WELD_Graph

직접 에너지 증착을 이용한 Ni625 합금의 가공 범위

Yusufu Ekubaru, Takuya Nakabayashi, Tomoharu Fujiwara, Behrang Poorganji

Abstract

Herein, a process window is developed for Ni625 alloy fabricated using a Nikon Lasermeister laser powder direct energy deposition (LP-DED) unit. The process map illustrates the relationship between the laser power, scan speed, and effective energy density, established by examining the correlation between the microstructure and mechanical properties. All samples exhibit a bimodal microstructure comprising equiaxed and columnar dendrite grains, and the dendrite arm spacing decreases with increasing scan speed. The tensile behavior of each sample demonstrates minimal variation, and the values are comparable to those reported previously. The ultimate tensile and yield strengths range from 1008 ± 2 to 941 ± 9 and 682 ± 11 to 640 ± 7 MPa, respectively. This study highlights the remarkable manufacturability of Ni625 alloy for additive manufacturing across diverse parameter sets, demonstrating that a single ideal process set does not exist for each material and machine. Instead, multiple “recipes” may be employed to achieve similar outcomes.

1. Introduction

Metal additive manufacturing (AM) is an excellent technology for part fabrication, offering distinct advantages over conventional manufacturing methods. With significant cost and lead-time reductions and the capability to develop complex geometrical features,[1–3] metal AM has rapidly garnered interest from key industries such as aerospace, automotive, military, and biomedical sectors.[3–5] Metal AM entails various techniques, including material jetting, sheet lamination, laser powder bed fusion, binder jetting, and direct energy deposition.

Laser powder direct energy deposition (LP-DED) presents unique advantages over other AM processes, including alloy design, repair capabilities, surface modifications, and the synthesis of large-scale components with adequate dimensional accuracy.[6] These capabilities have been increasingly demonstrated and recognized in various fields, particularly in the aerospace industry.[4] The in-situ alloying of elemental powders offers an effective alternative to the use of pre-alloyed powders, which are cost- and time-intensive to produce using traditional atomization methods. By mixing pure elemental powders of Ni, Cr, Mo, Nb, and Fe, Wang et al.[7] demonstrated the high-quality fabrication of Ni625 alloy components using LP-DED and in situ alloying. Wilson et al.[8, 9] repaired defective voids in turbine blades, illustrating the effectiveness of LP-DED in repair. These studies highlight the adaptability of DED to a wide range of defective parts, as well as its capabilities in repair and maintenance. Balla et al.[10] applied a tantalum coating onto titanium using LP-DED, a notable achievement considering the extremely high melting point (>3000 °C) of Ta, which poses challenges for traditional melt-cast methods. Ta-coated Ti exhibits favorable interactions with bone cells, indicating promising biocompatibility. Gradl et al.[1, 2, 11] utilized LP-DED to manufacture a large-scale rocket nozzle for aerospace applications. The growing recognition of LP-DED is reflected in the significant increase in the number of patents and scientific publications dedicated to this technology, highlighting its importance in academia and industry.[5, 12]

Furthermore, the anticipation of an expanding market for AM has spurred intense competition among AM machine manufacturers, resulting in the development of various AM systems.[4] In this context, the Nikkon Advanced Manufacturing Department in Japan developed an LP-DED system named Lasermeister. Extensive empirical testing has been conducted on this machine with common AM materials, including Fe-, Ni-, and Ti-based alloys. Herein, we present our latest research findings, particularly focusing on the Ni625 alloy, also referred to as Alloy 625 or Inconel 625.

The Ni625 alloy has been utilized in various industries, including petrochemical, aerospace, chemical, marine, and nuclear sectors, due to its excellent strength and high corrosion and fatigue resistance.[12, 13] Moreover, its remarkable weldability has attracted considerable attention in AM, where it has been successfully produced using various process parameters in LP-DED, including laser power (P) (220–1500 W) and scan speed (V) (8.3–33.3 mm s−1), with the corresponding effective energy density (ED) ranging from 14 to 66 J mm−2.[7, 13–20]

The solidification microstructure of AM-produced Ni625 alloy is complex, featuring fine dendrites, micro-segregated elements, and various solidification phases.[21] The nickel-based superalloy, primarily strengthened by the solid hardening effects of refractory elements including niobium and molybdenum within a nickel–chromium matrix exhibits a face-centered cubic (FCC) structure.[14] These alloys are sensitive to the precipitation of strengthening intermetallic phases, including stable ordered FCC (L12)γ′-Ni3Al; metastable body-centered tetragonal γ″-Ni3Nb; stable orthorhombic δ-Ni3Nb; carbides (MC, M6C); and intergranular brittle Laves phases ((Nb, Mo)(NiCrFe)2) in the interdendritic region.[1, 12–14] The formation of these phases, particularly the Laves phases, consumes significant amounts of Nb and Mo, thereby reducing their content in the matrix, which diminishes solid solution and precipitation strengthening effects.[19] Further, the Laves phase induces crack nucleation and propagation, significantly deteriorating creep rupture properties and ductility.[19] Consequently, manufacturing components with reduced elemental segregation and fewer Laves phases has become critical.[22–24]

The mechanical properties of materials are primarily influenced by factors such as porosity, grain size, the behavior of precipitates, and dendrite spacing.[25] Generally, the mechanical properties can be improved by reducing their size, which essentially means creating a finer microstructure. Reducing porosity can enhance the material’s strength and durability as fewer pores mean less space for cracks to initiate. Smaller grain sizes often lead to increased hardness and strength due to the Hall–Petch relationship. Controlling the behavior of precipitates, such as reducing their size, can increase the material’s strength as smaller precipitates more effectively hinder dislocation movement.[25, 26] Lastly, smaller dendrite spacing can contribute to a more homogeneous microstructure, reducing segregation and enhancing various mechanical properties.[25, 26] One fundamental approach to achieving a finer microstructure is to increase the cooling rate, and it can be accomplished by using a smaller P, a higher V, or a combination of both.[27]

Based on this background, this study aimed to 1) develop the process windows for Ni625 alloy using the Lasermeister system and 2) establish a process window that expresses the relationship between P, V, and ED based on a series of simulations and experiments focusing on microstructural properties and mechanical performance.

This research demonstrated for the first time that using lower P values and smaller hatch spacings can significantly enhance the strength of Ni625 alloys by promoting substantial microstructure miniaturization. Additionally, DED process “recipes” for Ni625 in the lower P region were developed. These results are expected to significantly contribute to the DED fabrication of components such as precise, large, thin-walled structures that are vulnerable to thermal deformation, as well as the automation of gas turbine blade repairs, among other applications.

Advanced Engineering Materials

Research Article

Open Access

Processing Windows of Ni625 Alloy Fabricated Using Direct Energy Deposition

Yusufu Ekubaru, Takuya Nakabayashi, Tomoharu Fujiwara, Behrang Poorganji

First published: 21 June 2024

https://doi.org/10.1002/adem.202400962

Sections

PDF

Tools

Share

Abstract

Herein, a process window is developed for Ni625 alloy fabricated using a Nikon Lasermeister laser powder direct energy deposition (LP-DED) unit. The process map illustrates the relationship between the laser power, scan speed, and effective energy density, established by examining the correlation between the microstructure and mechanical properties. All samples exhibit a bimodal microstructure comprising equiaxed and columnar dendrite grains, and the dendrite arm spacing decreases with increasing scan speed. The tensile behavior of each sample demonstrates minimal variation, and the values are comparable to those reported previously. The ultimate tensile and yield strengths range from 1008 ± 2 to 941 ± 9 and 682 ± 11 to 640 ± 7 MPa, respectively. This study highlights the remarkable manufacturability of Ni625 alloy for additive manufacturing across diverse parameter sets, demonstrating that a single ideal process set does not exist for each material and machine. Instead, multiple “recipes” may be employed to achieve similar outcomes.

1 Introduction

Metal additive manufacturing (AM) is an excellent technology for part fabrication, offering distinct advantages over conventional manufacturing methods. With significant cost and lead-time reductions and the capability to develop complex geometrical features,[1–3] metal AM has rapidly garnered interest from key industries such as aerospace, automotive, military, and biomedical sectors.[3–5] Metal AM entails various techniques, including material jetting, sheet lamination, laser powder bed fusion, binder jetting, and direct energy deposition.

Laser powder direct energy deposition (LP-DED) presents unique advantages over other AM processes, including alloy design, repair capabilities, surface modifications, and the synthesis of large-scale components with adequate dimensional accuracy.[6] These capabilities have been increasingly demonstrated and recognized in various fields, particularly in the aerospace industry.[4] The in-situ alloying of elemental powders offers an effective alternative to the use of pre-alloyed powders, which are cost- and time-intensive to produce using traditional atomization methods. By mixing pure elemental powders of Ni, Cr, Mo, Nb, and Fe, Wang et al.[7] demonstrated the high-quality fabrication of Ni625 alloy components using LP-DED and in situ alloying. Wilson et al.[8, 9] repaired defective voids in turbine blades, illustrating the effectiveness of LP-DED in repair. These studies highlight the adaptability of DED to a wide range of defective parts, as well as its capabilities in repair and maintenance. Balla et al.[10] applied a tantalum coating onto titanium using LP-DED, a notable achievement considering the extremely high melting point (>3000 °C) of Ta, which poses challenges for traditional melt-cast methods. Ta-coated Ti exhibits favorable interactions with bone cells, indicating promising biocompatibility. Gradl et al.[1, 2, 11] utilized LP-DED to manufacture a large-scale rocket nozzle for aerospace applications. The growing recognition of LP-DED is reflected in the significant increase in the number of patents and scientific publications dedicated to this technology, highlighting its importance in academia and industry.[5, 12]

Furthermore, the anticipation of an expanding market for AM has spurred intense competition among AM machine manufacturers, resulting in the development of various AM systems.[4] In this context, the Nikkon Advanced Manufacturing Department in Japan developed an LP-DED system named Lasermeister. Extensive empirical testing has been conducted on this machine with common AM materials, including Fe-, Ni-, and Ti-based alloys. Herein, we present our latest research findings, particularly focusing on the Ni625 alloy, also referred to as Alloy 625 or Inconel 625.

The Ni625 alloy has been utilized in various industries, including petrochemical, aerospace, chemical, marine, and nuclear sectors, due to its excellent strength and high corrosion and fatigue resistance.[12, 13] Moreover, its remarkable weldability has attracted considerable attention in AM, where it has been successfully produced using various process parameters in LP-DED, including laser power (P) (220–1500 W) and scan speed (V) (8.3–33.3 mm s−1), with the corresponding effective energy density (ED) ranging from 14 to 66 J mm−2.[7, 13–20]

The solidification microstructure of AM-produced Ni625 alloy is complex, featuring fine dendrites, micro-segregated elements, and various solidification phases.[21] The nickel-based superalloy, primarily strengthened by the solid hardening effects of refractory elements including niobium and molybdenum within a nickel–chromium matrix exhibits a face-centered cubic (FCC) structure.[14] These alloys are sensitive to the precipitation of strengthening intermetallic phases, including stable ordered FCC (L12)γ′-Ni3Al; metastable body-centered tetragonal γ″-Ni3Nb; stable orthorhombic δ-Ni3Nb; carbides (MC, M6C); and intergranular brittle Laves phases ((Nb, Mo)(NiCrFe)2) in the interdendritic region.[1, 12–14] The formation of these phases, particularly the Laves phases, consumes significant amounts of Nb and Mo, thereby reducing their content in the matrix, which diminishes solid solution and precipitation strengthening effects.[19] Further, the Laves phase induces crack nucleation and propagation, significantly deteriorating creep rupture properties and ductility.[19] Consequently, manufacturing components with reduced elemental segregation and fewer Laves phases has become critical.[22–24]

The mechanical properties of materials are primarily influenced by factors such as porosity, grain size, the behavior of precipitates, and dendrite spacing.[25] Generally, the mechanical properties can be improved by reducing their size, which essentially means creating a finer microstructure. Reducing porosity can enhance the material’s strength and durability as fewer pores mean less space for cracks to initiate. Smaller grain sizes often lead to increased hardness and strength due to the Hall–Petch relationship. Controlling the behavior of precipitates, such as reducing their size, can increase the material’s strength as smaller precipitates more effectively hinder dislocation movement.[25, 26] Lastly, smaller dendrite spacing can contribute to a more homogeneous microstructure, reducing segregation and enhancing various mechanical properties.[25, 26] One fundamental approach to achieving a finer microstructure is to increase the cooling rate, and it can be accomplished by using a smaller P, a higher V, or a combination of both.[27]

Based on this background, this study aimed to 1) develop the process windows for Ni625 alloy using the Lasermeister system and 2) establish a process window that expresses the relationship between P, V, and ED based on a series of simulations and experiments focusing on microstructural properties and mechanical performance.

This research demonstrated for the first time that using lower P values and smaller hatch spacings can significantly enhance the strength of Ni625 alloys by promoting substantial microstructure miniaturization. Additionally, DED process “recipes” for Ni625 in the lower P region were developed. These results are expected to significantly contribute to the DED fabrication of components such as precise, large, thin-walled structures that are vulnerable to thermal deformation, as well as the automation of gas turbine blade repairs, among other applications.

2. Experimental Section

2.1 Ni625 Alloy Fabrication

Ni625 alloy powders were procured from Carpenter Additive Inc.; and their compositions and morphologies are summarized in Table 1 and Figure 1, respectively. An LP-DED unit (Nikon Lasermeister 100A, Japan) with a 915 nm 200 W laser diode module and a beam diameter (d) of 0.5 mm was utilized to fabricate the Ni625 alloy samples (Figure 1c). Two samples, namely, a 10 × 10 × 10 mm cube and a 10 × 10 × 55 mm rectangle, were fabricated along the x-, y-, and z-axes on a SUS304 substrate via the XY scanning strategy. Cubic samples were used for microstructural analysis, whereas rectangular samples were employed for tensile property testing (Figure 1d).[7, 20] The parameter values used for the experiment are listed in Table 3, where the laser hatch spacing was maintained at 0.2 mm.

PowderNiCrMoNb + TaFeAlTiCMn
Inconel625Bal.20–238–103.15–4.15≤5≤0.4≤0.4≤0.03≤0.01
Table 1. Chemical composition (wt%) of Ni625 alloy.
Figure 1 a) Morphology and b) powder size distribution of the Ni625 alloy powders used in laser powder direct energy deposition (LP-DED). c) Schematic of LP-DED and d) dimensions of the tensile test sample.

2.2 Microstructure Characterization and Mechanical Properties

Samples were cut from the substrate via electrical discharge machining to analyze their microstructures and mechanical properties. The YZ cross sections were first mechanically polished using emery paper up to a 4000 grade and subsequently chemically polished with colloidal silica to achieve mirror-polished sections for microstructural examination.

Optical microscopy (KEYENCE VHX8000) and scanning electron microscopy (SEM; Hitachi SU1500) were conducted to examine the microstructures. The bulk samples fabricated by the LP-DED Lasermeister were characterized via X-ray diffraction (XRD; Rigaku RINT2500) with Cu-K radiation at room temperature (RT). Crystallographic texture and elemental segregation were investigated using electron backscattered diffraction (EBSD) and energy-dispersive X-ray spectroscopy (EDS), respectively, with a scanning electron microscope (JEOL JSM-7900F). A tensile test (Minebea TGI-50KN) was conducted at RT, where the loading axis was parallel to the build direction (BD). The test was conducted thrice for each sample, and the results were averaged.

2.3 Simulations

The formation mechanism of the microstructure induced by LP-DED was explored through simulations focusing on thermal behavior and solidification characteristics. The thermal behavior calculations provided insights into the temperature distribution and the shape and size of the melt pool. Conversely, analyzing the solidification characteristics aided in understanding the development of grains, which could manifest as either equiaxed dendrites (ED) or columnar dendrites (CD).

These simulations were performed using the commercial software FLOW-3D v12.0 for a region measuring 10 × 7 × 3 mm in the X, Y, and Z directions. The region was discretized into a structural Eulerian mesh with a size of 0.025 mm.

2.3.1 Heat Source Model

where P0 is the laser power (100/120/160 W), r is the distance from the beam center, r is the laser radius (0.25 mm), rb is the effective laser radius (0.1 mm), hc is the heat transfer coefficient (9.5 W m−2 K),[28] T is the temperature, and T0 is the ambient temperature (298 K).

2.3.2 Powder Model

We employed the Lagrangian particle tracking method to model the powder particles. Particles entering the melt pool transformed into liquid cells upon surpassing the melting point. The amount of powder injected was calculated from the predetermined powder utilization efficiency. The powder was injected at a constant velocity from the vertical direction of the melt pool to ensure the melting of all particles.

2.3.3 Melt Pool (MP) Flow Governing Equations

The governing equations, which include mass, momentum, and energy conservation, are expressed in Equation (2), (3), and (4), respectively.

where ρ is density, t is time, v is flow velocity, RSOR is the amount of mass source due to powder particles, p is pressure, μ is viscosity, vp is particle velocity, Cv is specific heat, fs is the solidus rate, ISOR is the discharge of energy, and L is latent heat. The thermophysical parameters were calculated using the thermodynamic database of JmatPro (Sente Software) considering their temperature dependencies (Table 2).

Temperature intervalsThermal conductivitySpecific heatDensityViscositySurface tensionLatent heat of fusion
T [K]κ [W (m K)−1]Cv [J (kg K)−1]ρ [kg m−3]μ [kg (m s)−1]σ [N m−1]L [kJ kg−1]
29810.84068474––210
60015.94568373––
90020.95048253––
120025.85598117––
150030.171379311.39 × 10−21.84
180031.473774990.62 × 10−21.74
210035.874572350.38 × 10−21.62
240040.274869520.26 × 10−21.52
Table 2. Thermophysical properties of Ni625 calculated using JmatPro.

2.3.4 Solidification Parameter

The temperature gradient G and solidification velocity R represent spatial temperature variations and are expressed as:

where ε is the cooling rate, Ts is the solidus line temperature (1398 K), TL is the liquidus line temperature (1613 K), tS is the time below the solidus line temperature, tL is the time below the liquidus line temperature, and ∇ is the differential operator.

3 Results

3.1 Simulated Data

The aspect ratio (D/W), indicating the depth (D) to width (W) ratio of the MP, was assessed in both the experimental and simulated scenarios to verify the simulation model. Figure 2 displays the results of the single-track experiments and simulations at V values of 5 and 10 mm s−1, with constant P and powder feeding rate (Q) values of 120 W and 3 g min−1, respectively. The experimental dimensions of the MP were measured from the optical images, whereas the simulated sizes of the MP were determined by identifying a black solidus line on the temperature contour map. The aspect ratios decreased with increasing V, and the experimental aspect ratios were slightly higher than the simulated ones, with differences of <10%. It is considered that one possible reason for this difference is the thermal boundary conditions of the substrate in the simulation. Hence, this model was employed for additional simulations to generate a process map for the Ni625 alloy.

Figure 2 Comparison of the experimental and simulated MP: a,b) experimental optical images, a’,b’) simulated temperature contours, and c) aspect ratio. Scale bars: 200 μm.

Various conditions were simulated to assess fabrication feasibility using these process parameters. Figure3 illustrates the simulated temperature contour plots and the maximum temperature of the MPs under nine different conditions, accompanied by their respective sizes. As shown in Figure 3b,c, with an increase in P from 100 to 160 W (while V is constant at 5 mm s−1), the maximum temperature increases from 2335 to 2725 K, and the width of the MPs increases from 540 to 780 μm; by contrast, increasing V when P is constant causes both the maximum temperature and the width and depth of the MPs to remain almost constant. The highest temperatures and dimensions of the MPs indicated a significant dependence on P but less dependence on V. Consequently, MPs were formed under all conditions, and the maximum temperature exceeded the melting point of the Ni625 alloy at 1623 K,[29] which allowed us to proceed with the experiments.

Figure 3 Simulated MP of a) temperature contour plot, b) maximum temperature, and c) dimensions at varying process parameters.

3.2 Microstructural Analysis

The fabricated state, porosity, and cracks of the samples produced under the nine simulated conditions were investigated via cross-sectional image analysis using an optical microscope. All samples, except S7, were successfully manufactured, as shown in Figure 4a; however, S7 could not be completed because the powder adhered to the nozzle owing to the highest energy density input. The optical density shown in Figure 4b was measured from optical images of the polished surfaces of the samples. Five images were taken from different locations on the polished surface of each sample at 200× magnification. The optical density of these images was then measured using ImageJ software, and the average was calculated. As shown in Figure 4b, most samples, excluding S2 and S3, exhibited a dense structure without any visible cracks; this resulted in a satisfactory industrial density of over 99.5%.[3, 30, 31] However, samples S2 and S3 showed noticeably lower density values with irregularly shaped pores caused by the lack of fusion owing to the lower energy density input. It can be generally observed that densification increases with increasing P and decreases with increasing V. This behavior is more significant in samples S1 to S3 at 100 W, while it is less pronounced in samples S4 to S9 at 120 and 160 W. This suggests that at lower P settings, the impact of V on densification is more pronounced, whereas at higher P settings, the effect of V becomes less significant. Consequently, optimizing P and V parameters is crucial for achieving desired densification levels in different samples.

Figure 4 a) Appearance of the LP-DED fabricated samples and b) optical density.

The microstructure of AM materials can be explained by the MP microstructure using Hunt’s columnar-equiaxed transition criteria.[27, 32, 33] As shown in Figure 5a, MPs typically exhibit a bimodal microstructure comprising two types of grains: ED at the top with no preferential crystallographic orientation and CD at the bottom that show a preference for growing from the bottom part to the center along the direction of the thermal gradient.[21, 33-35] This is attributed to the higher G/R ratio at the bottom part of the MP and the lower G/R ratio at the top, as illustrated in Figure 5b, where G/R is the grain morphology factor determining either ED or CD, and G × R is the cooling rate that determines the size of the grain. Typically, the extremely high G and G × R values in the AM process foster directional solidification, and enhance the textures of the microstructures of alloys.[36, 37]

Figure 5 a) Schematic of the MP microstructure and b) columnar-equiaxed transition criteria, adapted from ref. 27 with permission.[27, 32, 33]

The dendrite microstructural features, including the PDAS size and shape of the grains of the samples, were characterized by observing the SEM images of the aqua regia-etched YZ cross-section. PDAS is one of the factors in influencing mechanical properties and it was known that smaller PDAS increases various mechanical properties.[25, 26] As shown in Figure 6a, among the samples, S7 yielded the highest PDAS with a value of 3.7 ± 0.1 μm, while S3 yielded the lowest PDAS with a value of 1.7 ± 0.3 μm; consequently, the PDAS increased as the P increased and V decreased. In contrast, as shown in Figure 6b, all samples exhibited a bimodal grain microstructure consisting of CD and ED regions. Samples S7–S9, fabricated with the highest P of 160 W, exhibited a predominance of CD, while samples S1–S3, fabricated with the lowest P of 100 W, displayed an almost exclusive ED presence, and resulted in a trend that shifted from an ED-dominant to CD-dominant microstructure with increasing P and decreasing V, respectively; namely, high P values increased the dendrite structure, which is consistent with other research.[14]

Figure 6 SEM images of the YZ plane of the samples with a) higher magnitudes containing PDAS and b) lower magnitudes containing CD and ED regions.

The elemental microsegregation of the samples was analyzed using EDS mapping. Figure 7a,b illustrate the distributions of the main elements (Ni, Cr, Fe, Nb, and Mo) in samples S7 (with the highest energy density) and S3, respectively. The Mo and Nb contents in the interdendritic regions were higher than those in the dendritic regions, as indicated by the yellow arrow. Both samples exhibited significant Mo and Nb segregation with no clear differences in their segregation behaviors. Based on the obtained results and previous reports, it can be concluded that the observed phase corresponds to the Laves phase.[7, 16, 19]

Figure 7 EDS maps of samples of a) S7 and b) S3.

XRD analysis was conducted on the polished YZ cross-section of the samples to confirm the phase states. As shown in Figure 8, all the samples exhibited peaks corresponding to the reference Ni (PDF #04-0850) in the XRD analysis. Interestingly, in sample S7, the relative intensities of the (111) and (200) peaks were similar, even though (111) has the highest-intensity peak, indicating that (100) tends to be oriented in the BD (z-direction), which is in agreement with other studies.[3, 7, 38, 39] However, all samples exhibited a minor peak shift to a lower diffraction angle compared with Ni (PDF #04-0850), implying the presence of residual stress in the samples.[3, 38]

Figure 8 XRD patterns of the LP-DED fabricated samples.

One of the key features of AM that influences the mechanical properties is the crystallographic texture,[36] which was investigated using EBSD. As shown in Figure 9a, by increasing P and decreasing V, directional grain growth occurs along the z-direction with a {100} crystallographic orientation, which is an easy growth direction for the FCC crystal structure,[3, 36] which was observed in the samples. The values of the texture strength measure, MUD, increased as P increased and V decreased; however, apart from sample S7, no distinguishable crystallographic textures were observed for the samples, and S7 exhibited the highest texture with most grains aligned in the {100} crystallographic orientation; this finding is consistent with the XRD results shown in Figure 8.

Figure 9 EBSD a) inverse pole figure maps and b) the corresponding {001} pole figures with multiples of uniform distribution (MUD) values of the YZ plane.

3.3 Tensile Properties

A tensile test was performed at RT, and the results showed trends corresponding to the features of the microstructure. As shown in the optical images in Figure 4, the porosity increased with V in the sample fabricated at the lowest P of 100 W, whereas the elongation (El) of these samples decreased, as shown in Figure 10a. However, with an increase in V, the minor decreases in the PDAS and grain size shown in Figure 6 and 9 led to a minor monotonic increase in the ultimate tensile strength (UTS) for the samples produced at P = 120 and 160 W. Consequently, the tensile properties exhibited negligible variations because fewer changes were observed in the microstructure.

Figure 10 Tensile stress–strain curves of the samples.

3.4 Process Window

A process map illustrating the relationship between P, V, ED, and the feasibility of sample fabrication was established based on the experimental data obtained in this study. Figure 11 illustrates that the pink region represents high ED, while the blue region represents low ED. Additionally, samples S2 and S3, located in the low ED area, exhibited higher porosity owing to insufficient fusion. Conversely, sample S7, situated in the high ED area, was not fully produced because of powder adhesion in the nozzle. Consequently, the approximate optimal region is indicated by a yellow line.

Figure 11 P–V process map with ED contour.

4 Discussion

4.1 Pore Formation and Mechanical Property Impact

Pores are one of the major defects that significantly affect the mechanical properties of parts; which can primarily occur owing to both high- or low-energy input, as well as the insufficient overlap of laser tracks.[40]

High-energy input during the melting process can result in the formation of an unstable MP at extremely high temperatures and severe Marangoni convection, which in turn leads to the generation of spherical pores either by trapping the protective gas (Ar) or metallic vapor.[3, 40] As shown in Figure 4b, samples S7 and S1 fabricated with a higher energy density showed spherical pores with a maximum diameter of 40 μm. These pores were primarily formed owing to the trapping of Ar gas and were unlikely attributed to metallic vapor because of the high melting points of all the main elements of the Ni625 alloy. It is known that spherical pores with diameters <130 μm have negligible detrimental effects on the mechanical properties of the material.[3] Moreover, as illustrated in Figure 10a, sample S1 displayed satisfactory tensile properties, despite the presence of spherical pores.

Low-energy input cannot completely melt the metallic powder in the previously deposited layer, thus leading to irregularly shaped lack of fusion pores, as shown in Figure 4b. Samples S2 and S3 produced with lower energy input contained irregularly shaped pores with sizes over 100 μm; these samples exhibited lower elongation tensile properties, as shown in Figure 10a.

Insufficient overlap among laser tracks can also cause a lack of fusion pores, which may be attributed to a large hatch distance and/or layer thickness.[40] However, in this study, the primary cause of the lack of fusion pores was identified as low-energy input, predominantly due to low P.

4.2 Effects of P and V on Grain Size and Morphology

P and V are the primary process parameters used to adjust the energy density to tailor the microstructure, and they significantly affect the MP solidification process parameters G and R.[27, 32, 33] Therefore, a comprehensive understanding of G and R is crucial for predicting or explaining the microstructural features observed in experimental samples, and simulations are an effective tool for their calculation.[27, 41, 42]

As shown in the solidification map in Figure 5b, G × R is the cooling rate that determines the size of the grain, whereas (G/R) is the morphology factor that determines the shapes of the grains. In this study, a maximum cooling rate of 3.5 × 104 K s−1 was achieved for sample S3, which is close to the intrinsic cooling rate of LP-DED, which ranges from 103 to 104 K s−1.[21]

At increasing P and decreasing V values, the PDAS increased while the grain shapes shifted from being predominantly ED-dominant to CD-dominant, as shown in Figure 6b. It is believed that these behaviors can be attributed to the changes in G × R and G/R, as illustrated in Figure 12.

Figure 12 Simulated a) average G × R and a’) G × R contours, and b) average G/R and b’) G/R contours.

As shown in Figure 12a, the impact of P on G × R is minor at low V values but becomes significant at high V. Therefore, the G × R values of the samples are almost the same at V = 5 mm s−1, and the PDASs of these samples do not change significantly, as shown in Figure 6a. Conversely, G × R increased as a function of V as also proven by other researchers,[27, 41, 42] and the highest and lowest G × R values were obtained for S3 and S7, respectively; accordingly, S3 and S7 respectively exhibited the lowest and highest PDAS values equal to 1.7 ± 0.3 and 3.7 ± 0.1 μm, as shown Figure 6a.

As shown in Figure 12b, G/R is less affected by P but is significantly affected by V; additionally, G/R decreases as V increases, thus suggesting that CD increases with decreasing V. Correspondingly, the directional grain growth along the z-direction with the {100} crystallographic orientation is most significant in the samples with the lowest V of 5 mm s−1, as shown in Figure 9.

As shown in Figure 12a’,b’, higher G/R and lower G × R values are observed at the bottom of the MP; in contrast, lower G/R and higher G × R were obtained at the top of the MP[27, 41, 42] and these behaviors are most significant at low V, thus indicating that the morphology of the microstructure is prone to CD. Correspondingly, the texture strength measure MUDs were higher in fabricated samples with the lowest V, as shown in Figure 9.

4.3 Verification of Tensile Properties

Although there were no dramatic differences in the tensile behavior of each sample in this study, the results were still comparable to the tensile results from other existing studies. As illustrated in Figure 13 and as listed in Table 3, the UTS and yield strengths (YS) of samples exhibited minor changes, with UTS changing from 1008 ± 2 to 941 ± 9 MPa and YS changing from 682 ± 11 to 640 ± 7 MPa. However, in the samples fabricated with the lowest P of 100 W, the elongation noticeably decreased as V increased owing to the higher porosity caused by the lack of fusion, as shown in Figure 4. Conversely, according to the reference data in Table 3, it is known that the Ni625 alloy can be fabricated using a broad range of process parameters (for example, P may change from 220 to 1500 W and V from 8.3 to 33.3 mm s−1) yielding higher tensile properties than casting.

Figure 13 Comparison of tensile properties in this study with those obtained in other research studies.
LabelPVQdE D = P/(Vd)UTSYSEl
[W][mm s−1][g min−1][mm][J mm−2][MPa][MPa][%]
S11005.02.00.540.0951 ± 7655 ± 1542 ± 2
S210010.04.00.520.01008 ± 2682 ± 1136 ± 1
S310015.04.00.513.31005 ± 7674 ± 1328 ± 4
S41205.02.00.548.0941 ± 9640 ± 742 ± 2
S512010.04.00.524.0959 ± 3666 ± 741 ± 1
S612015.04.00.516.0989 ± 4669 ± 1237 ± 1
S71605.02.00.564.0–––
S816010.04.00.532.0944 ± 4670 ± 1042 ± 1
S916015.04.00.521.3960 ± 6672 ± 940 ± 1
[7]2208.32.30.466.01020.9675.823.1
[14]33033.37.00.414.11073 ± 5723 ± 2326 ± 2
[15]50012.52.51.233.3882 ± 7480 ± 2036 ± 5
[19]150015.07.5520.0733.7500.429.4
[29]Casting485.0275.025.0
Table 3. Comparison of process parameters and tensile properties in this study with those obtained in references.

In addition, based on the literature data listed in Table 3, the UTS decreases at increasing P. A higher P not only increases the evaporation[21] of Al, Cr, Fe, and Co in the Ni625 alloy by increasing the MP temperature, but also accelerates precipitation growth owing to a lower cooling rate, thus leading to a degradation of mechanical properties. Therefore, using P values as small as possible is advantageous for the microstructure and mechanical properties of the material and machine maintenance. In this study, tensile properties similar to those reported in other research studies[7, 13, 14, 18] were obtained by using a lower P combination with a small hatch space, as shown in Figure 13. A small hatch space increases remelting, which reduces the lack of fusion[43] and increases the ED grains.[3]

This study is believed to be the first report on the optimization of the strength and ductility of Ni625 alloys using a relatively low P value, thus demonstrating that high-performance Ni625 alloys can also be fabricated with lower P.

5 Conclusions

Extensive empirical testing on the Lasermeister was performed with common AM materials, including Fe-, Ni-, and Ti-based alloys. Herein, to develop process maps for the Ni625 alloy specific to this machine, the processability, microstructure, and mechanical properties of the alloy were experimentally and numerically investigated under various fabrication parameters. Key findings include: 1) A simulation model was established to predict the MP thermal history, including the dimensions and G and R rates; 2) The dimensions and highest temperatures of the MP were considerably affected by P but less affected by V, leading to high P values and increased size and maximum temperature of the MP; 3) Fully dense Ni625 alloy parts (>99.5%) were fabricated under conditions where P was >100 W and V was in the range of 5–15 mm s−1; 4) As P increased and V decreased, a corresponding increase in the dendritic structure and texture was observed. Notably, the sample synthesized with the highest P value of 160 W and lowest V value of 5 mm s−1 exhibited the most pronounced dendritic structure and texture; 5) A positive correlation was observed between the microstructure and tensile properties with lower elongations for finer microstructures. In particular, sample S3, which had the finest microstructure and highest porosity, exhibited the lowest elongation; 6) P ranged from 100 to 160 W, V varied between 5 and 15 mm s−1, and a corresponding process map for ED was established; and 7) The samples showed tensile strength values comparable to those in other research studies, with UTS and YS ranging from 1008 ± 2 to 941 ± 9 MPa and from 682 ± 11 to 640 ± 7 MPa, respectively.

This study demonstrated that a combination of lower P values and smaller hatch spacings can effectively strengthen Ni625 alloys. It was also found that there several parameters can be set to achieve similar outcomes. Indeed, these findings pave the way for the formulation of various “recipes” in the future tailored to the shape and complexity of different parts, thus opening new avenues for part development.

References

  1.  P. R. Gradl, T. Teasley, C. Protz, C. Katsarelis, P. Chen, in AIAA Propulsion and Energy Forum 2021.
  2. P. R. Gradl, C. Protz, T. Wammen, in AIAA Propulsion and Energy Forum and Exposition 2019.
  3. Y. Ekubaru, O. Gokcekaya, T. Ishimoto, K. Sato, K. Manabe, P. Wang, T. Nakano, Mater. Des. 2022, 221, 110976.
  4. D. Svetlizky, M. Das, B. Zheng, A. L. Vyatskikh, S. Bose, A. Bandyopadhyay, J. M. Schoenung, E. J. Lavernia, N. Eliaz, Mater. Today 2021, 49, 271.
  5. D. Svetlizky, B. Zheng, A. Vyatskikh, M. Das, S. Bose, A. Bandyopadhyay, J. M. Schoenung, E. J. Lavernia, N. Eliaz, Mater. Sci. Eng. A 2022, 840, 142967.
  6. A. Bandyopadhyay, K. D. Traxel, M. Lang, M. Juhasz, N. Eliaz, S. Bose, Mater. Today 2022, 52, 207.
  7. J. Wang, Y. Wang, Y. Su, J. Shi, Mater. Sci. Eng. A 2022, 830, 142296.
  8. T. Ünal-Saewe, L. Gahn, J. Kittel, A. Gasser, J. H. Schleifenbaum, Procedia Manuf. 2020, 47, 1050.
  9. J. M. Wilson, C. Piya, Y. C. Shin, F. Zhao, K. Ramani, J. Clean. Prod. 2014, 80, 170.
  10. V. K. Balla, S. Banerjee, S. Bose, A. Bandyopadhyay, Acta Biomater. 2010, 6, 2329.
  11. P. R. Gradl, C. S. Protz, Acta Astronaut. 2020, 174, 148.
  12. S. Pratheesh Kumar, S. Elangovan, R. Mohanraj, J. R. Ramakrishna, Mater. Today Proc. 2021, 46, 7892.
  13. A. Poudel, P. R. Gradl, S. Shao, N. Shamsaei, Mater. Sci. Eng. A 2024, 889, 145826.
  14. J. Nguejio, F. Szmytka, S. Hallais, A. Tanguy, S. Nardone, M. Godino Martinez, Mater. Sci. Eng. A 2019, 764, 138214.
  15. M. Rombouts, G. Maes, M. Mertens, W. Hendrix, J. Laser Appl. 2012, 24, 052007.
  16. Y. L. Hu, Y. L. Li, S. Y. Zhang, X. Lin, Z. H. Wang, W. D. Huang, Mater. Sci. Eng. A 2020, 772, 138711.
  17. G. Marchese, X. Garmendia Colera, F. Calignano, M. Lorusso, S. Biamino, P. Minetola, D. Manfredi, Adv. Eng. Mater. 2017, 19, 2.
  18. Y. L. Hu, X. Lin, X. B. Yu, J. J. Xu, M. Lei, W. D. Huang, J. Alloys Compd. 2017, 711, 267.
  19. Y. L. Hu, X. Lin, S. Y. Zhang, Y. M. Jiang, X. F. Lu, H. O. Yang, W. D. Huang, J. Alloys Compd. 2018, 767, 330.
  20. R. Savinov, Y. Wang, J. Shi, Mater. Sci. Eng. A 2023, 884, 145542.
  21. N. Kwabena Adomako, N. Haghdadi, S. Primig, Mater. Des. 2022, 223, 111245.
  22. G. P. Dinda, A. K. Dasgupta, J. Mazumder, Mater. Sci. Eng. A 2009, 509, 98.
  23. Z. Tian, C. Zhang, D. Wang, W. Liu, X. Fang, D. Wellmann, Y. Zhao, Y. Tian, Appl. Sci. 2020, 10, 81.
  24. D. K. Gorai, T. K. Kundu, IOP Conf. Ser. Mater. Sci. Eng. 2018, 338, 012041.
  25. J. M. V. Quaresma, A. Carlos, A. Garcia, Metall. Mater. Trans. A: Phys. Metall. Mater. Sci. 2000, 31, 3167.
  26. K. S. Cruz, E. S. Meza, F. A. P. Fernandes, J. M. V. Quaresma, L. C. Casteletti, A. Garcia, Metall. Mater. Trans. A: Phys. Metall. Mater. Sci. 2010, 41, 972.
  27. J. Shao, G. Yu, X. He, S. Li, R. Chen, Y. Zhao, Opt. Laser Technol. 2019, 119, 105662.
  28. Y. Chen, Y. Yan, B. Li, Automot. Innov. 2020, 3, 328.
  29. B. K. Anil Kumar, M. G. Ananthaprasad, K. Gopalakrishna, IOP Conf. Ser. Mater. Sci. Eng. 2016, 149, 012116.
  30. O. Gokcekaya, T. Ishimoto, T. Todo, R. Suganuma, R. Fukushima, T. Narushima, T. Nakano, Crystals 2021, 11, 9.
  31. Y. Ekubaru, O. Gokcekaya, T. Nakano, Crystals 2022, 12, 1348.
  32. T. DebRoy, H. L. Wei, J. S. Zuback, T. Mukherjee, J. W. Elmer, J. O. Milewski, A. M. Beese, A. Wilson-Heid, A. De, W. Zhang, Prog. Mater. Sci. 2018, 92, 112.
  33. W. Zhao, Y. Sun, P. Che, Z. Ning, H. Fan, H. Yang, J. Sun, P. K. Liaw, A. H. W. Ngan, Y. Huang, Mater. Des. 2024, 237, 112538.
  34. Z. Zhou, L. Huang, Y. Shang, Y. Li, L. Jiang, Q. Lei, Mater. Des. 2018, 160, 1238.
  35. Z. Tong, X. Ren, J. Jiao, W. Zhou, Y. Ren, Y. Ye, E. A. Larson, J. Gu, J. Alloys Compd. 2019, 785, 1144.
  36. O. Gokcekaya, T. Ishimoto, S. Hibino, J. Yasutomi, T. Narushima, T. Nakano, Acta Mater. 2021, 212, 116876.
  37. S. H. Sun, K. Hagihara, T. Ishimoto, R. Suganuma, Y. F. Xue, T. Nakano, Addit. Manuf. 2021, 47, 102329.
  38. Y. L. Hu, X. Lin, Y. L. Li, S. Y. Zhang, X. H. Gao, F. G. Liu, X. Li, W. D. Huang, Mater. Des. 2020, 186, 108359.
  39. R. J. Vikram, A. Singh, S. Suwas, J. Mater. Res. 2020, 35, 1949.
  40. J. N. Zhu, E. Borisov, X. Liang, E. Farber, M. J. M. Hermans, V. A. Popovich, Addit. Manuf. 2021, 38, 101802.
  41. M. S. Pham, B. Dovgyy, P. A. Hooper, C. M. Gourlay, A. Piglione, Nat. Commun. 2020, 11, 749.
  42. Z. Gan, G. Yu, X. He, S. Li, Int. J. Heat Mass Transf. 2017, 104, 28.
  43. L. E. dos Santos Paes, M. Pereira, F. A. Xavier, W. L. Weingaertner, L. O. Vilarinho, J. Manuf. Process. 2022, 73, 67.
논문 페이지로 이동