3.1 Parameter optimization of SLM
To obtain the best joint bonding properties, it is necessary to set up a multi-parameter window to screen out the best process parameters (laser power and scanning rate) combination of the two alloys. Based on Table 3, the parameters for forming 316L alloy were unified (320 W and 650 mm/s), while a total of nine sets of parameters (laser power: 250 W, 300 W and 350 W; scanning speed: 600 mm/s, 650 mm/sand 700 mm/s, respectively) were set to fabricate Inconel 718 components. It could be seen from Fig. 2(a) that the specimens under all nine combinations could be formed into a well-bonded multi-material, indicating the wide forming compatibility of steel and nickel alloys. Therefore, forming the respective dense parts became a priority, and finally, the parameter combination when forming the experimental samples (Fig. 2(b)) was selected as: 320 W and 650 mm/s of 316 alloy; 300W and 650 mm/s of Inconel 718.
3.2 Phase and microstructure analysis
Figure 3 depicted the XRD pattern of the SLM multi-material sample. There was almost exact matching of the characteristic peaks of the γ/γ'/γ'' phase. Besides, the weak diffraction (2θ = 46.3°) marked in the (211) plane could be identified in the specimen, which confirmed the existence of precipitated phase in the nickel alloy. Weak diffraction peaks of carbides (NbC) were also detected. At the same time, the strong diffraction peaks corresponding to the γ-Fe crystal planes (111), (200) and (220) could be clearly identified in the sample. In addition, no peaks of the martensite phase were detected.
Figure 4 showed the OM and SEM images of microstructure on XZ plane for the as-built sample. Figure 4(a), the OM image tested on the XZ plane of polished specimen, exhibited the combined effect and the metallurgical defects of 316L and Inconel 718. The OM image (Fig. 4(b)) after metallographic corrosion showed that the XZ plane could be divided into three distinct regions (distinguished by white dotted lines): the 316L region, the transition region and the Inconel 718 region, where the thicknesses of transition region accounted for approximately three scanning layers, which could be estimated as 150 µm. Overall, the as-built sample possessed solid metallurgical bonding as the transition area with clear boundaries existed between 316L and Inconel 718 region. Subtle pores could be observed in the 316L region while cracks were the dominant defects occurred in the transition region (defined as continuous and separate cracks, respectively), which was usually attributed to the insufficient relief of residual thermal stress [34–36].
Figure 4(c) and Fig. 4(d) were the typical SEM images corresponding to area A and area B marked by green dotted line in Fig. 4(b). From Fig. 4(c), columnar dendrite and cellular dendrite patterns occupied the separate molten pool of Inconel 718 in the transition region while the proportion of columnar dendrite far exceeded that of cellular dendrite. At the same time, columnar dendrite region Ⅰ and Ⅱ with different growth direction, which turned to be perpendicular to the boundary of molten pool, could be explained by the preferential growth of grains under different temperature gradients [37]. Moreover, common Laves phases of Inconel-based alloys were also found in columnar region with the long chain shape [38]. The formation of the Laves phase consumed the useful alloying element Nb in the matrix, thereby inhibiting the precipitation of the strengthening phases γ' and γ"[39]. Secondly, the brittle Laves phase provided conditions for the nucleation and growth of cracks under the action of residual stress or other external stresses, resulting in a significant decrease in the tensile properties, fracture properties and fatigue properties of the shaped parts [40]. In addition, the formation of low-melting eutectic Laves phase between dendrites was likely to cause thermal cracking in the additive manufacturing process.
An elevated percentage of cellular dendrite could be observed in Fig. 4(d), which totally located in Inconel 718 region. An overlapping region connected the cellular region Ⅰ and Ⅱ, with the average grain size calculated to be 1.2 and 0.7 µm, respectively. It had been reported that the fine equiaxed cellular dendrites were beneficial to reduce the segregation of Nb and form separated Laves phase particles, thus enhancing γ" phase strengthening phenomenon [41].
As shown in Fig. 5, to measure the grain orientations and textures of Inconel 718/316L multi-material sample in the as-built condition, EBSD mappings, as well as Z (BD) direction inverse pole figure (IPF) color mappings were constructed from the EBSD data to visualize both the grain morphologies and the crystallographic orientations. Figure 5(c) indicated the variation tendency of grain size related to the 316L region, transition region and Inconel 718 region along the Z (BD) direction. Typical grain morphologies of equiaxed fine grain and columnar grain were observed in 316L and Inconel region, while a visible grain coarsening transition region appeared between these two regions, and the average grain size were calculated to be 0.96 µm, 3.13 µm and 28.5 µm along the building direction. The coarsened grains proved to be the result of continued growth of 316L steel grains, which was attributed to the re-melting and recrystallization caused by the heat input effect when forming the Inconel 718 part and the continued growth of the grains caused by the temperature gradient and cooling rate declining [42] and finally, exacerbating the weakening of the interface performance.
The IPFs indicated that the Inconel 718 region of as-built sample had < 001>// Z (BD) as the preferential crystallographic orientation. Because the texture intensity in < 001 > direction was measured to be 16.25, characterized by strong fiber texture. The texture orientation of the Inconel 718 depended on the competition between the local heat flow direction and the multiple preferential growth directions of the material itself. Primary columnar dendrites that were parallel or nearly parallel to the heat flow direction (< 100 > orientation) would preferentially grow and other columnar grains, deviating from it greatly, would be eliminated [43]. Meanwhile, the texture of the grain coarsening transition region exhibited a weak {001} <100 > cubic texture belonging to the main FCC metal recrystallization texture, and its texture intensity was merely 3.32 (Fig. 5(g)).
3.3 Mechanical properties
Figure 6 illustrated the corresponding relationship between the micro-hardness variation of the Inconel 718/316L multi-material interfacial bonding region and the indentation image of sample after Vickers hardness test. At least three points were selected from each region, Fig. 6(b) showed the ten representative indentation points. The micro-hardness values of 316L region, transition region and Inconel 718 region were measured to be 383.7HV, 303.8HV and 320.7HV, respectively. Compared with the Inconel 718 region, the lower hardness value of the transition region with its microstructure dominated by columnar dendrite grains of Inconel 718 was attributed to the suppression of the γ" phase strengthening [41] caused by the formation of the Laves phase and the segregation of Nb as well as the crack initiation, which had been discussed in Fig. 4(c).
Figure 7 showed the tensile samples before test and the stress-strain curves in four states. The average ultimate tensile strength (UTS) of 316L and Inconel 718 samples were measured to be 1007.5 ± 21.3 MPa (Fig. 7(a)) and 986.2 ± 33.5 MPa (Fig. 7(b)), with elongation of 12.3 ± 0.7% and 22.4 ± 1.3%, respectively. In order to investigate the performance of the overall mechanical properties when the multi-material samples were subjected to different direction of loading force, relevant test on tensile specimens in horizontal and vertical combination were carried out. As shown in Fig. 7(c), the average UTS and elongation of horizontally combined Inconel 718/316L samples were 795.4 ± 2.3 MPa and 16.3 ± 0.9%. A reduction of approximately 21.1% and 19.3% occurred in average UTS comparing to the parent 316L and Inconel 718 alloy. This phenomenon could be explained as: (ⅰ) initial cracks existed in the transition region (Fig. 2(a)) propagated and grew rapidly when the samples were subjected to external load, and these places became the preferential fracture failure sites as a result of stress concentration; (ⅱ) a lack of γ" phase strengthening [41], as well as a weaker fine grain strengthening [44] led by coarser average grain size in transition region, which had been discussed above.
combined Inconel 718/316L; (d) vertically combined Inconel 718/316L.
Figure 8 demonstrated the tensile fracture morphology of horizontally combined Inconel 718/316L samples in details. Clear and defect-free boundaries found in transition region (Fig. 8(b)) further confirmed a dense metallurgical bonding happened. Typical cleavage and tongue patterns of brittle fracture were observed around the transition region, and the fracture mechanism could be concluded as: the rough and interconnected long-chain Laves phase particles in the sample provided a favorable location for the micro-pore nucleation and macroscopic crack propagation [45] and finally, leaded to brittle trans-granular fracture.
The average UTS and elongation of vertically combined Inconel 718/316L samples were 867.2 ± 21.3 MPa and 20.5 ± 0.3%. However, the fracture sites (as shown in Fig. 9) of all the samples were not located in the interface bonding area which meant the decline in average UTS of vertically combined sample was not related to the bonding effect of 316L and Inconel 718, and indicated a strong metallurgical bonding obtained in Inconel 718/316L bimetallic multi-material.
Three-point bending test could be used to characterize the deformation ability of multi-materials subjected to Z-axis load and reflect the internal defects of the multi-material parts. Long strip specimens of 80 mm in length, 10 mm in width and 10 mm in thickness (the thickness of 316L and Inconel 718 section was 5 mm each) were tested at a rate of 1 mm/min by an indenter with a radius of 10 mm in a three-point bending tester. And bending deformation occurred until the specimens were bent to broken, with the support span of 54 mm. Figure 10 demonstrated typical bending stress-strain curves for each type of specimen. The results of flexural strength and elongation were taken as the average values of three bending stress-strain curves. It could be seen that the flexural strength of 316L, Inconel 718, Inconel 718/316L of 316L at the bottom and Inconel 718 at the bottom were 2455.8 ± 39.7 MPa, 2443.9 ± 31.8 MPa, 1623.6 ± 27.7 MPa and 2462.9 ± 40.7 MPa, respectively.
Inconel 718/316L specimen of 316L at the bottom; (d) Inconel 718 at the bottom.
Moreover, Fig. 11 showed typical object diagram of three-point bending test specimens with curved fracture cracks. The elongation of 316L specimen, Inconel 718 specimen, Inconel 718/316L specimen of 316L at the bottom and of Inconel 718 at the bottom were 1.09 ± 0.13 mm/mm, 2.88 ± 0.14 mm/mm, 0.31 ± 0.05 mm/mm and 1.90 ± 0.16 mm/mm, respectively. The average flexural strength and elongation obtained in Inconel 718/316L of 316L at the bottom were the minimum compared with that of the other three types. The results showed that defects such as cracks at the joints of multi-material samples did have a detrimental effect on the overall bending performance of the components. And when the sample received a bending load, the crack was more likely to grow towards the 316L side, further causing bending failure. At the same time, no macro cracks appeared along the direction of the bonding interface even if the multi-material specimen failed to bend, which meant these two specific alloys were tightly bonded.
The joint shear strength of two SLM specimens and the entity graph before stretching were shown in Fig. 12. The samples dimensions were 30 × 10 × 10 mm, which 316L and Inconel 718 accounted for 5 mm, respectively. The elevating rate of the indenter loading force was 300 N/m. The shear strength results were obtained when the 316L and Inconel 718 portions were completely separated. Sample 1 and sample 2 achieved the shear strength of the SLM joint to approximately 461 and 438 MPa. Typical Inconel/steel multi-materials formed by specific method and its shear strength [46–49] were concluded in Table 4. The average shear strength of the obtained sample was 449.5 MPa. Comparing the table data, it could be found that the shear strength of Inconel718/316L multi-material obtained by SLM was relatively high, which fully proved the high metallurgical bonding of dissimilar alloy joints.
Table 4
Summary and comparison of shear strength values involved in different processing methods followed by post-processing.
Fabricating method | Experimental subjects | Post-processing | Shear strength (MPa) | Ref |
Inconel alloy | Steel alloy |
Induction brazing | Inconel X-750 | 304 SS | − | 483 | [46] |
Gas tungsten arc welding | Inconel 617 | 310 SS | − | 445 | [47] |
Explosive welding | Inconel 625 | P355NH | − | 572 | [48] |
Stress relief annealing | 591 |
Normalizing | 383 |
Infrared brazing | Inconel 601 | 422 SS | − | 362 | [49] |
Selective laser melting | Inconel 718 | 316L SS | − | 449.5 | Our work |