of lath martensite. The molten pool area near the bottom of the new cladding layer was obviously affected by heat, being heated to a higher temperature and undergoing continuous rapid cooling and solidification, and finally granular bainite was formed.
Figure 10b shows the microstructure of the SR sample, which shows that the microstructure of the sample had no obvious change (500×) compared with the deposited state after SR.
3.3. Test Results and Analysis of Surface Residual Stress
The test results of residual stress on the surface of samples in different states are shown in
Table 8. Since the scanning method was 67° rotation scanning, the sample was formed more uniformly, so the longitudinal residual stress and the transverse residual stress of the deposited sample were not much different.
The residual stress on the surface of the deposited sample was tensile stress, with a peak value of (453 MPa)~(583 MPa), which was about 50% of the yield strength of the material. After SR, the residual stress of the sample became compressive stress, and the peak value was concentrated in (−11 MPa)~(−22 MPa). The residual stress on the deposited surface was caused by the excessive temperature gradient. During the forming of the SLM, the cladding layer experienced a sharp increase and decrease in temperature within a short period of time, resulting in the temperature near the molten pool being much higher than others area, so a large temperature gradient was formed. In the subsequent cooling and solidification stage, the larger temperature gradient restricted the shrinkage and deformation of the cladding layer in the surrounding colder area, resulting in inconsistency of volume shrinkage and expansion, mutual restriction and large residual stress. It has been reported in literature that large tensile stress will lead to premature fatigue failure of parts [
21,
22,
23]. The SR heat treatment provided an activation energy, which helped the atoms that deviated from the equilibrium position to return to the equilibrium position, and promoted the precipitation of supersaturated solutes in the lattice, so the residual stress decreased. After SR, because the long holding time and slow cooling speed of the sample after heating, the cooling rate and time of the surface layer and the core were not much different, and the stress distribution tended to be consistent.
3.4. Test Results and Analysis of Gradient Residual Stress
Figure 11a, shows the results of the gradient residual stress test at the center of the deposited sample. The characteristics of the residual stress distributed along the thickness direction were as follows: the gradient residual stress had periodic fluctuations along the depth direction. The residual stress in different deposition layers was all tensile stress, and the closer to the surface it was, the greater the residual stress.
Figure 11b shows the gradient stress at the edge of the deposited specimen. The edge area distribution was the same as tensile stress, and the peak stress reached 730 MPa, indicating that the tensile stress at the surface edge of the sample was relatively large. Because the edge position of the sample was affected by the protective airflow and scanning strategy during the printing process, the residual stress fluctuated greatly. It can be seen from
Figure 11b that the change of residual stress in the edge region of the sample was similar to that in the central area, and it gradually decreased from the surface of the sample to the base layer. The uneven distribution of temperature caused different volume shrinkage in space, resulting in gradient residual stress. After each layer was formed, it cooled and shrank until the newly laid layer of powder was formed. At this time, the heat of the newly laid powder layer flowed into the previous forming layer, and the residual stress of the existing layer and the newly laid layer were readjusted. In addition, the stress caused by the deposition in front of the deposition layer was relaxed to meet the balance. Therefore, the gradient stress of the as-deposited sample decreased layer by layer from the surface to the substrate.