| Cite this article as: |
Miaomiao Chen, Renhai Shi, Zhuangzhuang Liu, Yinghui Li, Qiang Du, Yuhong Zhao, and Jianxin Xie, Phase-field simulation of lack-of-fusion defect and grain growth during laser powder bed fusion of Inconel 718, Int. J. Miner. Metall. Mater., 30(2023), No. 11, pp.2224-2235. https://doi.org/10.1007/s12613-023-2664-z
|
Inconel 718, a nickel-based superalloy, is widely used in high-temperature structural parts in aerospace, nuclear power plants, and ships due to its excellent corrosion resistance, fatigue resistance, and mechanical strength [1–3]. Laser powder bed fusion (L-PBF) allows rapid manufacture of components with complex geometries and has attracted extensive attention in the aerospace industry in recent years [4–6]. Compared with traditional casting and forging, Inconel 718 alloy components fabricated via L-PBF such as turbine disks, supports, and fasteners in aerospace engines have fine grain structure, which could improve their mechanical properties [6–9].
Due to the high cooling rate and temperature gradient during L-PBF, columnar dendrites with <100> orientation mainly grow epitaxially along the building direction. The anisotropy of mechanical properties can significantly reduce the life of components in practical applications (e.g., hot-end parts of the disk and shaft) [10–11]. Thus, controlling columnar dendrite epitaxial growth and refining the grain structure to reduce anisotropy are primary challenges in the field of superalloy additive manufacturing [12]. The microstructure of samples prepared via L-PBF are affected by the laser power, scanning rate, and scanning strategy. The microstructure anisotropy can be reduced by adjusting the process parameters. Wan et al. [13] obtained bimodal grain structure and oriented columnar grain structure with layer rotations of 0° and 90°, respectively. Gokcekaya et al. [14] obtained single-crystal, lamellar, and equiaxial microstructure by adjusting the laser power and scanning speed. The results showed that decreasing laser power or increasing scanning speed could suppress epitaxial grain growth and produce an equiaxed microstructure. However, low laser energy density can lead to defects such as lack-of-fusion (LoF) as a result of insufficient melting, reducing the mechanical properties [15].
Thus, the L-PBF process window that suppresses the epitaxial grain growth and prevents formation of LoF defects must be established. However, establishing the L-PBF process window using experimental methods is costly and time-consuming. The integrated computational materials engineering (ICME) approach can simulate microstructure evolution during L-PBF with a wide range of process parameters, reducing cost and accelerating process design [16]. Of the microstructure simulation methods, the phase-field method (PFM) can provide a detailed microstructure description and can be coupled with other physical fields to more accurately describe microstructure evolution. It has become one of the most commonly used modeling techniques to simulate microstructure [16–21]. The finite element method (FEM) can effectively simulate the temperature field of the L-PBF process with different process parameters for microstructure simulation [16,19].
A cross-scale simulation method combining the FEM and PFM is used in this study to simulate the temperature field and microstructure evolution of Inconel 718 alloy during L-PBF. A computational model describing LoF defects and epitaxial grain growth is established for microstructure and defect control. The research strategy is presented in Fig. 1. The phase-field (PF) model is combined with the FEM to establish the dependence between laser-scanning parameters and the solidified microstructure and defects. The L-PBF scanning parameters are optimized to eliminate LoF defects, suppress epitaxial grain growth, and improve the strength of as-deposited samples. The strategies proposed in this study could provide guidance for the microstructure and defect control of the alloys that prepared by L-PBF in future.
The temperature field of the L-PBF process was simulated via a moving heat-source model and birth-death element technology in ABAQUS software [22]. Local rapid heating and melting and a deep molten pool are the main characteristics of the L-PBF process. The laser heat source was loaded on the powder layer by means of heat flux, and its intensity conformed to a Gaussian distribution. Thus, a Gaussian heat source was selected as the heat-source model. Loading of the ellipsoid heat source and scanning path was realized by writing and importing a Fortran subroutine. The ellipsoid heat source model is expressed as follows [23].
| Q(x,y,z)=6√3λPabc×3√πexp(−3x2a2−3y2b2−3z2c2) | (1) |
where Q(x,y,z) is the heat flux in three-axis coordinates of x, y, and z; λ and P denote the absorption rate of the laser energy and the laser power, respectively; a, b, and c are parameters related to the shape of the heat source distribution, as shown in Fig. 2(a).
This study only uses the FEM to simulate the temperature evolution during the L-PBF. To simplify the calculation process and meet the requirements for temperature field calculation, the following assumptions are adopted for the FEM calculation model: (i) There is no energy exchange with the outside in the calculation domain. (ii) The flow of liquid metal in the melt pool is ignored, and only the heat transfer process of the workpiece is considered. In addition, the thermal conductivity, specific heat capacity, and material density are isotropic and temperature-dependent, and they vary with temperature as shown in Table 1 [24]. Other thermal property parameters are assumed to be constant, with a latent heat of 2.27 × 105 J·kg–1, a solidus temperature of 1170°C, and a liquidus temperature of 1337°C. (iii) Material evaporation, Marangoni effect, surface tension, and defects formed in the L-PBF are ignored. (iv) The birth-death element technology is used to simulate the layer-by-layer powder spreading process. The model used for ABAQUS finite element calculation is a geometric body with eight layers, 400 μm × 400 μm × 320 μm, as shown in Fig. 2(b). Convective and radiative boundary conditions are created in the Interact module of ABAQUS, with a convective heat transfer coefficient of 20 W·m–2·°C–1 and a radiative heat transfer coefficient of 0.7. The initial temperature field of the workpiece is set as 20°C.
| Temperature / °C | Density / (g·cm–3) | Thermal conductivity / (W·m–1·K–1) | Specific heat capacity / (J·kg–1·K–1) |
| 25 | 8.24 | 13.45 | 408 |
| 300 | 8.14 | 17.59 | 502 |
| 600 | 8.01 | 21.77 | 575 |
| 900 | 7.87 | 25.96 | 723 |
| 1200 | 7.72 | 30.18 | 819 |
| 1500 | 7.32 | 30.95 | 856 |
DownLoad:
CSV
The multiphase-field model, as implemented in the commercial phase-field software MICRESS (version 6.400, Access e.V., Aachen, Germany), is coupled with the CALPHAD. The model description has been presented [25–27]. The MICRESS software is coupled with Thermo-Calc software through the TQ interface to obtain thermodynamic and kinetic data from the TCNI9 and MOBNI5 databases. To simplify the composition, the Inconel 718 alloy is defined as a seven-component system in the simulation, as shown in Table 2, and the influence of other trace elements on the solidified microstructure is ignored. The two-dimensional PF model is used to simulate the grain evolution. In the computational domain, the temperature field data during the L-PBF are obtained from the ABAQUS simulated results. The positions and temperature distribution curves of each node are shown in Fig. 2(c)–(d). To simulate the grain evolution across multiple layers during L-PBF, a longitudinal section calculation domain with a size of 200 μm × 200 μm was selected, with a mesh size of 0.5 μm, an interface energy of 1.2 × 10–5 J·cm–2, and a kinetic coefficient of 0.025 cm4·J–1·s–1. The diffusion coefficient and molar volume of each element were obtained from Thermo-Calc.
| Ni | Cr | Fe | Mo | Nb | Ti | Al |
| 54.7 | 19.0 | 16.7 | 3.2 | 5.0 | 1.0 | 0.5 |
DownLoad:
CSV
In the computational domain, periodic boundary conditions are used to solve the left and right boundary grid nodes to eliminate the boundary effect of simulated results and construct a complete continuum. Insulating boundary conditions are used at the bottom and top boundary grid nodes to achieve oriented growth of grains from the bottom to top in the additive manufacturing process. The main focus of the phase-field model calculation is the process of grain growth, rather than the more complex microstructure of dendritic structures. Thus, the following assumptions are adopted in the PF model: (i) The effects of fluid flow and material shrinkage on microstructure evolution during L-PBF are ignored. (ii) Only the temperature evolution of each layer along the building direction is considered to investigate the influence of the depth of the molten pool on the grain growth, however, the influence of the transverse shape of the molten pool is ignored. Furthermore, the liquid–solid/grain transformation depends only on the temperature and concentration fields.
Reducing laser energy density is an effective way to prevent grain growth, but laser energy density that is too low results in insufficient melting and formation of LoF defects. During laser powder bed fusion, due to the approximate unidirectional heat flow and remelting caused via the next laser scanning, it is easy to obtain coarse columnar grains. If fine grain regions are presented between two adjacent layers, it means that the remelting is insufficient, which is easy to produce LoF defects. This is consistent with our experimental observation. As can be seen from Fig. 3, there are fine grain (the average size is 2.3 μm) regions near LoF defects.
In view of the fact that it is very difficult to simulate the formation and evolution of defects directly with the phase-field method. In this work, an indirect modeling idea for “lack-of-fusion defect” is proposed: the LoF defect is described by very fine grain structure. To establish the LoF model, two types of matrix phase nucleation are applied in the PF model, as shown in Fig. 3(a). Seed type I is nucleated due to local undercooling during L-PBF, seed type II is nucleated at the top of each layer. The new nucleated grains gradually grow in the subsequent laser scanning process, whereas the nucleated grains grown from seed type I are retained in the solidified microstructure in the form of stray grains. As shown in Fig. 3(a), the fine grains grown from seed type II are located at the position of the powders laid in the next layer (N+1). These fine grains are melted during the next laser scanning, and their melting state represents the melting state of the powders. If these fine grains are not fully melted in the subsequent laser scanning process, a small grain enrichment region forms, indicating the presence of LoF defects in the as-deposited microstructure.
To determine the nucleation parameters of seed type II, the microstructure of the sample with LoF defects was analyzed. Fig. 3(b) shows the electron backscatter diffraction (EBSD) orientation map of the as-deposited sample along the building direction with a laser power of 150 W and scanning speed of 960 mm·s–1. Lower energy density (E = 39.1 J·mm–3) resulted in LoF defects in the as-deposited microstructure. In the region with LoF, there were fine equiaxed grain accumulation zones with an average grain size of 2.26 μm and unlabeled black regions. Thus, the minimum nucleation spacing of seed type II was set as 2 μm. To determine the nucleation parameters of seed type II, the PF simulated results were compared with the experimental results in terms of LoF defects. It was determined that a large number of seed type II nucleate uniformly at the top of each layer when the solid fraction is 85vol%. The determination and verification of the nucleation conditions are detailed in Section 4.2.
To clarify the relationship between the laser scanning parameters and grain characteristics, an orthogonal experiment was conducted. The laser power, scanning speed, and rotation angle are indicated in Section 4.1. Seven values were determined for each parameter, as shown in Table 3, for the orthogonal experimental design. An orthogonal table of L49(78) was used to design 49 groups of simulations shown in Table S1 (see the supplementary material).
| Laser power / W | Scanning speed / (mm·s–1) | Layer rotation angle / (°) |
| 150, 200, 250, 275, 300, 325, 350 | 700, 800, 900, 960, 1000, 1200, 1400 | 0, 15, 30, 45, 67,75, 90 |
DownLoad:
CSV
Spherical Inconel 718 powders with a size range of 15–53 μm were prepared via gas atomization technology, and the chemical composition are presented in Table 4. The powders were dried at 100°C for 8 h and deposited on 304L stainless steel substrate with a size of 250 mm × 250 mm. As-deposited Inconel 718 samples were fabricated on a BLT A300 machine equipped with a maximum 500 W fiber laser operating in the continuous wave mode with a Gaussian intensity distribution and a 76 μm spot size (1/e2). The process was carried out under a constant flow of ultrahigh purity Ar gas (99.999vol%) to reduce the oxygen content below 200 × 10–6. The zigzag scanning strategy was used during L-PBF.
| Ni | Cr | Fe | Mo | Nb | Ti | Al | Co | Si | Cu | Mn | C |
| 54.66 | 18.98 | Margin | 3.15 | 5.0 | 0.98 | 0.48 | 0.12 | 0.07 | 0.034 | 0.027 | 0.024 |
DownLoad:
CSV
A Phenom ProX scanning electron microscope (SEM) and an Axio Imager metallographic microscope were used to observe the microstructure and molten pool morphology. The polished samples were electropolished at a voltage of 20 V for 15 s using the reagent (10vol% HClO4 + 90vol% C2H5OH). A ZEISS Gemini SEM 500 scanning electron microscope and a Symmetry S2 scanning electron microscope equipped with electron backscatter diffraction (EBSD) were used to examine the microstructure. The grain morphology, size, and orientation of the samples and the required orientation and polar diagrams were obtained via Channel 5 software. The size of tensile samples with a thickness of 1.5 mm at room temperature is shown in Fig. 1. According to the GB/T2281—2010 standard, mechanical properties were measured using a CTM2500 microcomputer-controlled electronic universal testing machine equipped with an extensometer with a drawing speed of 1.0 mm·min–1.
In our previous work [28], the influence of scanning speed (V = 720–1200 mm·s–1) and laser power (P = 200–350 W) on the density and mechanical properties of the Inconel 718 fabricated via L-PBF was studied via trial-and-error experimentation. The process parameters (laser power of 300 W, scanning speed of 960 mm·s–1, layer thickness (T) of 40 μm, and interlayer rotation angle of 90°) yielding superior comprehensive mechanical properties were experimentally screened. The microstructure is shown in Fig. 4(a)–(b). The relative density of the sample was approximately 99.9% and only a few gas pores were observed in the microstructure, as shown in Fig. 4(a). The grains grew into coarse columnar dendrites with an average grain size of 13.4 μm. The ultimate tensile strength (UTS), yield strength (YS), and elongation of the as-deposited sample were (1021 ± 10) MPa, (742 ± 6) MPa, and 27.7%, respectively. Fig. 4(c) summarizes the distribution of mechanical properties of Inconel 718 alloy fabricated via L-PBF in the literatures (without distinguishing the tensile direction), which was also shown in Table S2. As can be seen, UTS and YS were distributed within the range of 800–1100 MPa and 500–800 MPa, respectively. The elongation was distributed within the range of 10%–40%. Considering the coarse columnar dendritic microstructure of the Inconel 718 alloy fabricated via L-PBF in previous experiments, there is still room for further strength improvement.
The laser power (P), scanning speed (V), and rotation angle (A) are the key process parameters that affect the grain morphology and size of the as-deposited Inconel 718 alloy [29–32]. The relative density with different energy densities (E = P / (VTH), where H is the hatch spacing, T is layer thickness) of samples is shown in Fig. 4(d), and detailed parameters are presented in Table S3. The results indicate that the energy density with the relative density exceeding 99% is in the range of 40–125 J·mm–3. Due to the LoF defects caused via lower energy density and the large number of pores caused via higher energy density, the relative density of as-built samples at higher (E > 125 J·mm–3) or lower (E > 125 J·mm–3) energy density is lower (RD ≤ 99%). Thus, the optimal energy density was in the range of 40–125 J·mm–3, with the laser power between 150 and 350 W. The scanning speed was in the range of 700–1400 mm·s–1. An orthogonal experimental method was used to design simulation cases within the investigated ranges of laser power, scanning speed, and interlayer rotation angle A (0°–90°), as presented in Table S1, to accelerate process optimization.
Accurate simulation of the temperature field during the L-PBF formation process is essential for subsequent microstructure simulation. The morphology and size of the molten pool with a single scanning track were calibrated by adjusting the powder absorption rate (λ) and heat source-related parameters (a, b, c). The temperature region higher than the melting point (1337°C) of the Inconel 718 alloy is displayed in gray, and is the molten pool region. With λ = 0.52, a = 72 μm, b = 72 μm, and c = 150 μm, the simulated and measured melt pool morphologies with different laser powers (150–300 W) are shown in Fig. 5. Both the depth and width of the molten pool increased with increasing laser power. When the laser power increased from 150 to 300 W, the width (W) of the remelted molten pool in the solidified part affected by the next layer laser increased from (75.8 ± 0.63) to (116.4 ± 4.71) μm. The molten pool depth (D) increased from (26.1 ± 0.13) to (69.6 ± 2.02) μm. With a single scanning track, the calculated W and D of the remelted molten pool were within 5% of the experimental error, indicating that the simulated results are accurate and reliable.
The depth of the molten pool is a key factor affecting the growth of columnar dendrites and powder melting. However, the temperature field in the MICRESS phase-field model is one-dimensional. Thus, temperature variation along the forming direction was applied in the PF simulation to investigate the influence of melt pool depth on epitaxial grain growth along the building direction. The effect of melt pool width perpendicular to the forming direction on grain evolution was neglected. To establish the relationship between the L-PBF scanning parameters and the distribution characteristics of grain and LoF defects, five layers (layer thickness of 40 μm) were selected in the PF model to simulate grain evolution in the calculation domain within a length range of 200 μm along the building direction. Additionally, the LoF model was designed in the PF simulation using the region enriched with fine grains after L-PBF to represent the LoF defects. The temperature field in the PF simulation was obtained from the FEM simulated results. The positions of each node and the temperature distribution curve are shown in Fig. 2(c)–(d). The evolution process of temperature field and grain structure with a fine grain enrichment zone (LoF defect) in the simulated results is shown in Fig. 6. The initial grains grew directionally from the bottom. During the evolution of the temperature and concentration fields, the grains nucleated due to undercooling. The nucleated grains uniformly distributed at the top of each layer were gradually grown, remelted, and epitaxially grown. As the laser scanning paths were different from each layer, the melting depth of each layer grain structure was different along the building direction. As shown in Fig. 6, regions enriched with fine grains (LoF defects) were found at the bottom of the fourth and fifth layers due to insufficient melting depth. The results indicated that the PF method can effectively simulate the layer-by-layer nucleation, growth, remelting, epitaxial grain growth, and possible occurrence of LoF defects in L-PBF.
Uniform nucleation in the LoF model begins in the late stage of solidification. Thus, in the range of 70%–100% solid phase volume fraction (fs) in the late solidification stage of each layer, with an interval of 5% (fs = 70%, 75%, …, 100%, respectively), the uniform seeds are nucleated. Comparison the simulation and experimental results in different parameter conditions, when the solid fraction of each layer of solidification was 85vol%, uniform nucleation in the LoF model began at the top of each layer. If there are regions enriched with fine grains in the final calculated results, LoF defects appear in the experimental results in the corresponding experimental conditions. Fig. 7 shows the PF simulated results and EBSD image with different laser power P (150 and 275 W) and interlayer rotation angle A (90° and 30°). The results show that the PF model can effectively simulate regions enriched with fine grains at the bottom of the molten pool, the LoF defects under low laser power, and the columnar dendrites growing through the multilayer under high laser power. As shown in Fig. 7(a) and (e), the fine grain-enriched regions in the simulated results correspond to the LoF pores in the EBSD results. When the laser energy density is sufficient, the powders are completely melted. The solidified layer is remelted to a certain extent, leading to epitaxial grain growth. If the energy density is too high, a large number of columnar dendrites grow across multiple layers in the as-deposited structure, aggravating the anisotropy of the alloy. As shown in Fig. 7(b)–(h), when the laser energy density is 71.6 J·mm–3 (P = 275 W, V = 960 mm·s–1), a large number of columnar dendrites grow throughout the entire computational area with different interlayer rotation angles, which demonstrating that the interlayer rotation angle plays an important role on grain evolution. The simulation and experimental results indicated that the LoF model in the phase-field method can simulate epitaxial grain growth and prediction of LoF defects with different laser scanning parameters.
To establish the L-PBF process window for Inconel 718 alloy that prevents the growth of grains and formation of LoF defects, the simulated results of an orthogonal experimental design within the parameter range were statistically analyzed. Fig. 8 shows the simulated results under different process parameter conditions. The solid black points (●) represent the absence of LoF defects in the results, and the hollow blue points (○) represent the presence of LoF defects in the results. In Fig. 8(b), it is observed that when the laser energy density E is less than 40 J·mm–3, LoF defects are present in the sample with any interlayer rotation angle. When E ＞ 72 J·mm–3, LoF defects are not present in the sample. The optimal grain structure with refined grains and no LoF defects is located in the green region shown in Fig. 8(b). Thus, to obtain the optimal grain structure, it is essential to determine the critical laser energy density Ec, which defines the occurrence of LoF defects. The range of laser energy density where Ec may exist was further reduced to 40–72 J·mm–3 to cover the range (green region shown in Fig. 8(b)) for all interlayer rotation angles.
The evolutions of grains with a laser energy density of 40–72 J·mm–3 were simulated based on the orthogonal experiments and the verified PFM parameters. Some of the simulated results are shown in Fig. 9, showing the grain distribution with different scanning parameters. When the melting depth was insufficient (P = 200 W and V = 900 mm·s–1), there were several regions enriched with fine grains (LoF defects in the PF simulated results). Conversely, when the melting depth was excessive (P = 350 W, V = 1000 mm·s–1), a large number of grains exhibited epitaxial growth. To prevent epitaxial grain growth, it is necessary to accurately control the melting depth. According to the simulated results, the melting depth is related to laser power, scanning speed, and scanning strategy. Thus, the relationship between scanning parameters (laser energy density and interlayer rotation angle) and characteristics of grains in the computational results was investigated to optimize the solidified structure of Inconel 718 alloy fabricated via L-PBF.
A structure with more grains, fewer long columnar dendrites, and no fine grain-enrichment region (LoF defects) in the simulated results is the optimal solidification structure. Thus, the number of grains and the ratio of columnar dendrites with a length (along the growth direction) greater than 200 μm (the length of the calculation domain) were counted to analyze the grain structure with different scanning parameters. The distribution of the number of grains and the ratio of columnar dendrites with laser energy density of 40–72 J·mm–3 and interlayer rotation angle of 0°–90° are shown in Fig. 10(a)–(b). The solid points (●) represent the absence of LoF defects in the simulated results, the hollow points (○) represent the presence of LoF defects in the results, the black boxes (□) represent verifications of the possible range of critical laser energy density Ec, and the red stars (☆) represent experimental verifications of optimal solidification structure. As shown in Fig. 10(a)–(b), multiple simulations were added within the range of Ec in order to narrow the range. The range of Ec with different interlayer rotation angles was controlled within 5 J·mm–3. The results indicate that as the laser energy density increased, the grain number density gradually decreased, and the ratio of longer columnar dendrites increased.
The density of the contour lines of grain number reflects the degree of the influence of laser energy density on the microstructure. As shown in Fig. 10(a), with different layer rotation angle conditions, the density of contour lines of grain number is different in the range (green area) of Ec. With a layer rotation angle between 15° and 67°, most of the contour lines of grain number are relatively dense, indicating that the grain structure is more sensitive to laser energy density within the range of Ec. Small changes in energy density can lead to significant changes in the grain structure. However, with layer rotation angles of 0°–15° and 67°–90°, the contour lines are sparse. In this range, the grain structure can be controlled via laser energy density to obtain an as-deposited structure with refined grains and without LoF defects. Three black boxes (□) at points with sparse grain number contour lines shown in Fig. 10(a) were selected to verify the accuracy of the simulated results. The three data points are the left boundary points in the range of Ec with interlayer rotation angles of 15° and 67°, respectively, and the right boundary point with sparse grain number contour lines. The results of the three experiments are shown in Fig. 10(c)–(e). LoF pores were present in the as-deposited structure of Fig. 10(c) and (e) at the left boundary of the Ec range, whereas only a few gas pores were present in Fig. 10(d) at the right boundary of the Ec range, which is consistent with the simulated results. The results show that Ec for L-PBF Inconel 718 alloy with refined grains and without LoF defects occurs in the range of 55.0–62.5 J·mm–3 with any layer rotation angle. When E < Ec, LoF defects are present in L-PBF Inconel 718 alloy; when E > Ec, there are no LoF defects in L-PBF Inconel 718 alloy.
To obtain L-PBF scanning parameters with the target solidification structure, three points (A1–A3) were verified for microstructure and mechanical properties within the range of laser scanning energy density of 55.0–62.5 J·mm–3 in Fig. 10(b). A1 and A3 (ratio = 3.0) with low and similar ratios of columnar dendrites with interlayer rotation angles of 15° and 67°, and A2 (ratio = 2.6) with a finer grain structure but possibly containing LoF defects. According to the laser scanning parameters of Fig. 10(c)–(e) shown in Fig. 10, the scanning speed was adjusted to match the laser energy density of A1–A3. The laser scanning parameters are shown in Table 5.
| Point | Laser power / W | Scanning speed / (mm·s–1) | Energy density / (J·mm–3) | Interlayer rotation angle / (°) |
| A1 | 280 | 1160 | 61.42 | 67 |
| A2 | 280 | 1175 | 59.57 | 52 |
| A3 | 200 | 880 | 56.82 | 15 |
DownLoad:
CSV
Mechanical properties and EBSD maps with three parameters A1–A3 are shown in Fig. 11. The UTS of (1111 ± 3) MPa and YS of (820 ± 7) MPa at point A1 were the highest. The elongation at point A1 was 28.0% ± 0.5%. With the parameters of point A2, the grains were the smallest, with an average grain size (radius) of 5.34 μm. However, there were some LoF pores (black areas), thus, the mechanical properties were poor. With the parameters of points A1 and A3, there were no LoF defects. The average grain size (7.0 μm) of point A1 was lower than that of point A3 (7.72 μm), that is, the number of grains in the same size domain was greater with the parameter conditions of point A1. Comparison of A1 and A2 shows that the number of grains is not a decisive factor affecting the properties of the as-deposited alloy. The LoF defects result in a more significant loss of mechanical properties. Comparison of the results for A1 and A3 shows that with no LoF defects, more grains result in a smaller average grain size and higher strength. With the basic parameters (laser power = 300 W and scanning speed = 960 mm·s–1), the UTS and YS were (1021 ± 10) MPa and (742 ± 6) MPa, respectively. As shown in Fig. 4(c), compared with the basic parameters, the UTS and YS of point A1 increased by 8.8% and 10.5%, with the computationally screened parameters, respectively.
In summary, calculation combining the FEM and PFM and the simulation of epitaxial grain growth across multiple layers during L-PBF was realized in this work. Additionally, LoF model was proposed in the PF model to predict LoF defects. When the solid phase fraction of each layer was 85vol%, a large number of fine grains nucleated at the top of the layer and remelted, epitaxially grew, or formed LoF defects during subsequent processes. Using FEM and PFM together with the LoF model, the optimal grain structure could be screened out with limited tests. This work provides a feasible calculation approach that is helpful in tuning and optimizing the grain structure of parts fabricated via L-PBF.
A simulation combining the phase-field method and finite element method was conducted to investigate the effect of laser scanning parameters on epitaxial grain growth and LoF defect formation in Inconel 718 alloy. The following conclusions were drawn:
(1) The critical laser scanning energy density (Ec) defines the occurrence of LoF defects. When the interlayer rotation angle was 0°–90°, Ec varied in the range of 55.0–62.5 J·mm–3. When E < Ec, LoF defects were produced in the Inconel 718 alloy; when E > Ec, the grain coarsened with an increase in energy density.
(2) A refined grain structure with an average grain size of 7.0 μm without LoF defects was obtained in the simulation when the parameters were optimized as laser power of 280 W, scanning speed of 1160 mm·s–1, rotation angle of 67°, and Ec of 61.4 J·mm–3. The ultimate tensile strength ((1111 ± 3) MPa) and yield strength ((820 ± 7) MPa) of the as-deposited Inconel 718 alloy were increased by 8.8% and 10.5%, respectively, compared with specimens fabricated with parameters determined via the conventional trial-and-error approach.
(3) When the solid fraction of each layer reaches 85vol%, a large number of fine grains nucleate at the top of each layer. If these fine grains are not sufficiently remelted during L-PBF, it is likely to produce LoF defects.
This work was supported by the National Key Research and Development Program of China (No. 2021YFB3700701), the National Natural Science Foundation of China (Nos. 52090041 and 52022011), the National Major Science and Technology Projects of China (No. J2019-VI-0009-0123), and the Key-area Research and Development Program of Guangdong Province (No. 2019b010943001).
The authors declare that they have no known competing interests or personal relationships that could have appeared to influence the work reported in this paper.
The online version contains supplementary material available at https://doi.org/10.1007/s12613-023-2664-z.
| [1] |
M.D. Sangid, T.A. Book, D. Naragani, et al., Role of heat treatment and build orientation in the microstructure sensitive deformation characteristics of IN718 produced via SLM additive manufacturing, Addit. Manuf., 22(2018), p. 479.
|
| [2] |
S.C. Luo, W.P. Huang, H.H. Yang, J.J. Yang, Z.M. Wang, and X.Y. Zeng, Microstructural evolution and corrosion behaviors of Inconel 718 alloy produced by selective laser melting following different heat treatments, Addit. Manuf., 30(2019), art. No. 100875.
|
| [3] |
H.Z. Deng, L. Wang, Y. Liu, X. Song, F.Q. Meng and S. Huang, Evolution behavior of γ″ phase of IN718 superalloy in temperature/stress coupled field, Int. J. Miner., Metall. Mater., 28(2021), No. 12, p. 1949. DOI: 10.1007/s12613-021-2317-z
|
| [4] |
E. Hosseini and V.A. Popovich, A review of mechanical properties of additively manufactured Inconel 718, Addit. Manuf., 30(2019), art. No. 100877.
|
| [5] |
H.Y. Chen, D.D. Gu, Q. Ge, et al., Role of laser scan strategies in defect control, microstructural evolution and mechanical properties of steel matrix composites prepared by laser additive manufacturing, Int. J. Miner., Metall. Mater., 28(2021), No. 3, p. 462. DOI: 10.1007/s12613-020-2133-x
|
| [6] |
N. Li, S. Huang, G.D. Zhang, R.Y. Qin, et al., Progress in additive manufacturing on new materials: A review, J. Mater. Sci. Technol., 35(2019), No. 2, p. 242. DOI: 10.1016/j.jmst.2018.09.002
|
| [7] |
K. Gruber, W. Stopyra, K. Kobiela, B. Madejski, M. Malicki, and T. Kurzynowski, Mechanical properties of Inconel 718 additively manufactured by laser powder bed fusion after industrial high-temperature heat treatment, J. Manuf. Process., 73(2022), p. 642. DOI: 10.1016/j.jmapro.2021.11.053
|
| [8] |
M. Zhang, B. Zhang, Y. Wen and X. Qu, Research progress on selective laser melting processing for nickel-based superalloy, Int. J. Miner., Metall. Mater., 29(2022), No. 3, p. 369. DOI: 10.1007/s12613-021-2331-1
|
| [9] |
F.Y. Lu, H.Y. Wan, X. Ren, L.M. Huang, H.L. Liu, and X. Yi, Mechanical and microstructural characterization of additive manufactured Inconel 718 alloy by selective laser melting and laser metal deposition, J. Iron Steel Res. Int., 29(2022), No. 8, p. 1322. DOI: 10.1007/s42243-022-00755-x
|
| [10] |
A. Zinoviev, O. Zinovieva, V. Ploshikhin, V. Romanova, and R. Balokhonov, Evolution of grain structure during laser additive manufacturing. Simulation by a cellular automata method, Mater. Des., 106(2016), p. 321. DOI: 10.1016/j.matdes.2016.05.125
|
| [11] |
S.Y. Liu, H.Q. Li, C.X. Qin, R. Zong, and X.Y. Fang, The effect of energy density on texture and mechanical anisotropy in selective laser melted Inconel 718, Mater. Des., 191(2020), art. No. 108642. DOI: 10.1016/j.matdes.2020.108642
|
| [12] |
J.Y. Shao, G. Yu, X.L. He, S.X. Li, R. Chen, and Y. Zhao, Grain size evolution under different cooling rate in laser additive manufacturing of superalloy, Opt. Laser Technol., 119(2019), art. No. 105662. DOI: 10.1016/j.optlastec.2019.105662
|
| [13] |
H.Y. Wan, Z.J. Zhou, C.P. Li, G.F. Chen, and G.P. Zhang, Effect of scanning strategy on grain structure and crystallographic texture of Inconel 718 processed by selective laser melting, J. Mater. Sci. Technol., 34(2018), No. 10, p. 1799. DOI: 10.1016/j.jmst.2018.02.002
|
| [14] |
O. Gokcekaya, T. Ishimoto, S. Hibino, J. Yasutomi, T. Narushima, and T. Nakano, Unique crystallographic texture formation in Inconel 718 by laser powder bed fusion and its effect on mechanical anisotropy, Acta Mater., 212(2021), art. No. 116876. DOI: 10.1016/j.actamat.2021.116876
|
| [15] |
B. Stegman, A.Y. Shang, L. Hoppenrath, et al., Volumetric energy density impact on mechanical properties of additively manufactured 718 Ni alloy, Mater. Sci. Eng. A, 854(2022), art. No. 143699. DOI: 10.1016/j.msea.2022.143699
|
| [16] |
Y.C. Wang, J. Shi, and Y. Liu, Competitive grain growth and dendrite morphology evolution in selective laser melting of Inconel 718 superalloy, J. Cryst. Growth, 521(2019), p. 15. DOI: 10.1016/j.jcrysgro.2019.05.027
|
| [17] |
R. Acharya, J.A. Sharon, and A. Staroselsky, Prediction of microstructure in laser powder bed fusion process, Acta Mater., 124(2017), p. 360. DOI: 10.1016/j.actamat.2016.11.018
|
| [18] |
M. Yang, L. Wang, and W.T. Yan, Phase-field modeling of grain evolution in additive manufacturing with addition of reinforcing particles, Addit. Manuf., 47(2021), art. No. 102286.
|
| [19] |
W.J. Xiao, S.M. Li, C.S. Wang, et al., Multi-scale simulation of dendrite growth for direct energy deposition of nickel-based superalloys, Mater. Des., 164(2019), art. No. 107553. DOI: 10.1016/j.matdes.2018.107553
|
| [20] |
C. Kumara, A. Segerstark, F.B. Hanning, et al., Microstructure modelling of laser metal powder directed energy deposition of alloy 718, Addit. Manuf., 25(2019), p. 357.
|
| [21] |
C. Kumara, A.R. Balachandramurthi, S. Goel, F.B. Hanning, and J. Moverare, Toward a better understanding of phase transformations in additive manufacturing of Alloy 718, Materialia, 13(2020), art. No. 100862. DOI: 10.1016/j.mtla.2020.100862
|
| [22] |
J.P.M. Cheloni, E.B. Fonseca, A.H.G. Gabriel, and É.S.N. Lopes, The transient temperature field and microstructural evolution of additively manufactured AISI H13 steel supported by finite element analysis, J. Mater. Res. Technol., 19(2022), p. 4583. DOI: 10.1016/j.jmrt.2022.06.143
|
| [23] |
J. Goldak, A. Chakravarti, and M. Bibby, A new finite element model for welding heat sources, Metall. Trans. B, 15(1984), No. 2, p. 299. DOI: 10.1007/BF02667333
|
| [24] |
Y.H. Cheng, Numerical Simulation and Experimental Research of Selective Laser Melting on Nickel Based Alloy Powder GH4169 [Dissertation], North University of China, Taiyuan, 2016.
|
| [25] |
I. Steinbach, F. Pezzolla, B. Nestler, et al., A phase field concept for multiphase systems, Phys. D, 94(1996), No. 3, p. 135. DOI: 10.1016/0167-2789(95)00298-7
|
| [26] |
J. Eiken, B. Böttger, and I. Steinbach, Multiphase-field approach for multicomponent alloys with extrapolation scheme for numerical application, Phys. Rev. E, 73(2006), No. 6Pt2, art. No. 066122.
|
| [27] |
I. Steinbach, Phase-field models in materials science, Modell. Simul. Mater. Sci. Eng., 17(2009), No. 7, art. No. 073001. DOI: 10.1088/0965-0393/17/7/073001
|
| [28] |
Q. Peng, Study on Microstructure and Properties of Nickel-based Superalloy by Selective Laser Melting [Dissertation], University of Science and Technology Beijing, Beijing, 2020.
|
| [29] |
M. Zheng, L. Wei, J. Chen, et al., On the role of energy input in the surface morphology and microstructure during selective laser melting of Inconel 718 alloy, J. Mater. Res. Technol., 11(2021), p. 392. DOI: 10.1016/j.jmrt.2021.01.024
|
| [30] |
M. Balbaa, S. Mekhiel, M. Elbestawi, and J. McIsaac, On selective laser melting of Inconel 718: Densification, surface roughness, and residual stresses, Mater. Des., 193(2020), art. No. 108818. DOI: 10.1016/j.matdes.2020.108818
|
| [31] |
W.Q. Wang, S.Y. Wang, X.G. Zhang, F. Chen, Y.X. Xu, and Y.T. Tian, Process parameter optimization for selective laser melting of Inconel 718 superalloy and the effects of subsequent heat treatment on the microstructural evolution and mechanical properties, J. Manuf. Process., 64(2021), p. 530. DOI: 10.1016/j.jmapro.2021.02.004
|
| [32] |
H.Y. Wang, L. Wang, R. Cui, B.B. Wang, L.S. Luo, and Y.Q. Su, Differences in microstructure and nano-hardness of selective laser melted Inconel 718 single tracks under various melting modes of molten pool, J. Mater. Res. Technol., 9(2020), No. 5, p. 10401. DOI: 10.1016/j.jmrt.2020.07.029
|
| [1] | Zhao-zhen Cai, Miao-yong Zhu. Non-uniform heat transfer behavior during shell solidification in a wide and thick slab continuous casting mold [J]. International Journal of Minerals, Metallurgy and Materials, 2014, 21(3): 240-250. DOI: 10.1007/s12613-014-0901-1 |
| [2] | Guo-fa Sui, Jin-shan Li, Hong-wei Li, Feng Sun, Tie-bang Zhang, Heng-zhi Fu. Investigation on the explosive welding mechanism of corrosion-resisting aluminum and stainless steel tubes through finite element simulation and experiments [J]. International Journal of Minerals, Metallurgy and Materials, 2012, 19(2): 151-158. DOI: 10.1007/s12613-012-0531-4 |
| [3] | Jian-xun Fu, Jing-she Li, Hui Zhang. Effects of the width and thickness of slabs on broadening in continuous casting [J]. International Journal of Minerals, Metallurgy and Materials, 2010, 17(6): 723-728. DOI: 10.1007/s12613-010-0380-y |
| [4] | Hejie Li, Jingtao Han, Huachun Pi, Zhengyi Jiang, Dongbin Wei, A. Kiet Tieu. Modeling texture development during cold rolling of IF steel by crystal plasticity finite element method [J]. International Journal of Minerals, Metallurgy and Materials, 2008, 15(6): 696-701. DOI: 10.1016/S1005-8850(08)60273-X |
| [5] | Qingjun Chen, Yonglin Kang, Hao Yu, Chunmei Wang, Chengxiang Li. Research on microstructural evolution and dynamic recrystailization behavior of JB800 bainitic steel by FEM [J]. International Journal of Minerals, Metallurgy and Materials, 2008, 15(3): 250-254. DOI: 10.1016/S1005-8850(08)60047-X |
| [6] | Huachun Pi, Jingtao Han, Chuanguo Zhang, A. Kiet Tieu, Zhengyi Jiang. Modeling uniaxial tensile deformation of polycrystalline Al using CPFEM [J]. International Journal of Minerals, Metallurgy and Materials, 2008, 15(1): 43-47. DOI: 10.1016/S1005-8850(08)60009-2 |
| [7] | Xianhua Cheng, Qianqian Shangguan. Investigations of high-frequency induction hardening process for piston rod of shock absorber [J]. International Journal of Minerals, Metallurgy and Materials, 2005, 12(1): 85-89. |
| [8] | Yanli Wang, Junpin Lin, Zhi Lin, Xinfa Cui, Guoliang Chen. Deformation characterization of conductive particles in anisotropically conductive adhesive simulated by FEM [J]. International Journal of Minerals, Metallurgy and Materials, 2005, 12(1): 48-53. |
| [9] | Zhanghua Chen, Jiajian Jin, Jiumei Xiao. Numerical simulation of strain localization and damage evolution in large plastic deformation using mixed finite element method [J]. International Journal of Minerals, Metallurgy and Materials, 2004, 11(3): 273-277. |
| [10] | Zhiyuan Shi, Wanjun Wang, Zhigang Kan, Jinglin Wen, Xinhua Wang. Simulation of Temperature and Flow Field by Three Dimension Finite Element Method for Castex Process of AS Wire [J]. International Journal of Minerals, Metallurgy and Materials, 1999, 6(4): 262-267. |
| 1. | Shakti Kumar, Indrani Mukherjee, K. Pavan Chandu, et al. Multi‐Phase Field Model of Laser Powder Bed Fusion Based Additive Manufacturing of IN718 Superalloy. Advanced Theory and Simulations, 2025. DOI:10.1002/adts.202401031 |
| 2. | Yeon-Joo Lee, Pil-Ryung Cha, Hyoung-Seop Kim, et al. Review of “Integrated Computer-Aided Process Engineering Session in the 17th International Symposium on Novel and Nano Materials (ISNNM, 14–18 November 2022)”. MATERIALS TRANSACTIONS, 2025, 66(1): 144. DOI:10.2320/matertrans.MT-M2024133 |
| 3. | H. C. Madhu, V. Sampath. Influence of Laser Power on Microstructure and Mechanical Behavior of Laser Powder Bed Fusion IN718 after Heat Treatment. Journal of Materials Engineering and Performance, 2024. DOI:10.1007/s11665-024-09916-3 |
| 4. | Xiawei Yang, Tingxi Meng, Qiang Chu, et al. A review of linear friction welding of Ni-based superalloys. International Journal of Minerals, Metallurgy and Materials, 2024, 31(6): 1382. DOI:10.1007/s12613-023-2782-7 |
| 5. | Lingzhi Wu, Cong Zhang, Dil Faraz Khan, et al. Unveiling the cellular microstructure–property relations in martensitic stainless steel via laser powder bed fusion. International Journal of Minerals, Metallurgy and Materials, 2024, 31(11): 2476. DOI:10.1007/s12613-024-2947-z |
| 6. | Xiaolian Liu, Yuanzhuo Liu, Wenchong Guo, et al. Influence of laser power on microstructure and high-temperature tribological properties of CoAlTiWTa RHEA coating on Inconel 718 superalloy. Surface and Coatings Technology, 2024, 479: 130573. DOI:10.1016/j.surfcoat.2024.130573 |
| 7. | Kaiyang Wang, Shaojie Lv, Honghui Wu, et al. Recent research progress on the phase-field model of microstructural evolution during metal solidification. International Journal of Minerals, Metallurgy and Materials, 2023, 30(11): 2095. DOI:10.1007/s12613-023-2710-x |
Figures(11) / Tables(5)
Copyright©2019 Editorial Office of International Journal of Minerals, Metallurgy and Materials 京ICP备13030111号
Supported by:
Beijing Renhe Information Technology Co., Ltd.
Email:
info@rhhz.net
| Ni | Cr | Fe | Mo | Nb | Ti | Al |
| 54.7 | 19.0 | 16.7 | 3.2 | 5.0 | 1.0 | 0.5 |
DownLoad:
CSV
| Laser power / W | Scanning speed / (mm·s–1) | Layer rotation angle / (°) |
| 150, 200, 250, 275, 300, 325, 350 | 700, 800, 900, 960, 1000, 1200, 1400 | 0, 15, 30, 45, 67,75, 90 |
DownLoad:
CSV
| Ni | Cr | Fe | Mo | Nb | Ti | Al | Co | Si | Cu | Mn | C |
| 54.66 | 18.98 | Margin | 3.15 | 5.0 | 0.98 | 0.48 | 0.12 | 0.07 | 0.034 | 0.027 | 0.024 |
DownLoad:
CSV
| Point | Laser power / W | Scanning speed / (mm·s–1) | Energy density / (J·mm–3) | Interlayer rotation angle / (°) |
| A1 | 280 | 1160 | 61.42 | 67 |
| A2 | 280 | 1175 | 59.57 | 52 |
| A3 | 200 | 880 | 56.82 | 15 |
DownLoad:
CSV
Search
