Crack Formation and Control in an AlCoCrFeNi High Entropy Alloy Fabricated by Selective Laser Melting

Abstract

The equiatomic AlCoCrFeNi high entropy alloy (HEA) is prone to cracking during the additive manufacturing process due to the high cooling rates observed, which limits its application to a large extent. In this study, the selective laser melting (SLM) technique was adopted to fabricate the alloy and the mechanism of crack formation was revealed. Most importantly, a new design strategy was proposed to suppress the generation of cracks, and the optimization of the preparation process was also studied in detail. It is found that the interlaminar crack is related to the heat input at the edge of the specimen, and the internal cracks are formed by solidification cracks. Alloys without interlaminar crack can be prepared by means of combination of the side inclination angle and the process parameters. Side inclination angle optimization provides a possibility for the preparation of crack-free AlCoCrFeNi HEA by SLM.

Introduction

High entropy alloys (HEAs) are a new class of materials that contains five or more principal elements, and the atomic ratio of each element varies between 5% and 35%.¹ The concept was first proposed by Yeh et al in 2004.¹ In the same year, Cantor et al also named it multi-principal-elemental alloys.² The design concept of HEA breaks the traditional design framework of single metal as a principal component, and its characteristics are dominated by each principal component.

The HEA results in a solid solution phase with a simple crystal structure such as body-centered cubic (BCC), face-centered cubic (FCC), hexagonal close-packing (HCP), or FCC+BCC after solidification, where the solute from solvent cannot be distinguished.³ In brief, the unique effects such as high entropy effect, lattice distortion effect, sluggish diffusion effect, and “cocktail” effect make HEA possess some excellent properties, that is, high strength, high hardness, high wear resistance, outstanding corrosion resistance, good thermal stability, and radiation resistance.^4–10

The AlCoCrFeNi is a dual-phase BCC (A2/B2) HEA with considerable potential for high-temperature structural applications and has attracted particular attention in the past few years.^11–14 Currently, HEAs are mainly fabricated by casting (such as vacuum arc melting) or powder metallurgy and followed by rolling, forging, and annealing treatment to improve their performance.^15–17 However, it is difficult to obtain ultrafine grain structures by these methods due to the low cooling rate during the casting process and other limitations associated with the subsequent processes.

Additive manufacturing (AM) techniques characterized by rapid heating and cooling rates are beneficial to obtain (super-saturated) solid solution and ultra-fine homogeneous microstructure.^18–21 At present, the two main representative categories for AM-HEA are directed energy deposition (DED) and powder bed fusion processes.²² Sistla et al²³ prepared Al_xFeCoCrNi_2-x HEA with thin wall structures by DED and reported that the microhardness increases with a decrease in average grain size during the laser cladding process.²⁴

Moreover, Wang et al reported that almost a B2 solid solution structure was obtained because of higher cooling rates observed during the direct laser deposition process.²⁵ In addition, wire arc additive manufacturing (WAAM) is another interesting technology for the fabrication of HEAs,²⁶ where Shen et al manufactured the thin-walled AlCoCrFeNi HEAs by combined cable WAAM.^27,28 Results showed that travel speed can be varied to modify the heat input (reduce) to achieve higher cooling rates that lead to refined microstructure in HEAs with improved mechanical properties. The alloys produced by selective electron beam melting exhibited both BCC and FCC phases due to the raster heating that takes place before the melting process, promoting the precipitation of the FCC phase at grain boundaries.^29,30

However, only very few reports exist that study the fabrication of AlCoCrFeNi-HEA by selective laser melting (SLM). The HEAs may be brittle and can crack easily during the SLM process.³¹ The fabrication of AlCoCrFeNi-HEA by SLM leads to the formation of both B2 and A2 phases, fine grains with a dendrite-like substructure.^32,33 Even though the studies deal with the microstructure of the AlCoCrFeNi-HEA, a detailed study on their mechanical properties is missing. Karimi et al reported the successful fabrication of AlCoCrFeNi-HEA with layered/porous structure using SLM³¹; nevertheless, the sample dimensions that can be produced without defects are still questionable.

It has been found that preheating treatment can effectively reduce cracks^34–37 at the expense of production time and processing costs. The intent of this study is to investigate the cracking mechanism of AlCoCrFeNi-HEA prepared by SLM and to suppress the crack formation through optimization of process parameters and scanning strategy. The side inclination angle of the specimen is taken as the influencing factor for the first time in addition to considering the process parameters, and the purpose is to suppress interlaminar cracking by changing the side inclination angle of the block.

Experimental

The gas atomized equiatomic AlCoCrFeNi pre-alloyed powders (VILORY, Jiangsu, China) were used as the raw material. As can be seen in Figure 1a and b, the powders are spherical in shape and are distributed between 15 and 53 μm in size with the average particle size of D₅₀ being ∼36 μm. The nominal chemical composition of the AlCoCrFeNi-HEA powder is shown in Table 1. The energy-dispersive spectroscopy results in Figure 1c show that the elements in the HEA are distributed uniformly in the powder without segregation.

FIG. 1.

(a) Morphology and (b) particle size distribution of HEA powders, (c) distribution of elements in powder particles. HEA, high entropy alloy.

Table 1.

Nominal Chemical Composition of AlCoCrFeNi High Entropy Alloy Pre-Alloyed Powders

Element	Al	Cr	Fe	Ni	Co
wt. %	10.66	20.51	22.06	22.99	Balance

The SLM device HBD-100 machine (from HDB, Guangdong, China) with a laser power of 200 W was used to explore the process parameters. The laser beam diameter was set to 60 μm. A stainless-steel plate with a size of 110 × 110 × 15 mm was selected as the base substrate, and argon was used as the protective gas during processing. A strip-based scanning strategy was adopted,²¹ and the width of the strip was set to 2.5 mm. A hatch-style rotation of 67° was adopted between layers to have minimal repetition in scan direction during the building process.^38,39 To reduce the interlaminar cracking, pyramid blocks with a base of 10 × 10 mm² and a height of 7 mm were prepared with the constant side inclination angle at 80°, 73°, and 63° respectively, as shown in Figure 2.

FIG. 2.

Schematics illustrating the shape of the pyramid blocks and their corresponding scanning strategies employed in the present research.

The volume energy density is a comprehensive parameter used during the SLM process, which affects the microstructure and mechanical properties of parts, and can be described as follows in Equation (1)^40,41:

where E is the energy density (J/mm³), P is laser power (W), v is scanning speed (mm/s), h is hatch spacing (mm), and t is layer thickness (mm). The SLM processing parameters used in the present study were listed in Table 2, and the volume energy density varied between 38 and 95 J/mm³. However, it must be noted that the volume energy density equation is just empirical, and the values observed are just guidelines and cannot be taken as absolute values.⁴⁰

Table 2.

Selective Laser Melting Processing Parameters for Fabricating AlCoCrFeNi High Entropy Alloy Samples

Parameters	Value
Laser power P (W)	40, 50, 60
Scanning speed v (mm/s)	300, 400, 500
Layer thickness t (μm)	30
Hatch spacing h (μm)	70

The density of the samples fabricated with different process parameters was measured by the Archimedes method. The defects (such as cracks and porosity) were observed using an optical microscope. X-ray diffraction (XRD; D/MAX2500) with Cu-Kα radiation was used to examine the structural information of the SLM samples. The polished surfaces of specimens were etched with aqua regia for 15–20 s and were characterized by scanning electron microscopy (HITACHI SU-1500). The compression experiment was carried out by a CMT-5205 testing machine, and the cylinder specimens were prepared with 3 mm diameter and 6 mm height. As the present study does not involve any human/animal subjects, IRB approval is not necessary and can be waived.

Results and Discussion

Mechanism of cracking formation

In the cooling process of each layer during SLM, the substrate or the previously deposited layer hinders the shrinkage of the solidifying layer, and the surface of each layer is subjected to tensile stress.⁴² Due to the poor ductility of this alloy and when the tensile stress reaches a certain degree, exceeding the interlaminar strain limit that the alloy can withstand, interlaminar cracking will occur at the edge of the sample, as shown in Figure 3a. Especially at the sharp corner/edges, the partial layer above the crack will first warp upward due to the stress concentration, as indicated by the red circle in Figure 3b.

FIG. 3.

Defects observed in SLMed HEA samples: (a) interlaminar cracks, (b) warp, (c, e) discontinuous holes along the grain boundary, (d, f) solidification cracks. SLM, selective laser melting.

This is consistent with the simulation results that the maximum tensile residual stresses are located near the top of the deposit.⁴³ As shown in Figure 3c and e, discontinuous holes (marked in red dotted line) can be seen in the grain boundary region in both the X-Y plane and the X-Z plane (50 W-300 mm/s sample). This may be due to the formation of the continuous or semi-continuous liquid film in the grain boundary region during solidification.⁴⁴

In the final stages of solidification, the liquid metal is not replenished and discontinuous holes along the grain boundaries are finally formed. The high cooling rate of the SLM process will lead to large thermal gradients/stress. When the liquid phase solidifies and shrinks, the connection between the liquid films will be broken first, thus forming solidification cracks.^45,46 Figure 3d indicates that the bending distribution of cracks is consistent with the grain boundary, and the inside of cracks is a concave/convex columnar sub-grain boundary.

In the X-Z plane, as shown in Figure 3e, the columnar sub-crystals mainly grow along the building direction, so the discontinuous holes also distribute along the depositing direction. Due to the large thermal stress, the cracks will propagate multiple layers with epitaxial columnar crystals, as shown in Figure 3f.

The fracture surface of the as-prepared HEA SLM specimen compressed along the building direction is shown in Figure 4a. A typical cleavage fracture along with the columnar sub-crystalline structure on the side of the solidification crack can be seen, as shown in Figure 4b. Figure 4c is the schematic diagram illustrating the molten pool section. Multiple grains grow epitaxially and perpendicular to the boundaries of the molten pool, and each grain is composed of columnar sub-crystals. When columnar sub-crystals in different directions intersect, the grain boundary (red area) is formed, which is a crack-sensitive region.

FIG. 4.

(a) Fracture surface of SLMed samples, (b) enlarged fracture image of dotted rectangular area in (a), (c) schematic diagram of molten pool section.

The grain along the maximum heat flow direction and <001> direction will grow preferentially, whereas the grain growth in other directions is inhibited.⁴⁷ When the liquid film and tensile stress exist at the same time, solidification cracks will occur at the grain boundary. The crack initiation could occur at the columnar grain boundaries during deformation.

The suppression of the cracks

The strip scanning strategy with a short scanning line length can reduce the cooling tensile stress of the single track.²¹ However, this shows little effect on the edge tensile stress. Figure 5a reveals the effects of the side inclination angle and volume energy density on the forming quality of the specimens. As can be seen, the sample is designed into a pyramid shape, which can effectively reduce the heat input of the edge of the part. When the side angle is 80°, the specimen can be fabricated with a maximum energy density of 63 J/mm³.

FIG. 5.

(a) Effect of the tilt angle on the side and the effect of volumetric energy density on the building quality of the SLMed HEA specimens, (b) effects of the process parameters on density, and (c) effects of volumetric energy density on the relative density of the SLMed HEA alloy sample.

However, due to the accumulation of thermal stress, a wide interlaminar crack will present every 2–3 mm (every 70–100 layers). When the volume energy density exceeds 63 J/mm³, the samples crack almost every 60 layers above the substrate. This will lead to an increase in the heat input continuously, as the heat above the crack is difficult to transfer through the substrate due to the air gap induced by the crack (solid-to-solid contact missing for efficient and reliable heat transfer), and finally, lead to warpage and failure of the solidified part next to the cracks.

This can be attributed to the differences in the physical parameters and the poor combination (compatibility) of the substrate and the printing material. The warping can be controlled by the side inclination angle of the sample. The sample fabricated with a higher volume energy density can be formed with an inclination angle of 73°, and the layer warpage occurs at the position far away from the substrate accordingly. Thus, the sample with a volumetric energy density of 95 J/mm³ can be successfully prepared by printing at 63°. The heat input at the edge of each layer decreases with the increase of the side inclination angle of the sample. Thus, the accumulation of tensile stress and the cooling shrinkage effect at the edge is reduced, and the interlaminar cracking behavior of the edge may be inhibited by increasing the side inclination angle. The density of samples fabricated with different parameters is shown in Figure 5b.

When the power and side inclination angle are constant, the density gradually increases with the decrease of scanning speed with the highest density reaching 6.82 g/cm³. When the process parameters are the same, the sample with an inclination angle of 73° shows the highest density. In this study, the optimal process parameter observed is as follows: laser power 50 W and scanning speed 300 mm/s. As shown in Figure 5c, the relative density increases with increased volume energy density, exceeding 95% when the volumetric energy density is higher than 70 J/mm³.

The increase of heat input per unit volume with the increase of the energy density is conducive to the full melting of the powder and reduces the pores caused by insufficient fusion, thus improving the densification of the formed parts. The sample with an inclination angle of 73° shows the highest relative density, followed by that of 80°, whereas the specimen with the best surface forming quality of 63° displays the lowest relative density, especially when the energy density is low. The possible reason is that better external forming quality leads to higher internal residual stress, resulting in more internal defects.

The microstructure of the sample fabricated with an inclination angle of 73° is displayed in Figure 6. At low volumetric energy density, the samples show irregularly shaped pores and cracks as internal defects. The formation of pores can be attributed to the insufficient melting of the powder particles due to insufficient energy input.⁴⁸ In addition, a high scanning speed will reduce the energy absorption and the molten pool track become thinner, which leads to poor overlapping of the adjacent tracks.

FIG. 6.

Optical microscopies of the SLMed HEA samples as a function of laser powers and scanning speeds.

With the printing process, the thermal stress gradually accumulates continuously, and a part of the residual thermal stress can be released through the deformation of the hole. Therefore, when the energy density is low, the defects mainly appear in the bulk (core/inner sections) of the sample, which can restrain the occurrence of edge cracking, and a better side morphology can be obtained. With the increase of energy density, the sample becomes densified gradually. At higher volume energy density, the large pores caused by insufficient fusion are basically eliminated, and the cracks also become small. However, the interlayer cracks on the side of the sample are serious due to the residual stress released by external deformation.

Previous studies have indicated that porosity is inversely proportional to crack density.⁴⁹ In this study, interlaminar crack is quite an important factor. Low-volume energy density can reduce interlaminar cracks, but more pores and inner cracks would be introduced. Increasing the volume energy density can greatly reduce the pores and internal cracks; however, the effect of the energy density on interlaminar cracks is more significant. With the increase of volume energy density, the surface of the sample above the interlaminar crack is warped when the inner crack cannot be eliminated, which affects the deposition process.

The interlaminar cracks can be eliminated at high-volume energy density through side tilt angle optimization, and then the inner cracks can be suppressed by increasing energy density. The compressive stress-strain curve at room temperature of specimens with a 63° tilt angle is shown in Figure 7. The compressive yield strength and plastic strain increase with the increase of volumetric energy density, which is consistent with the relative density of alloys with different parameters. When the volumetric energy density exceeds 79 J/mm³, the compressive strength exceeds 1 GPa.

FIG. 7.

Compression test of SLMed AlCoCrFeNi alloys at room temperature.

Phase and microstructure analysis

Figure 8a shows the XRD pattern of the atomized pre-alloyed powder and the SLMed sample (50 W-300 mm/s). The results illustrate that both the powders and the SLM alloy are composed of a double BCC phase consisting of a disordered BCC phase (A2) and an ordered BCC phase (B2). The presence of the (100) peak indicates the existence of the B2 phase, and other diffraction peaks of the B2 phase including the (110), (200), and (211) are overlapped with those of the disordered A2 phase.⁵⁰ This is consistent with previous reports.^32,34

FIG. 8.

XRD patterns of AlCoCrFeNi HEA: (a) pre-alloyed powders and SLMed sample and (b) enlarged view showing (110) peak. XRD, X-ray diffraction.

Moreover, it can be found from Figure 8b that the (110) diffraction peak shifts to a lower 2θ angle and becomes broader for the SLM sample compared with the pre-alloyed powders, especially with increasing volumetric energy density. The diffraction peak shifting toward the left can be ascribed to the existence of residual stress presented in the samples due to the high cooling rate observed during the SLM process, resulting in severe lattice deformation.⁵¹ The broadening of the diffraction peaks can be attributed to the formation of fine/refined grains. The (110) peak of the A2 and B2 phases for the original powders can be distinguished by comparing standard PDF cards, but their difference is hard to observe due to the broadening of the peaks in the SLMed samples.

Figure 9a depicts the morphology of the SLMed sample (50 W-300 mm/s) observed perpendicular to the building direction. The microstructure can be divided into two zones: the matrix and the overlapping zone—the boundary of the scanning track. The enlarged part is shown in Figure 9b; two typical sub-crystalline structures, cellular and columnar, are presented on both sides of the molten pool boundary (dashed line), and the average diameter of the subcellular and the width of the columnar substructure are ∼1 μm.

FIG. 9.

Scanning electron microscopy images of the SLMed samples: (a) top view (X-Y plane); (b) enlarged view of the selected area shown as red dashed square in (a); (c, d) side view (X-Z plane). Dashed lines represents the molten pool boundary.

Figure 9c shows the microstructure along the building direction. It is obvious that the columnar grains grow epitaxially from the inner side of the molten pool boundary to the center of the molten pool, as observed in Figure 9b and d. As is well known that the direction perpendicular to the melt-pool contour is the maximum heat flow direction,⁵² the edge of the scanning track displays a high cooling rate and temperature gradient, and the grains can grow preferentially through the molten pool. In the X-Y plane, the microstructure is mainly cellular, because the direction of the dendrite solidification structure is almost parallel to the building direction.⁵³

On the outside of the molten pool boundary, the new scanning track covers the edge of the previous track; therefore, there exists some difference in the cooling rate and hence a difference in the morphology of the microstructure.

Conclusion

In this study, the AlCoCrFeNi-HEA was successfully fabricated by SLM, and the following conclusions can be drawn:

• (1)The rapid cooling rate of SLM and the existence of liquid film result in the formation of solidification cracks at grain boundaries. The interlaminar cracks are related to the heat input at the edge of the specimen.
• (2)High relative density samples without interlayer cracks are prepared; both the process parameters and scanning strategy are optimized as laser power 50 W, scanning speed 300 mm/s, and side tilt angle 73°.
• (3)The SLM-ed AlCoCrFeNi HEA is composed of A2 and B2 phases, and the microstructure is composed of a dendritic sub-crystal structure.

Authors' Contributions

S.W.: Conceptualization, writing—review and editing, writing—original draft, formal analysis, investigation, methodology, resources, software, validation, and visualization. P.M.: Funding acquisition, methodology, project administration, and supervision. Y.F.: Formal analysis, data curation. Z.Z.: Software, methodology. Z.Y.: Formal analysis, investigation, and software. X.S.: Conceptualization, writing—review and editing, and visualization. K.G.P.: Conceptualization, funding acquisition, project administration, supervision, and writing—review and editing.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported by the State Key Laboratory of Advanced Welding and Joining, Harbin Institute of Technology (No. AWJ-22M05), and the National Natural Science Foundation of China (No. 51971129).