Selective Laser Melting of Non-Weldable Nickel Superalloy: Microstructure, Cracks and Texture

Abstract

Additive manufacturing, particularly selective laser melting, presents exciting possibilities for fabricating components from high-temperature nickel-based superalloys. Controlling microstructural features and minimizing defects in fabricated specimens are critical challenges. This study explores the influence of process parameters on microstructure and defect formation in directionally solidified nickel-based superalloy specimens. We conducted a comprehensive analysis of selective laser melting process variables, including interdendritic spacing, crystallization times, and volumetric energy density. Electron backscatter diffraction analysis was employed to assess the feasibility of obtaining a directional structure in single-crystal nickel-based heat-resistant alloy specimens using selective laser melting. The study shows a significant correlation between reduced interdendritic spacing and increased defect formation. Longer crystallization times and higher volumetric energy density lead to decreased defect volumes and sizes. Electron backscatter diffraction analysis confirms the maintenance of preferential growth direction across subsequent layers. Our research underscores the importance of optimizing selective laser melting parameters, balancing refractory elements in alloy composition, and adopting strategies for enhancing crystallization times to minimize structural defects. This comprehensive approach ensures both heat resistance and minimal defects, facilitating the production of high-quality components. These findings contribute to advancing selective laser melting applications in critical industries like aerospace and power generation, where heat-resistant materials are paramount.

1. Introduction

Nowadays, additive technologies are gaining increasing significance as an innovative means of creating complex and functional components. One of the crucial directions in the development of additive technologies is the expansion of the range of materials used. In this context, there is significant interest in researching the formation of directional structures in heat-resistant nickel alloys through the application of additive manufacturing methods.

Developments in the field of heat-resistant nickel alloy metallurgy over the past decade have shown a significant improvement in their operational characteristics with the introduction of various elements, including rare-earth metals, into their chemical composition. While directional structures in heat-resistant nickel alloys were traditionally achieved through casting methods, which allowed for structure formation in the axial temperature gradient of the heater, the growing utilization of this material in additive manufacturing now offers control over the direction of gradient properties within the structure. To manufacture complex geometric components from heat-resistant nickel alloys, selective laser melting (SLM) is employed, following thorough research into the intricacies of structure formation during synthesis on existing nickel alloys and the impact of these characteristics on mechanical properties [1].

Nickel-based heat-resistant alloys represent a distinct class of materials specifically developed to ensure reliability and efficiency under extreme temperatures and high mechanical stresses. This material class plays a pivotal role in the production of components for gas turbine engines, turbojet engines, aerospace structures, and other systems where the heat-resistant properties of the material significantly impact the overall performance of the power plant.

The history of heat-resistant alloys dates back to the 1920s when scientists and engineers recognized the need for materials capable of withstanding extreme operating conditions. During this period, intensive research was conducted on nickel-based alloys with the introduction of various additives and alloying elements. The primary objective was to enhance the heat resistance and high-temperature performance of these materials to improve their efficiency within power generation systems. Heat-resistant nickel alloys have undergone significant advancements over the years, driven by the ever-increasing demands for improved operational characteristics and continuous innovations in metallurgical production methods.

The breakthrough in the development of heat-resistant alloys came with the discovery of the high-temperature γ’ phase, which possesses unique stability properties at elevated temperatures. This discovery opened up new horizons for creating heat-resistant materials capable of successfully operating under high-temperature conditions. Subsequently, an entire family of heat-resistant nickel alloys was developed, which can be categorized into different generations based on their long-term strength characteristics. This classification reflects the continuous evolution and refinement of these alloys to meet the ever-increasing demands of high-temperature applications.

In recent years, there has been a growing interest in research focused on the feasibility of manufacturing components with directional structures from heat-resistant nickel alloys using the selective electron beam melting method. For instance, in the studies [2,3,4], a method for forming a single-crystal structure using selective electron beam melting in nickel alloys CMSX4 and Inconel 718 was demonstrated. This development suggests the potential for manufacturing turbine engine blades with directional and single-crystal structures. The relevance of producing turbine engine blades using additive manufacturing methods is driven by the ability to exert local control over the structure and properties of the manufactured part [3,5,6]. Additionally, these methods offer fewer constraints on the topology of the designed and fabricated components. Furthermore, the utilization of additive methods in the production of turbine engine blades with directional structures demonstrates high economic efficiency in small-batch production and achieves comparable economic efficiency in large-scale production when compared to traditional directional casting technology.

Despite the high technological and economic appeal of additive manufacturing, not all metallic materials are suitable for this technology. Often, when attempting to use new materials, difficulties arise due to the formation of various defects. In particular, heat-resistant nickel alloys with an increased γ’ phase content pose challenges in additive manufacturing [2,7,8]. Cracks frequently appear in the structure of samples due to the formation of significant temperature gradients during high-energy laser or electron beam processing, as well as the low material elongation values [2,7,8,9,10]. The low weldability and the complexity of the chemical composition of alloys also contribute to the formation of hot cracks and high porosity in materials. As a result, a specific structure with non-equilibrium characteristics is formed. Research into ways to prevent the occurrence of such defects is an important direction that will expand the range of materials and ensure the required properties in products. Several studies have investigated the formation and reduction of defects [2,7,8,9,11,12,13]. It has been found that one of the main causes of defect formation in the form of cracks is the segregation of alloy components and their precipitation at grain boundaries, as well as the formation of shrinkage voids [2,13]. One of the most effective ways to influence the material structure is through the energy parameters of the radiation source. These parameters affect the distribution of stresses that occur during material melting. The ability to move the electron beam at speeds of up to 8000 m/s allows for high-temperature heating of the surface layer up to 1000 °C during selective electron beam melting [2,7,8,9,11,13], which reduces the number of microdefects and cracks. However, in selective laser melting technology, achieving a similar beam speed is impossible due to the limitations of the moving mirror system, which deflects the laser beam. In [13], research was conducted on the possibility of reducing the number of cracks in an alloy during selective laser melting by selecting optimal technological parameters.

The determination of the feasibility of manufacturing samples with a directional structure from nickel-based heat-resistant alloy powder containing 4% Re using the selective laser melting method, as well as the study of the peculiarities of structure formation and the reasons for the occurrence of defects in the alloy produced by the SLM method, is an important part of the development of selective laser melting technology.

2. Materials and Methods

2.1. Powder Characteristics

The powder of non-weldable nickel-based superalloy was produced using vacuum induction melting (VIM) gas-atomization with Ar. The chemical composition of the powder containing Ni, Cr, Al, Mo, W, Co, Re, Ta, Nb, and C. The exact chemical composition is given in

Table 1. Chemical composition of nickel-based superalloy atomized powder (weight percent).

Element	Ni	Cr	Al	Mo	W	Co	Re	Ta	Nb	C	B
wt%	bal.	4.9	5.9	1.0	8.5	9.0	4.0	4.0	1.6	0.15	0.02

. The particle size of the powder was determined by laser diffraction (Fritsch Analysette 22 NanoTec plus), giving a mean particle size d50 = 55.8 µm (the particle size of the powder ranged from 20 to 93 µm).

Scanning Electron Microscopy (SEM) of the powder was performed in secondary electron (SE) and backscattered electron (BSE) modes. According to the results of the BSE analysis, it is possible to draw conclusions about the chemical uniformity of the alloy and the absence of numerous particles with surface oxide layers (Figure 1b). Based on the results of SE analysis, it can be concluded that there are a small number of satellites and irregularly shaped particles which will not affect the fluidity of the powder during the process (Figure 1a).

2.2. Sample Preparation

The powder was processed in an AconityMIDI (Aconity3D GmbH, Herzogenrat, Germany) L-PBF system. The laser scans the top of the powder bed in a predetermined pattern. The heat transfer process consists of radiation from the powder bed, convection between the powder bed and the surrounding medium, and heat conduction within the powder bed and between the powder bed and the substrate (Figure 2). The latent heat of fusion in SLM is large. The complexity associated with the phase transition of the powder and the corresponding change in its thermal properties during the SLM process also complicates the heat transfer problem [14].

The system is equipped with a 1070 nm wavelength fiber laser with a maximum power of 1000 W. To investigate the effect of sample microstructure on L-PBF parameters, rectangular samples with a width of 10 mm, depth of 10 mm, and height of 30 mm were fabricated. Polycrystalline substrate (Inconel 718) was preheated to 1100 °C and continuously controlled by a thermocouple attached to molybdenum platform located under the substrate. Preheating is decisively significant because non-weldable nickel-based superalloys are very difficult to sinter due to the formation of a stable Al₂O₃-oxide layer on the particle surface [2]. The process chamber was continuously flooded with high purity argon gas to achieve oxygen content in the chamber below 20 ppm. Parameters of the L-PBF process can be divided into 5 groups according to the volume energy density (E) (Equation (1)) [15]. The E ranged from 30 to 150 J/mm³ and does not change inside the group. Laser power (P) and scanning speed (S) were used as changed factors inside each group. Hatch distance (H) was set to 100 µm. Layer thickness (L) was set to 50 µm (

Table 2. Parameter sets of sample preparation by selective laser melting.

Group	P, W	S, mm/s	E, J/mm3
1	210–420	1400–2800	30
2	300–600	1000–2000	60
3	450–900	1000–2000	90
4	450–900	750–1500	120
5	450–900	600–1200	150

).

$$E=\frac{P}{S·H·L}$$

(1)

Meander stripe Cross snake scanning strategy with boarders, where melting within adjacent scan tracks takes place along two opposite directions, has been used in this work. Rotation of the scan direction from layer to layer create a helix type of solidification path and simulate grain selection process during the standard casting process [3].

2.3. Characterization

The specimens were cut and polished along the build direction. To highlight the microstructure, specimens were etched with CuSO₄, H₂SO₄, and HCl. A Leica DMI 5000 light optical microscope was utilized to evaluate the quantity of microdefects. A Mira 3 Tescan scanning electron microscope (SEM) was used for microstructure and microdefect surface analysis. Local chemical analysis was performed with Energy Dispersive Spectroscopy (EDS). The resulting images (SEM) were analyzed in ImageJ (software v. 1.53c) to determine the average γ/γ’ -phase cell width. The electron backscatter diffraction (EBSD) analysis was performed using a Mira 3 Tescan SEM at accelerating voltage of 20 kV and with a step size of 5 μm². A total of 18 individual maps (700 × 700 μm²) were stitched together to cover the entire area of interest. The data were analyzed with HKL Channel 5 (software v. 5.1).

2.4. Measurement of Residual Stresses

The residual stresses were measured by the XRD-sin2ψ method on a Rigaku SmartLab X-ray diffractometer using a 2D detector. The calculation was carried out from a linear extrapolation of the experimental data with obtaining a function of the form Θ = k∙sin2Ψ + b; the obtained calculated values of k correspond to the angle Θ and the values of b to Θ. The stresses in the corresponding direction are calculated by the following formula (Equation (2)) [16].

$$\sigma =-\frac{E}{1+\nu }ctg\theta ·\Delta \theta ,$$

(2)

where E is the elastic modulus of the sample material, ν is the Poisson’s ratio of the sample material Θ which is defined as the tg of the obtained linear dependence slope, Θ is the calculated diffraction angle which is defined as the point of the obtained linear dependence intersection with the ordinate axis.

3. Results

3.1. Defects

The analysis of the obtained samples, conducted using an optical microscope, revealed that within the initially selected range of process parameters for selective laser melting with a volumetric energy density below 120 J/mm³ (Table 2), macrocracks were formed, primarily in the central and peripheral regions of the specimens (Figure 3a,b). In this regard, specimens from groups 1 and 2 exhibited signs of critical failure, accounting for more than 30% of the surface area of the cut specimens. The proportion of failures in group 3 specimens averaged 3.5%, in group 4 specimens—0.75%, and in group 5 specimens—0.05%.

SEM was employed for the detailed examination of the surface characteristics of macrocracks within the samples. The resulting images reveal distinct features on the macrocrack surfaces. These features include zones of hot intracrystalline fracture marked by the presence of dendritic patterns. Additionally, there are areas of cold intracrystalline fracture characterized by the presence of facetted fracture surfaces, which impart a distinct “riverine pattern” appearance to the overall structure (Figure 3d). This intricate surface morphology provides valuable insights into the mechanisms and conditions of crack formation and propagation within the specimens.

3.2. Microstructure and Chemical Composition

The microstructure of the samples exhibits a highly organized pattern characterized by elongated columnar cells that extend along the build direction of the γ-solid solution. These columnar cells, often referred to as γ/γ’ cells, represent the primary dendrite arms, with secondary arms exhibiting degeneracy. Within each cell, the γ-phase matrix predominates, with cuboidal particles composed of γ’-phase scattered throughout. These cuboidal particles are formed based on the intermetallic compound Ni3Al (Figure 4a,b).

The primary dendrite arm spacing (PDAS) of the γ/γ’ cells exhibit a clear correlation with the applied energy density during the sample manufacturing process, as demonstrated in Figure 4c. Specifically, samples in Group 1 are characterized by the smallest PDAS of the γ/γ’ cells, measuring at just 3.6 µm. However, as the energy density (E) is increased, a notable trend emerges: the PDAS of the γ/γ’ cells steadily increase: group 2–5.9 µm, group 3–6.9 µm, group 4–11.2 µm, group 5–13.5 µm.

In the lower layers of the samples, a substantial concentration of topologically close-packed (TCP) phases is clearly evident in SEM-BSE images (Figure 5 SEM). Detailed EDS analysis has identified the elemental composition of these phases, which is primarily characterized by the presence of Nb, Re, Ta, W, and Mo. Interestingly, when examining elemental distribution through elemental mapping, it becomes apparent that Re, Mo, and W exhibit a more uniform dispersion throughout the cell volume. In contrast, Ta and Nb show a distinct localization within the regions where the TCP phases are present (Figure 5).

The continuous heating of the working platform over 1000 °C plays a pivotal role in the precipitation and growth of the γ’ phase from the solid solution within the material. However, the evolution of this process is not uniform throughout the sample and is influenced by the duration of exposure to elevated temperatures. Notably, the lower part of the sample, which was manufactured earlier and therefore subjected to prolonged heating, exhibits γ’ phase particles that have had ample time to grow to larger dimensions.

Moreover, as one progresses vertically through the sample, a noticeable gradient in the sizes of the γ’ phase particles becomes apparent. The particle sizes vary within a range of 10–200 nm, with smaller particles dominating the upper regions and larger particles prevalent in the lower regions relative to the growth direction.

Furthermore, a striking transformation in the morphology of the γ’ phase particles is observable. A transition from the original cube-like shape to a lamellar or raft-like structure occurs. These lamellae are orientated both along and across the growth direction, as shown in Figure 6b,c. The beginning of the γ’ phase particles coagulation process is indicated by this morphological change. Coagulation caused by the influence of internal tensile stresses and high-temperature conditions prevailing in the lower layers of the sample leads to the formation of a “raft” structure. The results of X-ray structural analysis have revealed the presence of residual tensile stresses with a magnitude of 415 MPa.

In Figure 6a, the results of the structural analysis and EBSD of this sample are depicted. It has been observed that when employing such a high energy density value (greater than 150 J/mm³), no defects in the form of cracks are formed within the sample. The dimensions of the examined EBSD area were 6 mm in height, 4 mm in width, focusing on the central part of the sample. Along the periphery of the sample (1–1.5 mm from the border), the microstructure exhibits a higher degree of disorientation compared to the central region. This phenomenon can be attributed to the unmelted powder at the edges of the sample, which acts as nucleation sites for crystal growth, thus promoting the development of new grains.

The results obtained from the EBSD analysis provide evidence of the absence of high-angle grain boundaries across the predominant cross-sectional area of the sample. The predominant crystal orientation throughout this area aligns with the <001> orientation. The imagery reveals the presence of individual crystals, typically measuring around 30–50 µm in size, which exhibit orientations differing from the predominant one. Their occurrence is likely linked to the formation of forced crystallization centers within the melt. These centers may originate from non-metallic inclusions introduced into the melt from the powdered material or result from reactions with residual oxygen within the working chamber.

It is worth noting that the length of these inclusions is relatively small and corresponds to 1–2 layer thicknesses. This observation suggests that, as a result of competitive growth, such crystals are displaced by others growing in directions either close to or matching the predominant orientation.

Based on the analysis of crystal orientation, it can be concluded that when a predominant growth direction is established, this direction is maintained in subsequent layers, and abrupt fluctuations in orientation are quickly suppressed, contributing to the preservation of the predominant orientation.

4. Discussion

Based on the obtained data, it can be concluded that the quantity of macro- and microdefects is directly related to the width of the γ/γ’ cells. Microdefects predominantly form at the boundaries between γ/γ’ cells, especially in regions with a high concentration of phase segregations. The crystallization process of the melt within the interdendritic zones, constrained by horizontal boundaries of growing dendrites and vertical boundaries of carbides, can lead to the isolation of certain volumes of the liquid phase. Subsequently, due to thermal shrinkage observed at the crystallization front in areas isolated from the overall liquid phase, micro-pores and other defects associated with shrinkage are formed. For high crystallization rates, a reduction in the width of γ/γ’ cells and an increase in the extent of boundaries are characteristic, which, in turn, increases the probability of defects associated with shrinkage occurring within the sample’s structure.

Furthermore, as previously mentioned, in the lower layers of the sample subjected to continuous heating of the platform to temperatures exceeding 1000 °C, the coagulation of γ’-phase particles and the formation of raft structures are observed. This indicates that the phase of the alloy undergoes a loss of its cuboidal morphology, with its particles coalescing and growing intensively in a direction perpendicular to [001]. This suggests creep behavior in the central portion of the lower layers of the material. Additionally, it can be inferred that, in the presence of a significant accumulation of “hot” defects under the influence of internal stresses, such a structure may lead to the cold fracture of the sample.

It is also worth noting that prolonged exposure of the lower layers of the sample to temperatures exceeding 1000 °C leads to the formation of phases with a topologically dense packing based on Nb, Re, Ta, W, and Mo. These phases have a detrimental effect on the mechanical properties of the alloy due to the local disruption of coherence at interphase boundaries in the regions of their formation, ultimately resulting in the formation of wedge-shaped microcracks during high-temperature creep of the material and subsequent alloy failure. Haibo Long et al. [17] demonstrated that the microstructure, thermal stability, and mechanical properties of Ni-based alloys strongly depend on alloying elements and their concentration. In their study, alloying was the primary strategy for stabilizing compositions, microstructures, and thermo-mechanical properties. It is known that Re is an effective strengthening element, and the heat resistance of the alloy increases by approximately 30 °C with the addition of 3 wt.% Re, as observed in the first three generations of Ni-based superalloys [18]. However, Re disrupts the equilibrium of the solid solution in local areas, leading to the precipitation of TCP phases. Excessive lattice mismatches between the TCP phases and the γ’ phase create internal stresses at the phase boundaries, promoting crack initiation [19,20]. To inhibit the precipitation of TCP phases, 3 wt.% Ru began to be added to fourth-generation alloys, resulting in an increase in the creep resistance temperature from ~1060 °C for third-generation superalloys to ~1080 °C for fourth-generation superalloys [21]. It is necessary to consider that the balance of various elements also influences not only the high-temperature properties but also the feasibility of using SLM for manufacturing components from these alloys.

Therefore, for the successful application of the SLM method in the production of components from heat-resistant nickel alloys, both the process’s technological parameters and the chemical composition of the alloy, especially the balance of refractory elements, must be taken into account to ensure not only heat resistance but also the production of defect-free components through SLM. Reducing the volume fraction and size of defects can be achieved by increasing the crystallization time, which leads to an increase in the distance between primary dendrite arms (PDAS) and reduces the overall length of boundaries. This, in turn, reduces the likelihood of defect formation. In the selective laser melting process, the crystallization time can be increased by deepening the sample’s melt during scanning, which can be achieved by increasing the volumetric energy density. It is also worth mentioning that there are other methods to increase the crystallization time. For example, in the work of Buchbender et al., the concept of cooling to reduce substrate temperature during the process to mitigate thermal stresses was proposed [22]. The investigation of crystallographic orientations has shown that the established preferred growth direction is maintained in subsequent layers, and sudden fluctuations in orientation are quickly suppressed, contributing to the preservation of the preferred orientation within the sample’s structure.

5. Conclusions

Based on the obtained data, it can be hypothesized that the number of macrodefects decreases with an increase in the average interdendritic spacing of γ-solid solution cells. During the crystallization of the melt in the interdendritic zones, confined horizontally by the boundaries of growing dendrites and vertically by carbides, some volumes of the liquid phase may become isolated. Subsequently, due to thermal shrinkage observed at the crystallization front in zones isolated from the overall liquid phase, micro-pores and other shrinkage-related defects are formed. At high crystallization rates, a reduced width of γ-solid solution cells and increased boundary length are characteristic, which, in turn, increases the probability of shrinkage-related defects in the specimen.

The volume fraction and size of defects can be reduced by increasing the crystallization time, which increases the distance between the axes of the first order and reduces the overall boundary length, thereby reducing the likelihood of defect formation. Increasing the crystallization time during selective laser melting (SLM) can be achieved by increasing the depth of specimen melting during scanning. This can be accomplished by increasing the volumetric energy density. It is worth noting that there are other ways to increase the crystallization time. For instance, the concept of cooling was proposed to lower the substrate temperature during the process to reduce thermal stresses.

The analysis of crystal orientation suggests that an established preferential growth direction is maintained in subsequent layers, which contributes to the preservation of the preferential orientation in the specimen’s structure.

In this study, it has been shown that it is possible to produce specimens with a directional structure from a powder of a heat-resistant nickel-based alloy using the selective laser melting (SLM) method. Possible reasons for the formation of macro- and microdefects and methods for their elimination through SLM process parameters have been identified. Technological parameters of the SLM process have been developed, allowing for the production of dense specimens with a directional structure. EBSD analysis of the specimens showed the absence of high-angle boundaries, indicating the possibility of obtaining a directional structure in single-crystal nickel-based heat-resistant alloy specimens using SLM.

In conclusion, this work highlights the importance of considering both the technological parameters of the process and the chemical composition of the alloy, especially the balance of refractory elements, when using the selective laser melting (SLM) method to manufacture components from heat-resistant nickel-based alloys. This not only ensures the necessary heat resistance but also minimizes defects in the specimen’s structure, which is critical for the production of high-quality components.