Characteristics of Metal Specimens Formed by Selective Laser Melting: A State-of-the-Art Review

Abstract

The process of selective laser melting (SLM) is complex and instantaneous, which communicates with a variety of physical phenomena such as solid–liquid phase transition. The microstructure and material properties are favorable evidence to explain the SLM forming mechanism. In this paper, the forming properties of the iron-based alloy, titanium-based alloys, and aluminum-based alloys are discussed, which are relatively mature in research. The influence of processing strategy on the density and quality of metal specimens is briefly introduced. The research progress of microstructure and mechanical properties of metal specimens formed by SLM is mainly introduced. Also, the influence of processing parameters and heat treatment on the microstructure and the effect of microstructure on mechanical properties are summarized. The corrosion behavior of metal specimens formed by SLM is analyzed. Furthermore, the research development of copper-based alloys, shape memory alloys, and high-entropy alloys is presented. Finally, the problems encountered in the alloys forming process are described, and the future development direction has prospected.

Introduction

The high-energy laser beam of the selective laser melting (SLM) equipment melts the metal powder dispersed on the substrate in a protective atmosphere along the planned three-dimensional model path (Ref 1). The metals change rapidly from liquid to solid. The process is repeated until the forming process is completed. The SLM forming method has broad application prospects because it is not restricted by complex structures and thin-walled components (Ref 2, 3). Besides, the SLM forming method significantly shortens the design and production time, and the recycling of metal powder improves the utilization rate of raw materials and reduces the production cost (Ref 4). The metal powder particles are melted to form a molten pool and solidify rapidly in a short time (usually several tens of microseconds). In this process, the temperature gradient reaches 10⁶ K/cm, and the cooling rate during the solidification of the molten pool amounts to (10³-10⁸ K/s) (Ref 5). Under the effect of rapid heating and cooling, the growth of grains can be suppressed, and the grain size can reach the micron level (Ref 6). Due to factors such as process instability, step effects, and sticky powder, the surface quality, and dimensional accuracy of specimens formed by SLM usually cannot meet the requirements of industrial production. To meet the user requirements, the specimens need to be processed, but this will increase the production time (Ref 7). The forming process of SLM involves complex powder–liquid–solid phase transitions, and many physical fields are mutually coupled, such as the absorption and scattering of laser radiation by powder, heat transfer between powder and solid, melt flow, evaporation and volatilization, phase change, etc. The differences in materials, the complexity of the forming process, and the instability of the process will produce factors such as holes, splash, spheroidization, undesired microstructure and mechanical properties, and residual stress in the specimens. Identifying and controlling the above-mentioned physical phenomena and their interaction mechanisms produced by multi-physics coupling are critical for producing full-density specimens.

The powder used in the SLM forming process can be subdivided into pre-alloyed powder, mixed powder, and elemental powder according to the type of composition. The pre-alloyed powder has always been a key application in the field of SLM, and SLM forming mixed powder is a promising cost-effective alternative. The flexibility of powder composition can be provided by in situ alloying. Therefore, in situ alloying is a convenient way to control material properties or create new alloys. Pertinent research shows that mixed powder formed by SLM can also obtain excellent properties (Ref 8, 9). The metal powder materials include iron-based alloys (Ref 10, 11), magnesium-based alloys (Ref 12, 13), aluminum-based alloy (Ref 14, 15), titanium-based alloys (Ref 16, 17), nickel-based alloy (Ref 18, 19), chromium-based alloys (Ref 20, 21), bronze alloys (Ref 22), and various other alloys (Ref 23, 24). Refraction absorption and scattering will occur when the laser acts on the powder. The laser absorption rate of powder is different, so the difficulty of forming different kinds of powder into a specimen is different. According to relevant research reports, most of the alloys formed by SLM can almost achieve a nearly fully relative density by adjusting process parameters, scanning strategies, remelting, enhancement phase, etc. The difference is that the morphological and mechanical properties of the specimens formed by SLM are different.

Research on Powder Properties

The cost and performance of metal powder for SLM have become a bottleneck restricting the development of the industry. The metal powder used in the SLM process needs to carry out the requirements of high purity, high sphericity, excellent fluidity, high bulk density, fine particle size, and narrow particle size distribution. The metal powder used in the SLM is generally spherical, nearly spherical, and a small number of irregular particles. The particle shape will affect the fluidity of the powder, the relative density, and the melting process, which will ultimately have a major impact on the properties of the additive manufacturing products. Attar et al. (Ref 8) confirmed that powder morphology has an essential influence on the relative density and mechanical properties of the specimens formed by SLM. The spherical particles improve the flowability of the powder, the surface and internal structure of the additive-manufactured products, mechanical properties of the SLM-processed parts (Ref 25) and reduce pore defects (Ref 26). On the contrary, the irregularly shaped particles spread on the substrate cause less compaction density, improving the porosity of the specimens formed by SLM (Ref 27).

To ensure the uniformity of powder distribution and the stability of the system providing powder, the particle sphericity and particle size ratio were required to optimize. As shown in Fig. 1, the powder with high sphericity can be easily spread on the substrate by a scraper. Particle size and its distribution will significantly affect the viscosity of the powder, which in turn affects its dispersibility and transportability. The particle gap is too large due to the large particle size of the powder, so it is not easy to form a fully dense specimen. The adhesion and agglomeration of the powder are more likely to happen when the particle is too small, so it could reduce the fluidity of the powder. The excellent forming effect can be obtained by combining the powder with coarse and fine particles. Bourell et al. (Ref 28) found that the particle size distribution of the powder has an important influence on the surface quality, relative density, and mechanical properties of the specimens formed by SLM. Spierings et al. (Ref 29) found that the layer thickness of the powder should be 50% higher than the 90% powder particle diameter, and the fine particle size particles should be enough to fill the gap between the coarse particle size particles. Liu et al. (Ref 30) researched the performance of different particle size distributions of 316L stainless steel formed by SLM. The results show that different behavior characteristics are formed by powder with different particle size distribution. Gu et al. (Ref 31) studied the microstructure and tensile strength of Ti-6Al-4V pre-alloyed obtained from three different suppliers. The study found that the thermal conductivity and the relative density of the same process parameters are different owing to the difference in particle size distribution and the manufacturing process of different suppliers. The microstructure and tensile strength of the specimens from different suppliers are basically the same. Besides, according to the experiments on the additive manufacturing process of Inconel 718 powder, Technology (Ref 32) released a research report showing that the longer the powder is used, the oxygen impurity content increases linearly. The material properties will suffer a lot of damage when the oxygen concentration exceeds 180 ppm, so the raw material should be replaced regularly.

Fig. 1

(Reproduced from Ref 33,34,35,36,37,38, Copyright 2020, 2019, 2018, 2020, 2019, 2015, respectively, with permission from Elsevier)

Metal powder. (a) 316L; (b) Ti6Al4V; (c) AlSi10Mg; (d) Cu-15Ni-8Sn; (e) Inconel 718; (f) NiTi.

In the process of forming the SLM specimens, the substrate is preheated to reduce the temperature gradient. After the powder is spread, the laser acts on the powder, and the energy absorbed by the powder is converted into thermal energy. The thermal energy absorbed depends on the compacted density and reflectivity of the powder. Lasers are also deflected and scattered due to interactions with material electrons and nuclei. Heat loss is caused by volatilization, evaporation, thermal radiation, and thermal convection (the interaction between volatile elements and protective gas). As the powder continues to absorb heat, the heat is conducted into the material, and the powder material is melted into liquid, which depends on the thermal conductivity of the material and the degree of substrate preheating. This process has multi-physics phenomena such as gravity, buoyancy, surface tension, capillary effects, and the Marangoni effect. These phenomena will have different effects depending on the materials and the forming processes. Because the exposure time is measured in microseconds, the molten pool has a short time and a low viscosity. The molten liquid metal solidifies rapidly when the laser moves to the next position. The solidified metal can be used as a support structure, and it also can transfer heat and reduce the temperature gradient. After solidification, the shrinkage of the material will produce stress in the surrounding material when the temperature decreases. In the subsequent processing, due to residual stress, it may lead to partial relaxation (Ref 39), resulting in specimen deformation. Owing to the constant change of temperature gradient and processing temperature, repeated heat treatment in the heat-affected zone around the molten pool may change the microstructure through solid-state transformation.

The forming process of SLM is to melt the powder into the liquid and then solidify into a solid, which has complex physical changes. The powder interacts with the laser beam, and the photons of the laser beam experience multiple reflection and absorption in the powder, and the powder is melted to form a molten pool. In this process, fluid flow, heat conduction, convection, and radiation occur at the same time. As shown in Fig. 2, the fluid flow depends on a variety of phenomena, such as gravity, buoyancy, surface tension, evaporation, and Marangoni convection (thermal capillary convection). The effects of gravity and buoyancy are negligible and will not be considered. The Marangoni convection is linked to surface tension. The energy density of the laser source obeys the Gaussian distribution, and the center temperature is higher than that of the two sides. As a result, the center temperature of the molten pool is high, but the temperature coefficient of surface tension is low. The edge temperature of the molten pool is relatively low, but the surface tension of the molten pool is high. Under the action of tension gradient, the melt flows radially outward. The intensity of fluid convection is characterized by the Marangoni number (M) (Ref 40):

$$M=-\frac{\mathrm{D}\gamma }{\mathrm{D}T}\frac{L\Delta T}{\mu \alpha }$$

(1)

where L is the width of the molten pool, ΔT is the temperature difference between the inside and outside of the molten pool, μ is the dynamic viscosity, α is the thermal conductivity of the material, and Dγ/DT is the temperature sensitivity of the surface tension. If there are other active elements, the surface tension will become a function of the interaction of temperature and other active elements (Ref 41):

Fig. 2

Schematic of the physical phenomena in the SLM process

$$\frac{\partial \sigma }{\partial T}=-A-R{\Gamma }_{\mathrm{s}}\mathrm{ln}\left(1+{K}_{\mathrm{seg}}{a}_{i}\right)-\frac{{K}_{\mathrm{seg}}{a}_{i}}{1+{K}_{\mathrm{seg}}{a}_{i}}\frac{{\Gamma }_{\mathrm{s}}\Delta {H}^{0}}{T}$$

(2)

Among them, $\partial\upsigma /\partial \mathrm{T}$ is the surface tension coefficient related to temperature and concentration of active elements; A is the constant; R is the gas constant; Γ_s is the saturated surface excess; K_seg is the equilibrium adsorption coefficient of active elements; ${a}_{i}$ is the activity of active elements (mass percentage); − ΔH⁰ is the standard adsorption heat. With the increase in the active concentration, the surface tension coefficient decreases, which leads to the convection flow from the outer edge of the molten pool to the center of the molten pool.

If the temperature exceeds the boiling point, the material will cause a large amount of evaporation, and the generated recoil pressure will promote the fluid movement. It will penetrate deeply into a certain thickness of the powder and form a keyhole when the high recoil pressure is applied to the molten pool (Ref 42). Under the condition of high temperature and melt flow, the keyhole will overflow the molten pool, thus improving the density of the parts. The component deviation caused by the selective evaporation of volatile elements will lead to the material’s inhomogeneity and performance degradation (Ref 43). Verhaeghe et al. (Ref 44) used numerical and experimental methods to study the change in molten pool size with material evaporation. The results showed that the evaporation should be included in the numerical model to make the results more realistic. Masmoudi et al. (Ref 45) established a three-dimensional model of powder as a continuum to study the effect of environmental pressure on evaporation behavior. The results showed that the evaporation of the material becomes obvious under low pressure because low pressure helps to reduce the difference between the evaporation point and the melting point. Wu et al. (Ref 46) established a three-dimensional mesoscopic model to study the behavior of the molten pool when evaporation occurred. The results showed that the shape of the molten pool becomes narrower and deeper under the action of recoil pressure so that the main direction of heat transfer in the molten pool becomes vertical.

One of the obstacles affecting the widespread application of metal additive manufacturing is the control of grain structure. The grain structure may affect properties such as hot crack sensitivity and produce anisotropic mechanical properties. Fine equiaxed crystals can reduce thermal cracking and improve performance. Due to the high cooling rate and temperature gradient, the main characteristics of the grains are columnar and textured microstructures. Controlling the formation of equiaxed grains is a challenge. The competition between columnar and equiaxed crystals depends on the Gⁿ/V_s value of the solid–liquid interface (Ref 47):

$${G}^{n}/{V}_{s}<{C}_{st}, \text{Equiaxed grains}$$

$${G}^{n}/{V}_{s}={C}_{st},\text{Columnar to equiaxed transition}$$

(3)

$${G}^{n}/{V}_{s}>{C}_{st}, \text{Columnar grains}$$

where Gⁿ is the temperature gradient, V_s is the local solidification rate, C_st is the criterion for the transition between the columnar and equiaxed crystals, and n is the power index of the temperature gradient. By fitting the curve, n is 4.7986 and C_stis 3.41 × 10³. The laser energy presents a Gaussian distribution with higher central temperature and lower temperatures on both sides. Owing to the existence of temperature gradients in different regions of the molten pool, the crystallization rate and subcooling distribution are different (Ref 48). Therefore, different grain morphologies are formed in the microstructure. As shown in Fig. 3, the planar grains are caused by the wide temperature gradient at the boundary of the molten pool and the slow solidification rate. As the grains grow toward the center of the molten pool, the temperature gradient decreases, and the rate of grain growth and the mass fraction of the solute and the structural supercooled zone gradually increase. The equiaxed grain will eventually form when the grains continue to grow toward the center of the molten pool (Ref 49).

Fig. 3

(Reproduced from Ref 37, Copyright 2018, with permission from Elsevier)

Grain map measured by electron backscattered diffraction technology.

Characteristics of Alloys Forming

Iron-based Alloys

Stainless steel, high-strength steel, tool steel, die steel, high-speed steel, iron-nickel alloys, and so on, are all iron-based alloys. Among them, the research and application of 316L stainless steel (316L SS) are the most common. 316L SS has the advantages of wide source, low cost, simple preparation, excellent formability, etc., and also has high structural strength, excellent mechanical properties, creep resistance, and corrosion resistance. The research on iron-based alloys mainly focuses on the relative density, surface quality, surface roughness, forming precision, corrosion resistance, mechanical properties (such as tensile strength, yield strength, elongation, hardness, etc.), and microstructure of the alloys. Iron-based alloys are widely used in aerospace, medical equipment, marine, industrial, and other fields (Ref 50). The following mainly introduces some research status of scholars on 316L SS.

The methods to improve the forming quality of SLM are summarized in Table 1. Wang et al. (Ref 51) studied the surface roughness of 316L SS formed by SLM. To improve the surface quality and reduce surface roughness, it is necessary to comprehensively consider the effects of laser power, scanning speed, hatch space, and layer thickness. It is necessary to adjust the process parameters (such as increasing the scanning speed and using the interlayer interlaced scanning strategy) to improve the surface forming quality when the layer thickness is increased. Strano et al. (Ref 52) analyzed the effects of the layer thickness and the particle size of the powder on the surface roughness of overhanging structures. Due to the laser forming process, a large amount of powder will adhere to the surface of the specimens. The “step effect” will be caused due to the inconsistency of layer-by-layer stacking when the overhanging structures are formed. The powder adhered to the surface of the overhanging structure causes the “step effect” to be more prominent when the layer thickness is equivalent to the particle size of the powder. Yang (Ref 53) proposed a scanning strategy combining contour scanning and orthogonal scanning to improve scanning accuracy. He used surface remelting to improve the surface roughness of 316L SS formed by SLM. Wang et al. (Ref 54) found that four different types of single track are caused by different energy densities. The single tracks include regular coarse track, regular fine track, regular but occasionally broken track, and irregular spheroidized track. The energy density (J) formula is as follows:

$$J=\frac{P}{h\cdot v\cdot l}$$

(4)

where $P$ is the laser power, h is the layer thickness, v is the scanning speed, and l is the hatch space. The smaller the laser power and the faster the scanning speed, the more regular and fine the resulting trajectory. Energy density is the product of comprehensive consideration of process parameters. Different kinds of metal powder have different optimum energy ranges. Only under the appropriate energy density, the formed track is regular and fine. Sun et al. (Ref 55) concluded that high-power laser (380 W) can significantly improve the forming efficiency (about 72%) and obtain high-density (> 99%) specimens with appropriate scanning strategies and scanning parameters. The final forming quality is linked to the cross-sectional density and surface quality, and the cross-section quality affects the final forming quality. Wang et al. (Ref 56) studied the formation of 316L SS with a layer thickness of 150 μm. The forming density can reach 99.99%, and the forming efficiency is almost ten times higher than before.

Table 1 Ways to improve forming quality

The mesoscopic structure of 316L SS formed by SLM is similar to laser welding. The powder absorbs sufficient heat to produce liquid melt (also known as the molten pool) when the metal powder is irradiated by the high-energy laser beam. There is a competitive relationship between the values of Gⁿ/V_s at the solid–liquid interface during the solidification of the melt, so different subgrain structures such as columnar crystal and equiaxed crystal will be formed. The submicroscopic structure of 316L SS formed by SLM is shown in Fig. 4(a) (Ref 57). The size of the molten pool depends on the diameter of the laser spot and the input of laser energy density. Remelting and superposition of the formed crystal structure appear due to the existence of point distance and hatch space. Equiaxed crystal (Fig. 4d) and columnar crystal structures are formed as shown in Fig. 4(c) due to the high cooling rate and nonequilibrium conditions. The elongation of the macrostructure corresponds to the elongation of the microstructure molten pool and equiaxed crystal along the tensile direction after the tensile test, as shown in Fig. 4(b), (e), and (f).

Fig. 4

(Reproduced from Ref 57, Copyright 2015, with permission from Elsevier)

Cross-sectional topography of SLM forming 316 stainless steel. (a) Molten pool topography of an as-built sample. (b) Morphology of the molten pool after the tensile test. (c) The SEM images of grain structure. (d) The SEM images of isometric grain structure. (e) Side view of equiaxed grain structure after the tensile test. (f) The cross-sectional view of equiaxed grain structure after the tensile test.

Analysis of material behavior on a macroscopic scale by considering microstructural properties is one of the advanced methods in computational materials science. Simonovski et al. (Ref 58) provided a computational framework for the finite element simulation of intergranular crack propagation in polycrystalline aggregates by using a cohesive zone model (CZM) in the grain boundaries to study the damage of the material. They (Ref 59) also studied the effect of grain boundary strength and cohesion parameters on the macroscopic response of stainless steel polycrystalline aggregates. As shown in Fig. 5, Ahmadi et al. (Ref 60) established a mathematical model to predict the influence of different microstructures on the mechanical properties of specimens and linked the macromechanical properties with the microstructure and processing parameters. Wang et al. (Ref 61) found that the crystallization types in different areas of the molten pool are different due to different temperature gradient and heat flux direction during the solidification process of SLM. At the bottom of the molten pool, the crystal grains first grow in a planar manner and then grow in the manner of cell dendrites. In the middle of the molten pool, only the cell-dendritic mode is performed. The cell-dendritic direction at the top of the molten pool changes along the laser scanning direction to the maximum heat flow direction. The Marangoni flow has a great influence on the growth orientation of the grains by changing the direction of the heat flux due to its heat and mass transfer. Suryawanshi et al. (Ref 62) studied the effects of a one-way scanning strategy and a checkerboard scanning strategy on the microstructure and mechanical properties of 316L SS. The checkerboard scanning strategy has better performance than the one-way scanning strategy. Jeon et al. (Ref 63) studied the anisotropic behavior of SLM forming 316L SS in tensile and compression tests and analyzed the influence of microstructure and pore defects on the mechanical properties of the material. He believes that the porosity effect is the main factor affecting the anisotropy of the tensile test through the analysis of fracture morphology and strain concentration. The strain concentration around the pores leads to local yield and crack growth, which leads to serious anisotropy of strength and ductility in different directions.

Fig. 5

(Reproduced from Ref 60, Copyright 2016, with permission from Elsevier)

Establishing the relationship model between microstructure and mechanical properties.

Table 2 summarizes the mechanical properties of 316L SS. Compared with cast specimens, the mechanical properties of specimens formed by SLM are significantly enhanced due to the fine microstructure. The large temperature gradient and rapid solidification rate in the process of SLM lead to grain refinement. A too high cooling rate will lead to a large number of dislocations. According to the dislocation theory, the grain boundary hinders the dislocation movement and expands when the grain is refined. Therefore, dislocation movement and grain deformation need more stress, so the mechanical properties are enhanced. Zhang et al. (Ref 64) showed that the 316L SS specimens formed by SLM can obtain the best mechanical properties while satisfying the optimal placement direction (the processing direction is perpendicular to the stretching direction) and the scanning angle (30°). Simson et al. (Ref 65) studied the effect of residual stress of the 316L SS formed by SLM and found that the original residual stress largely depends on the selected process parameters. The direction of the main stress component produced depends on the layer being inspected. At the top surface, the residual stress shows a higher value in the scanning direction than in the vertical direction. However, on the side, the maximum principal stress is perpendicular to the scanning direction.

Table 2 Mechanical properties of 316L SS

The final performance of the specimens formed by SLM is affected by the corrosion resistance. Sander et al. (Ref 74) found that the 316L SS specimens formed by SLM have a higher pitting potential than forged specimens, and therefore they have excellent corrosion resistance. Chao et al. (Ref 75) showed that the reason for the excellent corrosion resistance of 316L stainless steel formed by SLM is that the structure of the specimens does not contain MnS, so there is no Cr loss near MnS. Kale et al. (Ref 76) found that the corrosion resistance of the specimens formed by SLM is significantly higher than that of spark plasma sintering. The passivation film formed on the 316L SS specimens formed by SLM is thicker than the quenched specimens. Therefore, the 316L SS specimens formed by SLM have excellent corrosion resistance (Ref 77). Interestingly, Laleh et al. came to the opposite conclusion. Laleh et al. (Ref 78) found that the 316L SS specimens formed by SLM have poor corrosion resistance. This is mainly due to the presence of pores in the specimens. Lin et al. (Ref 79) studied the active screen plasma co-alloying treatment with nitrogen and platinum to modify the surface of 316 stainless steel. The results showed that the surface of the treated 316L SS forms a dense columnar single-phase Pt3Fe deposit. Due to its excellent electrical conductivity and corrosion resistance, Pt3Fe has significantly improved its surface conductivity and corrosion resistance of 316L SS. Lv et al. (Ref 80) studied the post-treatment of 316L SS formed by SLM for the surface quality of the specimens. Sandblasting, oxidation removal, and chemical mechanical polishing can effectively improve the surface quality of 316L SS specimens. Among them, the sandblasting can remove spherical particles and pores, thereby inhibiting the formation of corrosion cracks on the surface of the 316L SS specimens. After sandblasting, the passivation film is more stable. The boundary of the molten pool near the sandblasted surface is fuzzy, which can improve the corrosion resistance of 316L SS formed by SLM.

The research of multimaterial grafting technology is also the hot spot of current research. The related properties of the metal can be improved by grafting one metal onto the surface of another. The 316L/CuSn10 bimetal structure was formed by Chen et al. (Ref 81). The bimetal structure formed by SLM is mainly divided into three regions: 316L metal region, CuSn10 metal region, and fusion region, as shown in Fig. 6(a). Because the thermal conductivity of CuSn10 is higher than that of 316L SS, the crack growth is promoted by the heat concentration in the fusion zone. The crack growth starts from the fusion zone and continues to the 316L SS zone, as shown in Fig. 6(b), (c), and (f). The bonding strength of the bimetal structure can reach 423.3 ± 30.2 MPa by optimizing process parameters and scanning strategy. The study of Lyu (Ref 82) confirmed that the hardness and wear resistance of the TiAlN/TiN multilayer film on the SLM specimens are superior to the uncoated.

Fig. 6

(Reproduced from Ref 81, Copyright 2019, with permission from Elsevier)

Microstructure of SLM-formed 316L SS and CuSn10 bimetal: (a) entire fusion zone (100×), (b) area A of entire fusion zone (500×), (c) area B of entire fusion zone (600×), (d) area C of entire fusion zone (3000×), (e) area D of (c) (1500×), (f) schematic diagram of dendritic cracks.

Because of the advantages of low cost and easy forming, the iron-based alloy formed by SLM is the most extensive and in-depth research at present. By using the preheating process, the problems of temperature gradient can be effectively overcome. Density, residual stress, and mechanical properties can be improved by adjusting process parameters and scanning strategies (Ref 83). Through the optimization of boundary contour scanning and upper surface remelting, the forming accuracy and surface forming quality can be effectively improved. The forming efficiency can be effectively improved by increasing the thickness of the forming layer. Coating with heterogeneous materials can improve the hardness, corrosion resistance, and wear resistance of iron-based alloys.

Titanium-based Alloys

Titanium-based alloys (TBAs) are important nonferrous metal, with the advantages of low density, low thermal expansion coefficient, excellent corrosion resistance, high specific strength, perfect low-temperature brittleness resistance, excellent weldability, nontoxic and nonmagnetic, which is widely used in aerospace, marine, petroleum energy, national defense, industry, medical, and other fields (Ref 84). TBAs can be divided into three categories according to the microstructure: α type, β type, and α + β type. β type TBAs are suitable for human implantation. A lot of researches and applications have been carried out on α + β type TBAs. The Ti6Al4V to be described is the most widely studied TBAs.

Scholars have done a lot of research on the process parameters, microstructure, and mechanical properties of TBAs formed by SLM. Sun et al. (Ref 85) used statistical analysis to optimize the process parameters and obtained the Ti-6Al-4V specimens with relatively high density. Because of the unique microstructure of the Ti-6Al-4V specimens formed by SLM, its tensile strength is higher than that of traditional processed. Due to the existence of residual stress, internal porosity, poor surface quality, microstructure, and other factors, SLM formed Ti-6Al-4V specimen has a low fatigue life (Ref 86). To improve the mechanical properties, the microstructure is adjusted by heat treatment (Ref 87, 88), and the surface quality is improved by sandblasting, machining, electropolishing, and other operations (Ref 89). After treatment, the Ti-6Al-4V specimens formed by SLM can achieve the same properties as forging and casting. Due to the characteristics of rapid melting and solidification of SLM, the mechanical properties of Ti-6Al-4V specimens formed by SLM are different from those formed by traditional methods. The mechanical properties of the Ti-6Al-4V specimens formed by SLM can be well understood and analyzed by establishing the finite element model. Lu et al. (Ref 90) established a three-dimensional thermomechanical coupling finite element model of Ti-6Al-4V alloys prepared by SLM and found that the traditional Ti-6Al-4V material data are not suitable for SLM. Tao et al. (Ref 91) established the J-C model which can be used to predict the quasi-static deformation behavior at room temperature. However, due to the problem of consistency, the model constant may depend on the specific process conditions. His modified Arrhenius and BP-ANN models can accurately simulate stress–strain behavior in the process of thermal dynamic compression.

Although the strength of Ti-6Al-4V alloys formed by SLM is excellent (Ref 92, 93), its elongation is less than 10%. The properties of low elongation largely depend on the microstructure, including phase structure, morphology, distribution, and characteristic length scale. Ti6Al4V alloys formed by SLM are mainly composed of a large number of fine acicular α′ martensite and a small amount of initial β columnar crystal (Ref 94), as shown in Fig. 7. It solidifies to form a β-phase columnar crystal when Ti6Al4V alloys are heated to a melting state. The critical cooling rate of the martensite transformation is 410 K/s. At the high cooling rate, β phase columnar crystal cannot be transformed into equilibrium α-phase through diffusion, while it can transform into metastable α′ martensite with a dense hexagonal structure of α-phase supersaturation. The grain boundary of this initial β columnar crystal is prone to intergranular fracture, which reduces its plasticity. SLM parameters have a significant effect on the microstructure of Ti6Al4V alloys. The martensite phase decomposition and β phase precipitation can be realized in the process of additive manufacturing by adjusting the process parameters; thus, the Ti6Al4V specimens with high strength and high ductility can be formed (Ref 95). Xu et al. (Ref 96) found that changing the layer thickness, energy density, and laser focus shift distance can effectively control the content of α′ martensite in Ti6Al4V alloy and realize the in situ decomposition of α′ martensite, resulting in the structure containing only (α + β) lamellae. After heat treatment, the structure of Ti6Al4V alloys usually consists of initial β columnar crystals containing (α + β) lamellar structure. Table 3 shows the typical structural characteristics of TBAs formed by SLM (Ref 97, 98). It can be seen from the table that the structure of the Ti6Al4V alloys formed by SLM after heat treatment is mainly composed of the α + β phase in columnar prior-β grains. However, the β-type titanium alloys are mainly composed of the β phase.

Fig. 7

(Reproduced from Ref 95, Copyright 2015, with permission from Elsevier)

The microstructure of the Ti6Al4V specimens was formed by SLM. (a) Bottom of test piece, (b) molding direction, (c) middle of the specimens, (d) top of the specimens.

Table 3 Typical microstructure characteristics of SLM forming TBAs

The phase transformation between α and β phases can be achieved by adjusting techniques (such as thermomechanical processing or heat treatment), which can produce rich microstructure and mechanical properties. The Ti-6Al-4V alloys are the most widely used implant material because of its excellent performance. However, due to the presence of Al and V, it may cause allergic reactions, such as Alzheimer’s disease and neuropathy (Ref 107). Also, the mismatch of large Young’s modulus (10-30 GPa) between the alloy and the human skeleton can lead to stress shielding (Ref 108). The β-titanium alloy has high strength, superelasticity, excellent biocompatibility, and outstanding corrosion resistance. Liu et al. (Ref 106) found that the specimens formed at low energy density levels are mainly composed of the β phase and nonthermal ω phase, while the specimens formed vertically are nanosized α laths. Through the analysis of mechanical properties, it is found that the specimens without α-laths have excellent tensile strength, yield strength, and elongation. Chen et al. (Ref 109) found that Ti-Nb alloy (Ti-37Nb-6Sn only contains nontoxic elements, such as Nb, Ti, and Sn) is very suitable for biomedical applications.

Table 4 summarizes the mechanical properties of TBAs prepared by different processes. The low elongation of TBAs formed by SLM is due to the presence of a large amount of fine α′ martensite, but TBAs usually have excellent tensile strength and yield strength. The α′ martensite can be transformed into (α + β) lamellar structure by adjusting the process parameters or heat treatment, which can significantly improve the poor mechanical properties of Ti6Al4V alloys. The elongation of TBA exceeds 10% after heat treatment or adding the reinforcing phase.

Table 4 Room-temperature mechanical properties of TBAs and their composites prepared by different processes

To improve the related properties of Ti alloys, Zafari et al. (Ref 117) mixed Ti-64 (50 wt.%) and Ti-5553 powder (both 15-45 μm with D10 = 22, D50 = 35, and D90 = 46 μm) to form mixed alloys different from composite materials, referred to as HYbrid Ti Alloys (HYTAs). As shown in Fig. 8, the mixed Ti alloys have the excellent combination of high strength, uniform elongation, and work hardening rate. The mechanical properties of HYTAs formed by SLM showed three different stages: low work hardening, high work hardening, and decreasing work hardening followed by necking and fracture, as shown in Fig. 8(a). Besides, the properties, including yield strength and elongation, can be selectively adjusted by heat treatment. The Ti-Cu alloys prepared by Zhang et al. (Ref 118) can overcome the negative effect of high-temperature gradient during laser melting. The specimens of Ti-Cu alloys have excellent yield strength and uniform elongation due to their completely equiaxed fine-grained structure.

Fig. 8

(Reproduced from Ref 117, Copyright 2019, with permission from Elsevier)

(a) Tensile stress–strain curves, (b) tensile strength versus uniform elongation, (c) tensile stress–strain curves after post-SLM heat treatments, (d) tensile elongation versus strength for HYTA, conventional Ti alloys, and various high-strength steels.

Ti-Nb alloys are excellent corrosion resistance material. Its excellent corrosion resistance is attributed to the ability to quickly form a stable self-protecting oxide film on the metal surface, such as titanium oxide (TiO₂) and niobium oxide (Nb₂O₅). Wang et al. (Ref 119) found that the Ti-35Nb specimens formed by SLM after heat treatment are more excellent corrosion resistance than the specimens formed by SLM without heat treatment. This is mainly due to the more uniform chemical composition and Nb distribution of the specimens after heat treatment. The homogenization of the alloy elements can promote the development of the passivation film with uniform chemical composition. Dai et al. (Ref 120) showed that the corrosion resistance of Ti-6Al-4V alloys was determined by the amount of α′-phase and β-Ti phase in the microstructure. The corrosion resistance mechanism of the specimens formed by SLM on different planes is related to its microstructure characteristics. The inferior corrosion resistance of build direction is attributed to the presence of more α′ martensite and less β-Ti phase in the microstructure. However, the microstructure of the specimens can be adjusted by adjusting the scanning strategy and heat treatment to improve the mechanical properties (Ref 121). Qin et al. (Ref 122) improved the corrosion resistance of Ti-5Cu alloys by improving the inhomogeneous Ti₂Cu phase. He also found that the corrosion behavior of the Ti-24Nb-4Zr-8Sn (Ti2448) alloys prepared by SLM and forging is similar because the Ti2448 alloys formed by the two methods have a single β phase. It shows that the manufacturing method cannot determine the corrosion resistance of the alloys, but the composition of various phases in the microstructure may determine the corrosion resistance.

The defects of low thermal conductivity and large elastic deformation of TBAs can be effectively improved by SLM. TBAs formed by SLM are prone to defects such as cracks and micropores. Substrate preheating, remelting, and presintering can reduce the generation of thermal stress and effectively improve the forming density and mechanical properties. Although the TBAs formed by SLM can meet the forging standard after heat treatment, further research on the heat treatment process suitable for the TBAs formed by SLM is needed to improve its mechanical properties. Also, the SLM forming of precision large-scale components and the control of deformation still need to be studied.

Aluminum-based Alloys

Aluminum-based alloys (ABAs) have the advantages of light density, high specific strength, high specific rigidity, excellent conductivity, outstanding thermal conductivity, excellent corrosion resistance, etc. It is difficult to form ABAs by SLM due to the high thermal conductivity, high laser reflectivity (about 91%), poor fluidity, and strong oxidation of ABAs. The performance of most ABAs formed by SLM is not ideal. For example, due to the unstable element Zn contained in the Al-Zn series high-strength alloys, the forming process is unstable due to the violent reaction of pores and molten pool. Therefore, the Al-Zn series is generally not suitable for SLM. At present, ABAs formed by SLM are limited, generally AlSi10Mg, AlSi7Mg, AlSi12, and other series of alloys, of which the research of AlSi10Mg alloys is the most in depth. The AlSi10Mg alloys belong to the Al-Si alloys and have a composition close to the eutectic point. The melting point of ABAs is low, and the solidification range of solid–liquid phase is small. The formation of defects can be reduced by reducing the output of laser energy (Ref 123).

Kempen et al. (Ref 124) pointed out that the best process parameters (P = 200 W, v = 1400 mm/s, l = 105 µm) can obtain the smooth monorail shape and stable molten pool state of AlSi10Mg alloys formed by SLM. Buchbinder et al. (Ref 125) showed that process parameters have a significant effect on the density of AlSi10Mg alloys due to the existence of pores. To obtain the excellent relative density of specimens formed by SLM, higher laser power, higher scanning speed, and smaller hatch space should be used. The relative density of the specimens increases with the increase in laser energy density. The relative density will decrease due to overheating when the laser energy density exceeds a certain threshold. The appropriate laser energy density can reduce the spheroidization effect, reduce the porosity, and increase the relative density of the specimens.

The microstructure of AlSi10Mg alloys formed by SLM mainly contains fine cell dendritic structure, as shown in Fig. 9 (Ref 126). In the figure, the gray honeycomb feature is the Al element, and the white fiber is the Si element. There are three different areas in the molten pool, as shown in Fig. 9(d), which are the fine area (MP fine), the coarse area (MP coarse), and the heat-affected area (HAZ) around the molten pool. In the molten pool, dendrites grow to the center of the molten pool. Outside the molten pool, the intercell network is destroyed by coarsening the silicon phase into unique particles, which is caused by the increase in the diffusion rate of Si in the heat-affected zone. Olakanmi et al. (Ref 127) found that Mg/Si elements can affect the shape of the formed surface. The morphology and oxygen content of powder particles will affect the flow and solidification behavior of the molten pool during monolayer forming, while the irregular surface oxide film and particle shape will aggravate the agglomeration phenomenon and form a porous surface. Li et al. (Ref 128) found that TiB₂ nanoparticles can inhibit the grain growth and reduce the grain size of the formed specimens to 2 μm. The TiB₂ particles are distributed at the boundary of grain boundary and cellular structure, which increases the deformation dislocation. Fiocchi et al. (Ref 129) found that at 294 °C, after annealing for 2 hours, the mesh Si structure was broken and evenly distributed on the ABAs matrix, which could effectively improve the microstructure characteristics of the AlSi10Mg alloy.

Fig. 9

(Reproduced from Ref 126, Copyright 2013, with permission from Elsevier)

Microstructure of AlSi10Mg alloy was formed by SLM. (a) Full view of microstructure, (b) Area A in a. (c) Area B in a. (d) Area C in a.

The evolution of the structure during the solidification process plays a vital role in the mechanical properties of specimens formed by SLM. McDonald et al. (Ref 130) found that coarse-grained and acicular eutectic silicon phases induced cracks in a tensile environment, which deteriorated mechanical properties. Rapid solidification can refine microstructure, distribute eutectic silicon uniformly, improve mechanical strength and plasticity. The tensile strength and yield strength of SLM specimens are superior to the casting and forging process, but the plasticity is low (Ref 131). AlSi10Mg alloys can be hardened by precipitating the Mg₂Si phase by T6 heat treatment (completely artificially aged) (Ref 132). Table 5 summarizes the mechanical properties of specimens formed by SLM. It can be seen from the table that the elongation of the ABAs after heat treatment is increased, but the tensile strength and yield strength are reduced.

Table 5 Mechanical properties of Al-Si alloys

Prashanth et al. (138) found that the tensile properties of the formed specimens with double scanning and checkerboard scanning are excellent, and the tensile increased to up to 460 MPa. The results showed that the mechanical properties can be regulated by different scanning methods. The microscopic simulation of Gu et al. (139) showed that the pressure difference and centripetal force acting on the AlN reinforcing particle promoted the sufficient rearrangement of the AlN particles due to the combined effect of the convection vortex, the capillary force, and the gravity force. Regular distribution of AlN reinforcing particles in a ring-like structure was obtained within the finally solidified composites under the optimized laser energy density. Tang et al. (140) found that the porosity induced by oxides has a great influence on the fatigue properties of AlSi10Mg specimens. Brandl et al. (132) formed AlSi10Mg at the preheating temperature of 300 °C. He found that preheating reduces the generation of holes during the forming process, so it can effectively improve the high cycle fatigue performance of the formed structure. Tan et al. (141) found that the hardness of TiB₂ enhanced AlSi10Mg was higher than that of nonenhanced AlSi10Mg. Wang et al. (142) studied the SLM forming of TiC/AlSi10Mg nanocomposites. The results show that the hardness and tensile strength of the composites are much higher than those of the unreinforced AlSi10Mg (Table 6).

Table 6 Mechanical properties of CBAs

Leon et al. (Ref 150) compared the corrosion resistance of AlSi10Mg alloy formed by SLM and casting. The results showed that the AlSi10Mg alloys formed by SLM had a microstructure with more uniform sizes and distribution, so its corrosion resistance is more excellent than that of the AlSi10Mg alloys formed by casting. Xing et al. (Ref 151) used ultrasonic peening treatment (UPT) technology to treat AlSi10Mg alloys. The study found that UPT can effectively reduce the porosity of the AlSi10Mg specimens and increase the density of the material. The corrosion resistance of the UPT-treated specimens has been improved. Yang et al. (Ref 152) showed that the Al-12Si alloy formed by SLM has stronger corrosion resistance than the as-cast. This is attributed to the fact that the Al-12Si alloy formed by SLM has ultrafine silicon particles, which enhances the stability of the oxide layer.

The difficulty of forming ABAs is still the poor fluidity and high laser reflectivity. Poor fluidity causes difficulties in powder spreading. The high laser reflectance causes difficulties in melting and forming. The densification can be improved by optimizing the process parameters and scanning strategy. The composite of the nanoreinforced phase and ABAs can not only improve the laser absorption rate of ABAs but also improve the mechanical properties.

Copper-based Alloys

Copper-based alloys (CBAs) have superior thermal conductivity, low strength, and high density. However, the laser absorption of pure copper is poor, and the absorption rate is less than 2%. Pure copper cannot absorb enough energy due to the low absorption rate, so it is difficult to form pure copper by SLM. Current research on CBAs formed by SLM is mainly focused on tin bronze, shape memory alloys, brass, etc.

Scudino et al. (Ref 153) found that the microstructures of CBAs formed by casting and SLM (as shown in Fig. 10) are both α-Cu (Sn) phase nucleated dendrites and α + δ eutectoids. However, the size of α dendrite and (α + δ) eutectoids of the microstructure formed by SLM was 2.3 and 1.4 um compared to the microstructures formed by casting of 5.3 and 3.5 um. The refined structure has an important influence on the mechanical properties of Cu-10Sn bronze. As shown in Fig. 11, the matrix phase of Cu-15Ni-8Sn alloy formed by powders, cast ingots, and SLM is α-phase, retaining the FCC lattice structure of copper. Because the atomic diameter of Sn is larger than that of Cu and Ni, the significantly high supersaturated solute Sn in the sample formed by SLM caused an increase in the lattice constant, resulting in an increase in crystalline interplanar spacing. The solid solubility of Sn in the Cu-15Ni-8Sn alloy formed by SLM is significantly elevated, which means that the SLM process has a faster cooling rate compared with the gas-atomized powder and the laser cladding process (Ref 33).

Fig. 10

(Reproduced from Ref 153, Copyright 2015, with permission from Elsevier)

(a–c) Microstructure of SLM formed Cu-10Sn alloy, (d) Casting Cu-10Sn alloy. The bright phase is (α + δ)-eutectoid, and the dark phase is α-dendrites.

Fig. 11

(Reproduced from Ref 33, Copyright 2020, with permission from Elsevier)

XRD patterns of Cu-15Ni-8Sn alloy powders, cast ingots, and SLM-manufactured parts.

It is difficult to control the densification and porosity in the process of SLM forming CBAs, which is one of the problems to be solved urgently. The numerical simulation of CBAs formed by SLM is carried out to understand the crack and porosity from the depth. Controlling the grain size is very important to improve product performance and quality. The high reflectivity of copper is easy to damage the laser device. Although the laser absorptivity of pure copper is low, the laser absorptivity of the powder can be effectively improved by mixing the pure copper powder with other powder (Ref 154). The method of coating heterogeneous materials on the surface of pure copper can be used to improve the laser absorption rate of the powder (Ref 155). The absorption rate of pure copper can be improved by using the short-wavelength laser, but the replacement of the laser will inevitably lead to the problem of production efficiency (SLM equipment of the short-wavelength laser is not suitable for the preparation of other materials). There is almost no report on the corrosion resistance of CBAs.

Nickel-based Alloys

Nickel-based alloys (NBAs) contain more than 30% nickel, which has excellent yield property, tensile property, creep property, oxidation resistance, corrosion resistance, weldability, high strength, and excellent properties at 650-1000 °C. NBAs are often used in equipment manufacturing under special working conditions, such as high-temperature aviation engines and gas turbines, oil extraction equipment under marine corrosion environment, stress corrosion cracking resistance equipment in the chemical and process industries, etc. The nickel-based superalloy is one of the most widely used materials in the field of high temperature and corrosion resistance. Currently, Inconel 263, Inconel 625, Inconel 718, Inconel 738LC, GH3536, GH4196, Monel 400, Monel k-500, Hastelloy X, K4202, and Hastelloy C-276 are used in SLM, among which Inconel 718 and Inconel 625 are widely studied and applied.

The microstructure of nickel-based superalloy formed by SLM is mainly composed of matrix γ-phase, main strengthening γ″-phase, auxiliary strengthening γ′-phase, δ-phase, carbon (nitrogen) compound, and Laves. The structure of specimens formed by SLM is different from that of casting. Due to the lack of γ″, γ′, δ phases and the existence of a large number of Laves phases, the properties of specimens formed by SLM are poor. Generally speaking, with the increase in γ″-phase, the strength of the alloys will increase. However, the γ-phase will transform into δ-phase under a high-temperature environment for a long time due to metastable property. Proper increase in δ-phase can refine grains, reduce stress concentration, and improve the plasticity of the alloys. However, the existence of δ-phase is easy to reduce the ductility of the alloys. Therefore, optimizing the SLM forming parameters and heat treatment process parameters will help to optimize the content and distribution of the δ-phase.

Trosch et al. (Ref 156) compared the microstructure and mechanical properties of SLM, forging, and casting. The grain size of SLM is smaller than that of forging and casting. As shown in Fig. 12, δ-phase is deposited at the grain boundary of forged and cast specimens after heat treatment. In the specimens formed by SLM, the δ-phase is distributed not only at the grain boundary but also within the grain. This is due to Nb element segregation caused by rapid cooling. To improve the mechanical properties of SLM specimens, the formation of δ-phase must be reduced. The δ-phase within the grain is dissolved by heat treatment. The microstructure of specimens formed by SLM after heat treatment is similar to that of forging and casting. Li et al. (Ref 157) prepared Inconel 625 specimens by SLM and studied its microstructure. The results showed that the molten pool is composed of elongated columnar crystals, and the preferred orientation of the grains after heat treatment is obvious. Strößner et al. (Ref 158) found that heat treatment temperature affects the distribution of δ-phase in the structure of the alloys. The poor tensile properties of the alloys treated by “980 °C/1h + double aging” are caused by the segregation of Nb element which results in the appearance of δ-phase. After heat treatment of “1065 °C/1h + 980 °C/1h + double aging”, the content of δ-phase was reduced, and the excellent tensile property was obtained. Pröbstle et al. (Ref 159) found that the creep properties of the alloys can reach a fairly good effect after solution treatment at 1000 °C. After solution treatment, the content of δ-phase is reduced, and the small amount of δ-phase improves the resistance sensitivity and intergranular fatigue crack resistance. It can be seen from Table 7 that the NBAs formed by SLM need to be improved. The elongation cannot meet the requirements of casting and forging standards.

Fig. 12

(Reproduced from Ref 156, Copyright 2015, with permission from Elsevier)

Grain size distribution of SLM (measurements parallel (||) and perpendicular (⊥) to the building direction), forged and cast microstructure with corresponding SEM pictures.

Table 7 Mechanical properties of nickel-based alloy

The TiC-reinforced Inconel 718 composites formed by SLM have excellent corrosion resistance compared with the unreinforced Inconel 718, which is attributed to the improvement in the microstructure and the reduction in Cr-lean regions in the matrix. Zhang et al. (Ref 160) compared the corrosion behavior of three planes of TiC/Inconel 718 composites formed by SLM (the vertical direction: YZ-, XZ-planes, the horizontal direction: XY-plane). It is found that the TiC/Inconel 718 composites formed by SLM have electrochemical anisotropy. The YZ-plane shows high corrosion potential, low corrosion current density, and high polarization resistance, which is attributed to the passivation film of chromium oxide formed on the YZ-plane. However, the corrosion resistance of the XY-plane is poor, which is attributed to the large pores and uneven microstructure.

It is still necessary to carry out the research on NBAs formed by SLM. The systematic study of the forming properties of NBAs is not limited to meeting forging standards. The high cost of NBAs powder production is a major bottleneck restricting the study of NBAs formed by SLM. The complex microstructure of NBAs is composed of multiple secondary phases, and the microstructure and characteristics of NBAs are very sensitive to temperature. At present, only Inconel 718 and Inconel 625 are relatively mature in the application, and other NBAs need to be developed to meet the growing demand of all industries.

Shape Memory Alloys

Shape memory alloys (SMAs) can recover their original shape after deformation (under specific temperature/stress conditions) according to thermal or mechanical instructions. When austenite is stressed in a specific temperature range, it transforms into martensite. After unloading stress, martensite transforms into austenite because martensite is unstable without stress. Therefore, when the stress is unloaded, the shape of SMAs is restored by the reversible martensitic transformation. SMAs have been widely used in aerospace, machinery, electronics, architecture, automatic control, and other fields because of their shape memory effect and superelasticity.

NiTi shape memory alloys (NiTi-SMAs) are widely used in the industrial field due to their superelastic strain up to 8%, excellent mechanical properties, corrosion resistance, and biocompatibility, but they also have the disadvantages of high material cost and difficult processing (Ref 167). The industrial application of NiTi-SMAs is limited because the transition temperature is lower than 100 °C (Ref 168). The high-temperature shape memory alloys have been developed to improve the phase transition temperature. However, it is difficult to manufacture Ti-Ni-Pd, Ti-Ni-Pt, and Ni-Ti-Hf because of their high price relatively. NiTi-SMAs are not easy to be processed by conventional methods due to the extremely sensitive composition and poor machinability (Ref 169, 170). The nickel content of the alloys is low due to the evaporation of elements when NiTi-SMAs are prepared by SLM. The phase transition temperature of NiTi-SMAs formed by SLM increases by 30 K due to the decrease in nickel content. Low scanning speed can increase the transformation temperature and stabilize martensite (Ref 171). Increasing martensitic transformation temperature can produce better shape memory effect at a specific temperature through high energy density (Ref 172). Dadbakhsh et al. (Ref 34) used the combination of low laser power and low scanning speed (LP) and the combination of high laser power and high scanning speed (HP) to prepare specimens. The difference is that HP specimens are austenite at room temperature, while LP specimens are martensite at room temperature. After annealing, the phase transition temperature of HP specimens transitioned to the same range as LP. Therefore, the transition characteristics of nickel-titanium powder can be retained in the SLM-formed NiTi-SMAs specimens by selecting appropriate process parameters. Figure 13 shows the differential scanning calorimetry (DSC) curve of the specimen illustrating the start (Ms) and end (Mf) temperatures of martensite transformation and the start (As) and end (Af) temperatures of austenite transformation. Compared with the specimen without heat treatment, the specimen after 400 °C heat treatment has improved shape memory performance. With the use of this heat treatment condition, a high density of fine Ni₄Ti₃ metastable phase was produced. The specimens had the highest volume fraction of martensitic phase as well (Ref 173). Due to the effect of martensite stabilization, the reverse transformation temperature is different from the forward transformation (Ref 174).

Fig. 13

(Reproduced from Ref 173)

Fifth cycle of the differential scanning calorimetry testing of the (a) nonheat-treated, (b) heat-treated at 400 °C.

As shown in Fig. 14(a), the martensitic characteristics of the NiTi-SMAs are mainly caused by thermal stress. The formation of these characteristics can reduce the SLM residual elastic properties between laser scanning tracks, and it will reduce the unity of martensitic transformation, thus reducing the quality of shape memory response (Ref 175). The grain mainly grows in the direction of heat flow (SLM construction direction) (Ref 172), as shown in Fig. 14(b). At a higher magnification, fine martensitic features (Fig. 14c) and fine austenite subgrains (Fig. 14d) (Ref 176) can be observed.

Fig. 14

(Ref 176), reproduced with permission)

Microstructure characteristics of NiTi alloy parts formed by selective laser melting. (a) optical microscope image of SLM laser track containing martensite caused by thermal stress when laser thermal stress is high; (b) electron backscatter diffraction image of grain growth direction; (c) microscopic image of martensite characteristics; (d) microscopic image of austenite subgrain. Note: E, laser energy density; v, scan speed. (S. Dadbakhsh, M. Speirs, J. Van Humbeeck, J.P. Kruth, Laser Additive Manufacturing of Bulk and Porous Shape-memory NiTi Alloys: From Processes to Potential Biomedical Applications, MRS Bull., 41, 2016, p 765-774.

The shape memory response and mechanical properties of NiTi-SMAs with porous structure can be controlled by laser parameters. Dadbakhsh et al. (Ref 34) prepared octahedral porous scaffolds with different volume fractions and selected different parameters to affect the mechanical properties and shape memory response. As shown in Fig. 15, HP specimens need a higher load to break compared with LP specimens. This is mainly related to the larger volume fraction of porous scaffolds formed by HP. Bernard et al. (Ref 177) show that NiTi-SMAs can withstand 10⁶ compression fatigue cycles, and its stress is 1.4 times higher than the yield strength. Ma et al. (Ref 178) prepared NiTi-SMAs under different SLM process parameters and the obtained “U”-shaped components can achieve multistage deformation behavior. With the heating of the components, the reverse transformation from martensite to austenite occurs, resulting in two-stage shape change, in which different parts of the component activate the shape memory response at different temperature intervals and different rates, as shown in Fig. 16(a). As shown in the differential scanning calorimetry (DSC) of Fig. 16(b), the lower energy density process used on the left arm results in a lower phase transition temperature, and the shape recovery starts at about 0 °C. In contrast, the shape recovery in the right arm treated with higher energy density did not occur until 60 °C.

Fig. 15

(Reproduced from Ref 34, Copyright 2015, with permission from Elsevier)

Compressive stress–strain curves of the scaffolds made by (a) HP and (b) LP sets of parameters with different designed strut sizes (e.g., 180 mm) and actual solid volume fractions (%VF). Notice the scale difference.

Fig. 16

(Reproduced from Ref 178)

(a) Multistage shape recovery in a U-shaped additively manufactured NiTi build piece using selective laser melting. Two arms of the piece activate their shape recovery at different temperatures, creating a location-dependent active response; (b) the location-dependent active response is created by changing the SLM processing parameters (shown in the DSC curves on the right) at different sections of the build, which results in differences in the transformation temperatures in corresponding sections.

The adjustable range of transformation temperature and shape recovery of iron-based shape memory alloys are lower than those of NiTi-SMAs and copper-based shape memory alloys (Cu-SMAs). The Cu-SMAs not only have excellent thermal processing and shape memory effects but also have a low price. However, the traditional methods for forming Cu-SMAs have defects, such as unstable memory performance, poor machinability, coarse grains, and poor fatigue resistance (Ref 179). These defects hinder the research and application of Cu-SMAs, and the coarse grains were the main cause. The growth of grains can be inhibited because of the rapid heating and cooling effect of SLM. Therefore, SLM can be used to form high-performance Cu-SMAs, which has a broad application prospect. Gargarella et al. (Ref 180) confirmed the feasibility of forming Cu-SMAs (81.95Cu-11.85Al-3.2Ni-3Mn) by SLM. Although the larger stress–strain can be obtained, the relative density of the specimens reached 92% due to the existence of a large number of pores. Gustmann et al. (Ref 181) obtained 81.95Cu-11.85A-3.2Ni-3Mn alloys with relative density up to 99%, and the microstructure is mainly β-martensite. After that, Gustmann et al. (Ref 182) formed 81.95Cu-11.35Al-3.2Ni-3Mn-0.5Zr alloys through which Zr was added. The research showed that the addition of the Zr element can refine the grain and increase the phase transition temperature. Furthermore, Gustmann et al. (Ref 183) formed 81.95Cu-11.85Al-3.2Ni-3Mn by remelting. The relative density of specimens increased to 99.5 ± 0.3%, and the microstructure and phase transformation temperature can be changed by remelting. The mechanical properties of SMAs formed by SLM are shown in Table 8. SLM-fabricated SMAs have excellent tensile properties, and their compressive properties can reach a higher level.

Table 8 Mechanical properties of shape memory alloy

In general, there is no extra impurity in NiTi-SMAs formed by SLM. Compared with the traditional NiTi-SMAs, the NiTi-SMAs formed by SLM shows higher reversible strain, but their fracture strain and stress are lower. Therefore, it is important to further understand the evaporation, grain structure, texture, precipitation, customized transformation, and recovery behavior of NiTi-SMAs formed by SLM from the microscale. The phase transformation characteristics can be better controlled by controlling the process parameters of SLM. The corrosion resistance of SMAs has not been reported yet. In recent years, the rapid development of the intelligent system and the development of high-performance Cu-SMAs have significantly promoted the application of Cu-SMAs.

High-entropy Alloys

The customary definition of high-entropy alloys (HEAs) is that it is a system consisting of five or more main elements, each with a concentration between 5 and 35 at.%. The underlying mechanism of the HEAs is a minimization of the Gibbs free energy through a balance between entropy and enthalpy (Ref 188). HEAs have shown potentials for industrial applications due to many attractive properties, including high strength, high hardness, wear resistance, corrosion resistance, and other performance characteristics. As an emerging field of materials research, HEAs have high research value and broad application prospects.

The structure of HEAs after solidification is generally composed of the body-centered cubic phase (BCC) or the face-centered cubic phase (FCC). The HEAs containing Cu, Ti, Cr, Ni have excellent corrosion resistance because each corrosion-resistant component can play its corrosion resistance in the eutectic. Tian et al. (Ref 189) made statistics on the basic characteristics of more than 100 types of HEAs. The average valence electron concentration of BCC is between 4.33 and 7.55, of which FCC is between 7.80 and 9.50. Hardness varies with atomic size. The addition of elements with the larger atomic radius can promote the formation of BCC solid solution, but it may increase the brittleness. The addition of elements with smaller atomic radius helps to form FCC solid solution, but it may reduce the hardness. The maximum intensity can be obtained when the average valence electron number is about 6.5-7.0 and the average atomic size difference is about 6%. HEAs can adjust the content of elements at will to obtain a different crystal structure.

The relative density of HEAs formed by SLM still needs to be improved. For example, the maximum relative density obtained by Niu is 98.4% (Ref 190). Guo et al. (Ref 191) studied SLM-fabricated CoCrFeNiMn HEAs, and the relative density of the specimens reached 99.3%. Nonetheless, the surface quality and dimensional accuracy of CoCrFeMnNi HEAs formed by SLM cannot meet the requirements of industrial applications. AlCoCrFeNi HEAs contain BCC structure, BCC + FCC structure, and A2 + B2 structure at different cooling rates (Ref 192, 193). For engineering applications, FCC HEAs often show serious defects: insufficient yield strength and redundant ductility (Ref 194). Figure 17 shows the microstructure of AlCoCrFeNi HEAs formed by SLM. The forming process of SLM includes rapid solidification of the molten pool and reciprocating thermal cycle of multiple molten pools, resulting in elongated columnar grains extending into multiple layers. It can be seen from Fig. 17(b) and (c) that the A2 grain bundle epitaxial growth is perpendicular to the molten pool boundary, but the B2 phase is between the columnar A2 grains (Ref 190). Table 9 summarizes the phase structure of different types of HEAs. It can be found that the phase structure of the HEAs is FCC or BCC.

Fig. 17

(Reproduced from Ref 190, Copyright 2018, with permission from Elsevier)

SEM images showing the typical microstructure of SLM samples: (a) low magnification; (b) high magnification of (a); (c) and (d) high magnification of (b). The dashed lines highlight the melt pool boundaries.

Table 9 Phase structure of HEAs

The mechanical properties of SLM HEAs are excellent. As shown in Table 10, the tensile strength and yield strength are higher than casting, but the elongation needs to be improved. After heating treatment, the mechanical properties of HEAs formed by SLM are improved. Not only the tensile strength and yield strength have been improved, but also the elongation has been improved. Sun et al. (Ref 196) used mixed powder to form Al0.5CoCrFeNi HEAs and obtained excellent mechanical properties. It is found that the addition of a small amount of carbon is helpful to improve the mechanical properties of FeCoCrNiMn, AlCrFeMnNi, FeNiMnAl, CrCuNbTiY, and Al0.5CoCrCuFeNi HEAs (Ref 198). The addition of a small amount of carbon can lead to the transformation from dislocation slip-dominated plasticity to dislocation slip and winding mixed deformation mode (Ref 199). The increase in hardness is due to solution strengthening and carbide formation (Ref 200).

Table 10 Mechanical properties of high-entropy alloys

Zhang et al. (Ref 202) showed that the free corrosion current of NbMoTaW HEAs was 8.716 × 10-11 A, which was two orders of magnitude lower than that of 316L SS. Therefore, NbMoTaW HEAs have excellent corrosion resistance. Fujieda et al. (Ref 201) also concluded that the corrosion resistance of the CoCrFeNiTi HEAs formed by SLM is better than that of the traditional corrosion resistant alloys (such as duplex stainless steel and nickel-based superalloy).

Some scholars have used SLM to form HEAs, but there is no mature theory. Cracks, porosity, and other defects easily appear in the process of SLM. It is still necessary to further study the forming mechanism and control measures through process improvement and other methods. The high-temperature creep behavior of HEAs formed by SLM has not been reported. The type and content of the elements have a great influence on the structure and properties of the HEAs. It is necessary to study the effect of element content on the microstructure of the alloys to control the phase composition and mechanical properties of the HEAs.

Summary and Prospect

The densification of specimens formed by SLM can be improved by adjusting the process parameters. However, the optimal process parameters can be different due to the different optimal forming areas of different equipment. The changing trend of process parameters and the effect of heat treatment are available for reference, although the best process parameters of different equipment are different. It is still difficult to guarantee the high precision and relative density of the specimens, although considerable progress has been made in SLM. The laser absorptivity of the powder is different, so the forming difficulty and forming performance are different. The volatilization of volatile elements in materials leads to thermal radiation and thermal convection (the interaction between volatile elements and protective gases). The heat loss is caused by thermal radiation and thermal convection. It is easy to cause forming defects because of multiple physical phenomena and improper selection of process parameters. It is difficult to form a fully dense specimen due to the existence of micropores. At present, the amount of materials database of SLM is relatively less than that of reducing material processing, and only a small amount of powder has developed mature process parameters. The doping, process optimization, and heat treatment are mainly used to improve the consistency of the microstructure and improve the mechanical properties.

The manufacturers are committed to developing excellent materials. The development of powder with stronger mechanical properties and better forming properties is essential for the application of SLM. The crystal grains in the alloys with a completely equiaxed crystal structure grow uniformly in all directions to form a strong bond. Specimens with fully equiaxed crystal can withstand greater forces, and cracks or deformation defects that are easy to occur during the forming process will be significantly reduced. Therefore, one of the development directions of SLM in the future is to develop powder that can form completely equiaxed grains to improve the performance of specimens.

The surface quality and forming accuracy of SLM specimens can be improved by post-treatment. Sandblasting and high-temperature calcination can improve the forming quality, but the post-treatment process is complex. The high surface quality of the specimens can be manufactured directly by using laser cutting technology. The new forming process by combining SLM with laser cutting can not only avoid the generation of stress in mechanical processing but also process the fine structure which cannot bear the cutting force of the tool. Therefore, the use of laser composite processing technology can quickly and efficiently produce precision parts.

The forming scheme of SLM based on big data and artificial intelligence can ensure the real-time monitoring and correction of the forming process. The temperature stability of the molten pool is an important index of processing stability. The effective way to improve the geometric accuracy, mechanical properties, and microstructure of the specimens is to accurately control the temperature of the molten pool. The stability of the forming process can be guaranteed by monitoring of the forming process in real time. To ensure the accuracy and efficiency of SLM, it is also necessary to develop the methods of the molten pool morphology monitoring, acoustic monitoring, composition monitoring, and optical radiation monitoring.

Stainless steel, cobalt-based alloys, SMAs, and TBAs are commonly used medical metal materials. TBAs are the most representative human implant material because of their nontoxic, nonmagnetic, corrosion resistance, high strength, good toughness, low elastic modulus, and other excellent properties. TBAs implants formed by SLM can not only meet the requirements of biosafety but also be customized according to patients’ conditions. The implant can overcome the problem that the shape of traditional universal implants is not compatible with the human body and mechanical properties are not up to standard. The research and application of customized implant prostheses and implant teeth are more and more extensive, and the biological safety of human implants is becoming more and more important. Human implants need to meet the requirements of biocompatibility and biological function. At present, the molding accuracy and quality of SLM are not excellent due to the limitation of powder raw materials, process level, equipment conditions, and post-treatment. With the continuous optimization and improvement in material technology, information technology, and control technology, the process of SLM will become more and more sophisticated and high end, and the human implants produced will meet the human needs gradually.

Intelligent 4D printing is also an important development direction. 4D printing is developing in the direction of fast response, intelligent forming materials, and precise forming of composite materials. The phase transition of SMAs must be controlled accurately to realize high intelligence and precision forming. It is necessary to fully consider the conditions to be achieved after forming when designing the SMAs structure. The shape memory effect of the special structure needs to be considered when forming the specimens with a complex structure (such as overhang structure and lattice structure). The application of SMAs is mainly to control its deformation, and the deformation mainly depends on the regulation of tissue properties. Therefore, it is necessary to effectively control the distribution and growth of cell crystals, dendrites, equiaxed crystals, etc., through process control or the addition of nanoenhancement phases to improve its overall performance.