*3.2. Study of Droplet Spatter Ejected from "Liquid Base" of Melt Pool*

The melt pool is a critical feature of L-PBF. Numerous studies on the spattering from the melt pool have been done using a numerical simulation, which avoided the high cost and inefficiency of repeated experiments. Khairallah et al. [68] studied the mechanism of spatter generation at the powder scale using a 3D high-precision model. The metal vapor exerted pressure on the melt pool during L-PBF, causing the emission of liquid metal. When the liquid metal was stretched, the column grew thinner and decomposed into tiny droplets because the surface tension tended to minimize the surface energy. Additionally, it was discovered that at the start of the scanning, it was rather easy to generate large-sized back-ejected spatters [69]. They assumed that the laser scanning velocity could not be kept constant at the beginning and end of the trajectory due to inertia, resulting in a deposition of a nonuniform energy density and causing such spatters. They proposed a stability criterion to eliminate back-ejected spatter effectively. Altmeppen et al. [70] proposed a method to simulate time-dependent particles and heat ejection from the moving melt pool. This model can predict the direction and velocity of spatter emission and determine the size and temperature of a single particle by evaluating the direction and velocity of local laser scanning.

In order to verify the intrinsic mechanism of the spatter generation, experiments were applied to detect the spatter using X-ray imaging and high-speed imaging. The explosion caused by the instability of the front wall of the keyhole, which resulted from the vaporization of the L-PBF volatile element, induced much droplet spatter. Zhao et al. used X-ray imaging to study the spatter behavior of Ti-6Al-4V powder during L-PBF. As illustrated in Figure 10, they demonstrated how the bulk-explosion induced by the instability of the front wall of the keyhole in the melt pool resulted in a considerable

amount of droplet spatter [71]. Using in situ high-speed high-resolution imaging and thermodynamic analysis, Yin et al. investigated the vaporization and explosion behavior of alloy components in a Cu-10Zn alloy during L-PBF [72]. It was found that the explosion caused by the violent vaporization of a low boiling point also induced much droplet spatter and defects in the melt track. spatter [71]. Using in situ high-speed high-resolution imaging and thermodynamic analysis, Yin et al. investigated the vaporization and explosion behavior of alloy components in a Cu-10Zn alloy during L-PBF [72]. It was found that the explosion caused by the violent vaporization of a low boiling point also induced much droplet spatter and defects in the melt track.

imaging to study the spatter behavior of Ti-6Al-4V powder during L-PBF. As illustrated in Figure 10, they demonstrated how the bulk-explosion induced by the instability of the front wall of the keyhole in the melt pool resulted in a considerable amount of droplet

*Micromachines* **2022**, *13*, x FOR PEER REVIEW 14 of 42

**Figure 10.** MHz X-ray images of metal spattering of Ti-6Al-4V during laser processing. **Event No. 01 (sky blue dashed rectangles):** A protrusion forms at the top surface and runs down along the front keyhole wall, accompanied by the keyhole morphology changing from a J-like shape to a reverse-triangle-like shape. **Event No. 02 (purple dashed rectangles):** A following protrusion appears, grows, and collapses around the horizontal center of the keyhole. A mini keyhole on top of the protrusion is outlined by a light yellow dashed curve. **Event No. 03 (dark blue arrows):** The local curvature on the rear keyhole wall changes. **Event No. 04 (light green dashed and solid rectangles):** Melt ligaments form, elongate, and break up into spatters (light green dashed circles numbered SP01–SP05). **Event KP (sky blue solid rectangles):** describes the formation and vanishing of a keyhole pore (Reprinted with permission from Ref. [71]. Copyright 2019 APS Physics). **Figure 10.** MHz X-ray images of metal spattering of Ti-6Al-4V during laser processing. **Event No. 01 (sky blue dashed rectangles):** A protrusion forms at the top surface and runs down along the front keyhole wall, accompanied by the keyhole morphology changing from a J-like shape to a reverse-triangle-like shape. **Event No. 02 (purple dashed rectangles):** A following protrusion appears, grows, and collapses around the horizontal center of the keyhole. A mini keyhole on top of the protrusion is outlined by a light yellow dashed curve. **Event No. 03 (dark blue arrows):** The local curvature on the rear keyhole wall changes. **Event No. 04 (light green dashed and solid rectangles):** Melt ligaments form, elongate, and break up into spatters (light green dashed circles numbered SP01–SP05). **Event KP (sky blue solid rectangles):** describes the formation and vanishing of a keyhole pore (Reprinted with permission from Ref. [71]. Copyright 2019 APS Physics).

Using high-speed and high-resolution imaging technologies, Yin et al. [41] investi-

gated the spatter behavior of Inconel 718 powder during L-PBF. The subthreshold ejection phenomenon was detected in which droplets emitted from the droplet column fell back to the melt pool. Later, the authors also studied the correlation between the ex situ melt track characteristics and the in situ high-speed and high-resolution characterization. They showed that the protrusion of the head of the melt trajectory was caused by the combined action of the backward flowing melt and the droplet ejection behavior in the melt pool [34]. Moreover, as illustrated in Figure 11, the melt pool first forms a depression under the action of the recoil pressure of the vapor; a high-energy laser beam impinges on the front wall of the depression, causing the surface of the front wall to quickly vaporize and generate a metal vapor that is perpendicular to this surface; the metal vapor expands and Using high-speed and high-resolution imaging technologies, Yin et al. [41] investigated the spatter behavior of Inconel 718 powder during L-PBF. The subthreshold ejection phenomenon was detected in which droplets emitted from the droplet column fell back to the melt pool. Later, the authors also studied the correlation between the ex situ melt track characteristics and the in situ high-speed and high-resolution characterization. They showed that the protrusion of the head of the melt trajectory was caused by the combined action of the backward flowing melt and the droplet ejection behavior in the melt pool [34]. Moreover, as illustrated in Figure 11, the melt pool first forms a depression under the action of the recoil pressure of the vapor; a high-energy laser beam impinges on the front wall of the depression, causing the surface of the front wall to quickly vaporize and generate a metal vapor that is perpendicular to this surface; the metal vapor expands and impacts the rear wall of the depression; finally, the spatter is formed and ejected backwards. The vertical metal vapor plume was identified as the principal reason for the melt pool spattering.

Through in situ measurements of a typical forward spatter ejection angle, the vapor recoil pressure (approximately 0.46 atm) was quantified. various phenomena can be identified including 3D dynamics of melt pools, vapor plume dynamics, and spatter generation.

impacts the rear wall of the depression; finally, the spatter is formed and ejected backwards. The vertical metal vapor plume was identified as the principal reason for the melt pool spattering. Through in situ measurements of a typical forward spatter ejection angle,

The development of various advanced in situ characterization methods provides new directions for spatter research. Wang et al. [48] used a high-speed camera to investigate the characteristics of the droplet spatter of 316L stainless steel powder during L-PBF process. Gould et al. [73] reported an in situ method to analyze the L-PBF process of Ti-6Al-4V and W powders by using high-speed X-ray and high-speed infrared imaging simultaneously. Combining both imaging of high-speed X-rays and high-speed infrared imaging,

*Micromachines* **2022**, *13*, x FOR PEER REVIEW 15 of 42

the vapor recoil pressure (approximately 0.46 atm) was quantified.

**Figure 11.** Schematic of the correlation between the depression zone in melt pool and the backwardejected spatter during L-PBF. The inclined angles for the normal direction of the front depression wall *θFW* and the inclination angle of the rear depression wall *θRW* had the same trends with the average angle of the backward-ejected spatter (*θBE*). (Reprinted with permission from Ref. [34]. Copyright 2019 Elsevier B.V.). **Figure 11.** Schematic of the correlation between the depression zone in melt pool and the backwardejected spatter during L-PBF. The inclined angles for the normal direction of the front depression wall *θFW* and the inclination angle of the rear depression wall *θRW* had the same trends with the average angle of the backward-ejected spatter (*θBE*). (Reprinted with permission from Ref. [34]. Copyright 2019 Elsevier B.V.).

Surface tension and evaporation both have a noticeable effect on the melt pool. Dai et al. [74] studied the process parameters of the thermal behavior, fluid dynamics, and surface morphology in a melt pool using a mesoscopic simulation model. The results indicated that the evolution of the melt pool was highly sensitive to the melt viscosity, surface tension, and recoil pressure during L-PBF. Bärtl et al. [75] investigated the ability of the aluminum alloy powder materials Al-Cr-Zr-Mn, Al-Cr-Sc-Zr, and Al-Mg-Sc-Mn-Zr to produce lightweight and high-performance structures by L-PBF. They regarded that both The development of various advanced in situ characterization methods provides new directions for spatter research. Wang et al. [48] used a high-speed camera to investigate the characteristics of the droplet spatter of 316L stainless steel powder during L-PBF process. Gould et al. [73] reported an in situ method to analyze the L-PBF process of Ti-6Al-4V and W powders by using high-speed X-ray and high-speed infrared imaging simultaneously. Combining both imaging of high-speed X-rays and high-speed infrared imaging, various phenomena can be identified including 3D dynamics of melt pools, vapor plume dynamics, and spatter generation.

the surface tension and evaporation were potentially crucial factors dominating the melt dynamics, and the melt dynamics of materials with a lower surface tension and less evaporation were the most unstable. Table 4 summarizes the research on droplet spatter ejected from the "liquid" base of the melt pool. **Table 4.** Summary of research on droplet spatter ejected from the "liquid base" of the melt pool. **Generation Mechanism Materials References**  Surface tension Ti-6Al-4V, TiC Dai et al. (2020) [74] Al-Cr-Zr-Mn, Al-Cr-Sc-Zr & Al-Mg-Sc-Mn-Zr Bärtl et al. (2022) [75] Surface tension and evaporation both have a noticeable effect on the melt pool. Dai et al. [74] studied the process parameters of the thermal behavior, fluid dynamics, and surface morphology in a melt pool using a mesoscopic simulation model. The results indicated that the evolution of the melt pool was highly sensitive to the melt viscosity, surface tension, and recoil pressure during L-PBF. Bärtl et al. [75] investigated the ability of the aluminum alloy powder materials Al-Cr-Zr-Mn, Al-Cr-Sc-Zr, and Al-Mg-Sc-Mn-Zr to produce lightweight and high-performance structures by L-PBF. They regarded that both the surface tension and evaporation were potentially crucial factors dominating the melt dynamics, and the melt dynamics of materials with a lower surface tension and less evaporation were the most unstable. Table 4 summarizes the research on droplet spatter ejected from the "liquid" base of the melt pool.

Inconel 718 Yin et al. (2019) [41]

Vapor recoil pressure 316L Khairallah et al. (2016) [68]


**Table 4.** Summary of research on droplet spatter ejected from the "liquid base" of the melt pool.

#### *3.3. Study of Powder Spatter Ejected from "Solid Base" of Substrate*

Due to the entrainment effect of the gas flow, powder particles close to the laser zone of action are ejected and spattered. Ly et al. [64] performed an experimental comparison of the melt pool hydrodynamics of laser welding and L-PBF processes. In contrast to laser welding, the primary cause of spatter in L-PBF was not the laser-induced recoil pressure, but the entrainment effect of the ambient gas flow driven by the metal vapor on the microparticles. The high-speed X-ray video imaging of the defects and melt pool performed by Leung et al. [76] supported the Ly et al. hypothesis about the generation of cold and hot entrainment spatter during L-PBF. Chen et al. [77] built a multi-phase flow model to investigate the spatter generation during L-PBF. The spatter phenomena were shown to be the result of metal vapor- and ambient gas-induced entrainment, which supported the findings of Ly et al. [64].

Gunenthiram et al. [78] used high-speed camera techniques to investigate the dynamic behavior of 316L stainless steel powder and 4047 aluminum–silicon alloy powder during the generation of spatter in L-PBF. As shown in Figure 12 [61], due to the heat transfer from the surrounding powder bed, the powder particles in close contact with the front and sides of the melt pool tended to agglomerate to form larger droplets. Some of the agglomerates were subject to an entrainment gas flow, which in turn were ejected as spatter. To establish the correlation between the scanning velocity and spatter generation, Zheng et al. [51] used a high-speed camera technique to investigate the effect of the scanning velocity on the generation and evolution of the metal vapor plumes during L-PBF of 304 stainless steel powder. The results indicated that the powder spatter generations are more closely related with the stability/evolution of the vapor plume and resulting melt-track, rather than the changing of the volumetric energy density (VED). The trend of an increasing number of spatters with an increasing VED was reported by Gunenthiram et al. [78]. The droplet spatter generated at the commencement of the scan trajectory was found to be the consequence of coupling between the melt pool and the inclined metal vapor plume. Table 5 summarizes the studies of the spatter from the solid substrate ejection.

**Table 5.** A summary of the studies on spatter from solid substrate ejection.


**Figure 12.** Schematic of spatter ejection process and interaction between vapor plume and spatter behavior in L-PBF. (Reprinted with permission from Ref. [61]. Copyright 2022 Chinese laser press.). **Figure 12.** Schematic of spatter ejection process and interaction between vapor plume and spatter behavior in L-PBF. (Reprinted with permission from Ref. [61]. Copyright 2022 Chinese laser press).

#### *3.4. Study of Spatter Generation Mechanism in Multi-Laser-PBF Fabrication Process*

*3.4. Study of Spatter Generation Mechanism in Multi-Laser-PBF Fabrication Process*  Recently, a multi-laser beam based on L-PBF has been applied to fulfil the growing demand for large-sized part manufacturing in aerospace and energy fields. Andani et al. [79] investigated the spatter behavior of Al-Si10-Mg powder during dual-beam L-PBF using a high-speed camera technique. They showed that the number of operating laser beams significantly influences the spatter creation mechanisms during the SLM process. A higher number of working laser beams induces a greater recoil pressure above the melt-Recently, a multi-laser beam based on L-PBF has been applied to fulfil the growing demand for large-sized part manufacturing in aerospace and energy fields. Andani et al. [79] investigated the spatter behavior of Al-Si10-Mg powder during dual-beam L-PBF using a high-speed camera technique. They showed that the number of operating laser beams significantly influences the spatter creation mechanisms during the SLM process. A higher number of working laser beams induces a greater recoil pressure above the melting pools and ejects a larger amount of metallic material from the melt pools. However, there was no description of the interaction between the dual-beam laser and the material in the overlap region.

ing pools and ejects a larger amount of metallic material from the melt pools. However, there was no description of the interaction between the dual-beam laser and the material in the overlap region. The mechanism by which a dual-beam laser generates spatter is distinct from that of The mechanism by which a dual-beam laser generates spatter is distinct from that of a single-beam laser. Yin et al. [80] investigated the interaction between dual-beam lasers and the material in the overlap region during the dual-beam L-PBF of Inconel 718 alloy powder using a high-speed, high-resolution video imaging system. They proposed to use the spatter growth rate (*rs*) to quantitatively characterize the spatter behavior in multi-laser powder bed fusion (ML-PBF).

a single-beam laser. Yin et al. [80] investigated the interaction between dual-beam lasers and the material in the overlap region during the dual-beam L-PBF of Inconel 718 alloy powder using a high-speed, high-resolution video imaging system. They proposed to use the spatter growth rate (*rs*) to quantitatively characterize the spatter behavior in multilaser powder bed fusion (ML-PBF). According to experimental observations, Yin et al. [80] believe that most of the spatter in multi-laser L-PBF is due to metal vapor-induced entrainment (ejected from the "solid According to experimental observations, Yin et al. [80] believe that most of the spatter in multi-laser L-PBF is due to metal vapor-induced entrainment (ejected from the "solid baes" of the substrate) rather than the metal vapor recoil pressure (ejected from the "liquid baes" of the melt pool). In fact, the *rs* in the vapor entrainment dominant stages is one order of magnitude higher than that in the unstable melt pool dominant stage disturbed by the recoil pressure and the collision of the two melt pools. This proves that the entrainment effect is dominant in the cause of the multi-laser-PBF spatter, as shown in Figure 13. A summary of the studies on the mechanism of the spatter generation during an ML-PBF process is shown in Table 6.

baes" of the substrate) rather than the metal vapor recoil pressure (ejected from the "liquid baes" of the melt pool). In fact, the *rs* in the vapor entrainment dominant stages is one

by the recoil pressure and the collision of the two melt pools. This proves that the entrainment effect is dominant in the cause of the multi-laser-PBF spatter, as shown in Figure 13. A summary of the studies on the mechanism of the spatter generation during an ML-PBF

process is shown in Table 6.

**Figure 13.** Schematic diagram of the transformation of the main mechanisms of spatter generation, which changes from (**a**) the vapor-induced recoil pressure with an almost homogeneous distribution of spatter ejection angle, into (**b**) the vapor-induced entrainment that majority of spatters eject along the direction of the metal vapor propagation. (Reprinted with permission from Ref. [80]. Copyright 2021 Elsevier B.V.). **Figure 13.** Schematic diagram of the transformation of the main mechanisms of spatter generation, which changes from (**a**) the vapor-induced recoil pressure with an almost homogeneous distribution of spatter ejection angle, into (**b**) the vapor-induced entrainment that majority of spatters eject along the direction of the metal vapor propagation. (Reprinted with permission from Ref. [80]. Copyright 2021 Elsevier B.V.).

**Table 6.** A summary of the studies on the mechanism of spatter generation during ML-PBF process. **Table 6.** A summary of the studies on the mechanism of spatter generation during ML-PBF process.


#### **4. Disadvantage of Spatter**  Spatter is an unpreventable by-product of the complex heat transfer process between **4. Disadvantage of Spatter**

the laser and the metal powder in L-PBF [20,30,54]. Spatter brings a negative influence to the process stability and the efficiency of the energy, which reduces the quality of the manufactured object and can potentially damage the machine [68]. In accordance with the current research, the disadvantages posed by spatter in L-PBF can be classified into three categories: (1) The effect of spatter on the printing processing: spatter can affect the powder re-coating in the next layer, and reduce the energy input efficiency of the laser and the operation stability of the powder re-coating device [63,81] as well as the optical lens. (2) The effect of spatter on structure and performance: spatter is not conducive to controlling the structure (e.g., voids, roughness) and performance (e.g., tensile properties, oxygen contents) of printed parts. (3) The effect of spatter on powder recycling: recycled powder can entrain spatter particles, resulting in a significant deterioration of powder quality. The use of recycled powder for forming parts can lead to a reduction in part performance. *4.1. Effect of Spatter on Printing Processing*  Spatter is an unpreventable by-product of the complex heat transfer process between the laser and the metal powder in L-PBF [20,30,54]. Spatter brings a negative influence to the process stability and the efficiency of the energy, which reduces the quality of the manufactured object and can potentially damage the machine [68]. In accordance with the current research, the disadvantages posed by spatter in L-PBF can be classified into three categories: (1) The effect of spatter on the printing processing: spatter can affect the powder re-coating in the next layer, and reduce the energy input efficiency of the laser and the operation stability of the powder re-coating device [63,81] as well as the optical lens. (2) The effect of spatter on structure and performance: spatter is not conducive to controlling the structure (e.g., voids, roughness) and performance (e.g., tensile properties, oxygen contents) of printed parts. (3) The effect of spatter on powder recycling: recycled powder can entrain spatter particles, resulting in a significant deterioration of powder quality. The use of recycled powder for forming parts can lead to a reduction in part performance.

#### According to the generation mechanism of spatter, it can be found that spatter has a *4.1. Effect of Spatter on Printing Processing*

negative influence on powder re-coating and energy absorption during L-PBF processing. 4.1.1. Effect of Spatter on Powder Re-Coating According to the generation mechanism of spatter, it can be found that spatter has a negative influence on powder re-coating and energy absorption during L-PBF processing.

#### Spatter particles that redeposit onto the powder bed hinder the powder re-coating, 4.1.1. Effect of Spatter on Powder Re-Coating

and voids between the spatter particles and powder can induce part defects. Figure 14 shows how spatter generated during L-PBF introduces voids and internal defects in the printed part. Wang et al. [63] discovered that the re-coating powders were influenced by Spatter particles that redeposit onto the powder bed hinder the powder re-coating, and voids between the spatter particles and powder can induce part defects. Figure 14 shows how spatter generated during L-PBF introduces voids and internal defects in the printed

the spatter particles due to a small amount of spatter attached to the surface of the printed

part. Wang et al. [63] discovered that the re-coating powders were influenced by the spatter particles due to a small amount of spatter attached to the surface of the printed parts during stacking, and the spatter particles caused the deformation of the scraper (Figure 14a). When the redeposited spatter particles are smaller than the layer thickness, after laser scanning, the spatter particles melted completely and were metallurgically bonded to the powder and the underlying part. If the redeposited spatter particles' size exceeded the layer thickness, they did not melt completely, which induced voids between the powder and the spatter particles, as illustrated in Figure 14b. The voids remained after the scanning of the next layer, creating metallurgical defects, as illustrated in Figure 14c. Schwerz et al. [82] found the presence of spatter particles of approximately 136 µm in the cross-section of the part, illustrating how particles significantly larger than the nominal layer thickness were incorporated into the material despite recoating, and in the process, large spatter bumps of particles can cause damage to the scraper, as shown in Figure 14d. parts during stacking, and the spatter particles caused the deformation of the scraper (Figure 14a). When the redeposited spatter particles are smaller than the layer thickness, after laser scanning, the spatter particles melted completely and were metallurgically bonded to the powder and the underlying part. If the redeposited spatter particles' size exceeded the layer thickness, they did not melt completely, which induced voids between the powder and the spatter particles, as illustrated in Figure 14b. The voids remained after the scanning of the next layer, creating metallurgical defects, as illustrated in Figure 14c. Schwerz et al. [82] found the presence of spatter particles of approximately 136 µm in the cross-section of the part, illustrating how particles significantly larger than the nominal layer thickness were incorporated into the material despite recoating, and in the process, large spatter bumps of particles can cause damage to the scraper, as shown in Figure 14d.

*Micromachines* **2022**, *13*, x FOR PEER REVIEW 19 of 42

**Figure 14.** The effect of spatter on re-coating powder. (**a**) Pre-placing powders will be blocked when the next powder layer spreads; (**b**) spattering particles impede circulation of powders and cause voids nearby (before laser melting); (**c**) laser scanning can completely melt small spatter to achieve metallurgical bonding, while large spatter can only melt some of them. (Reprinted with permission from Ref. [63]. Copyright 2016 Elsevier B.V.). (**d**) A spatter particle of cross-section ~136 µm incites a solidification front (schematized with white arrows) that competes with the solidification fronts in the melt pool (schematized with black arrows). (Reprinted with permission from Ref. [82]. Copyright 2021 Elsevier B.V.). **Figure 14.** The effect of spatter on re-coating powder. (**a**) Pre-placing powders will be blocked when the next powder layer spreads; (**b**) spattering particles impede circulation of powders and cause voids nearby (before laser melting); (**c**) laser scanning can completely melt small spatter to achieve metallurgical bonding, while large spatter can only melt some of them. (Reprinted with permission from Ref. [63]. Copyright 2016 Elsevier B.V.). (**d**) A spatter particle of cross-section ~136 µm incites a solidification front (schematized with white arrows) that competes with the solidification fronts in the melt pool (schematized with black arrows). (Reprinted with permission from Ref. [82]. Copyright 2021 Elsevier B.V.).

In order to detect the distribution of the re-deposition of the spatters on the build

area, a long-exposure near-infrared in situ monitoring associated with image analysis was employed to determine the exact locations using the EOS EOSTATE Exposure OT system [82]. This system consists of a 5-megapixel sCMOS (scientific complementary metal-oxidesemiconductor) camera positioned on top of the build chamber and comprises the entire build platform area in its field of view. A bandpass filter of 900 nm ± 12.5 nm is placed on the camera to filter the detection of the reflected laser to avoid the detection of the environmental noise. A sample image representative of a single layer can be observed in Figure 15a, samples near the gas inlet (Figure 15b) and gas outlet (Figure 15c) are shown separately. The long-exposure images revealed deviations in the form of high-intensity spots preferentially distributed towards the gas outlet, as in Figure 15c, the re-deposition spatter can be extracted by algorithms (Figure 15d). The spatter deposited near the gas outlet has been identified as one of the factors responsible for the rise of internal defects, which will be discussed in Section 4.2. In order to detect the distribution of the re-deposition of the spatters on the build area, a long-exposure near-infrared in situ monitoring associated with image analysis was employed to determine the exact locations using the EOS EOSTATE Exposure OT system [82]. This system consists of a 5-megapixel sCMOS (scientific complementary metaloxide-semiconductor) camera positioned on top of the build chamber and comprises the entire build platform area in its field of view. A bandpass filter of 900 nm ± 12.5 nm is placed on the camera to filter the detection of the reflected laser to avoid the detection of the environmental noise. A sample image representative of a single layer can be observed in Figure 15a, samples near the gas inlet (Figure 15b) and gas outlet (Figure 15c) are shown separately. The long-exposure images revealed deviations in the form of high-intensity spots preferentially distributed towards the gas outlet, as in Figure 15c, the re-deposition spatter can be extracted by algorithms (Figure 15d). The spatter deposited near the gas outlet has been identified as one of the factors responsible for the rise of internal defects, which will be discussed in Section 4.2.

**Figure 15.** The results obtained by the monitoring system in conjunction with the spatter detection algorithm. (**a**) Sample long-exposure image consisting of the signals emitted during the exposure of a single layer on the entire build area. (**b**) A sample area near the gas inlet without any identified disturbances is highlighted for comparison. (**c**) Areas with disturbances are observed preferentially near the gas outlet. (**d**) A sample output from the spatter detection algorithm, in which the region shown in (**c**) is overlayed with detections. (Reprinted with permission from Ref. [82]. Copyright 2021 Elsevier B.V.). 4.1.2. Effect of Spatter on Energy Absorption **Figure 15.** The results obtained by the monitoring system in conjunction with the spatter detection algorithm. (**a**) Sample long-exposure image consisting of the signals emitted during the exposure of a single layer on the entire build area. (**b**) A sample area near the gas inlet without any identified disturbances is highlighted for comparison. (**c**) Areas with disturbances are observed preferentially near the gas outlet. (**d**) A sample output from the spatter detection algorithm, in which the region shown in (**c**) is overlayed with detections. (Reprinted with permission from Ref. [82]. Copyright 2021 Elsevier B.V.). **Figure 15.** The results obtained by the monitoring system in conjunction with the spatter detection algorithm. (**a**) Sample long-exposure image consisting of the signals emitted during the exposure of a single layer on the entire build area. (**b**) A sample area near the gas inlet without any identified disturbances is highlighted for comparison. (**c**) Areas with disturbances are observed preferentially near the gas outlet. (**d**) A sample output from the spatter detection algorithm, in which the region shown in (**c**) is overlayed with detections. (Reprinted with permission from Ref. [82]. Copyright 2021 Elsevier B.V.).

#### If spatter occurs in the laser path, it might result in an inefficient use of laser energy. 4.1.2. Effect of Spatter on Energy Absorption 4.1.2. Effect of Spatter on Energy Absorption

Several studies have been done on the influence of spatter on the energy required to melt the powder. Ferrar et al. [83] first reported on the influence of gas flow on L-PBF in 2012. They demonstrated that by-products of processing in the laser path could absorb and scatter the laser beam, inducing laser beam attenuation and the generation of a lack of fusion. Anwar et al. [84] came to a similar conclusion in the selective laser melting of Al-Si10-Mg, implying that laser energy might be squandered on spatter, as shown in Figure 16. The laser beam irradiated the spatter particles that entered the beam path and consumed a significant amount of energy, which induced the incomplete melting of the powder and defects [85]. The accumulated spatter in the powder bed inevitably consumed the energy required to melt the fresh powder [86]. If spatter occurs in the laser path, it might result in an inefficient use of laser energy. Several studies have been done on the influence of spatter on the energy required to melt the powder. Ferrar et al. [83] first reported on the influence of gas flow on L-PBF in 2012. They demonstrated that by-products of processing in the laser path could absorb and scatter the laser beam, inducing laser beam attenuation and the generation of a lack of fusion. Anwar et al. [84] came to a similar conclusion in the selective laser melting of Al-Si10-Mg, implying that laser energy might be squandered on spatter, as shown in Figure 16. The laser beam irradiated the spatter particles that entered the beam path and consumed a significant amount of energy, which induced the incomplete melting of the powder and defects [85]. The accumulated spatter in the powder bed inevitably consumed the energy required to melt the fresh powder [86]. If spatter occurs in the laser path, it might result in an inefficient use of laser energy. Several studies have been done on the influence of spatter on the energy required to melt the powder. Ferrar et al. [83] first reported on the influence of gas flow on L-PBF in 2012. They demonstrated that by-products of processing in the laser path could absorb and scatter the laser beam, inducing laser beam attenuation and the generation of a lack of fusion. Anwar et al. [84] came to a similar conclusion in the selective laser melting of Al-Si10-Mg, implying that laser energy might be squandered on spatter, as shown in Figure 16. The laser beam irradiated the spatter particles that entered the beam path and consumed a significant amount of energy, which induced the incomplete melting of the powder and defects [85]. The accumulated spatter in the powder bed inevitably consumed the energy required to melt the fresh powder [86].

condensate is the product of vaporized metal that quickly cools and condenses. (Reprinted with permission from Ref. [87]. Copyright 2016 Elsevier B.V.). **Figure 16.** Spatter and other by-products pass through the laser stream and squander laser energy, condensate is the product of vaporized metal that quickly cools and condenses. (Reprinted with permission from Ref. [87]. Copyright 2016 Elsevier B.V.). **Figure 16.** Spatter and other by-products pass through the laser stream and squander laser energy, condensate is the product of vaporized metal that quickly cools and condenses. (Reprinted with permission from Ref. [87]. Copyright 2016 Elsevier B.V.).

#### *4.2. Effect of Spatter on Structure and Performance* Spatter causes a loss of laser energy, moreover, spatter re-deposition and oxidation

*4.2. Effect of Spatter on Structure and Performance* 

*Micromachines* **2022**, *13*, x FOR PEER REVIEW 21 of 42

Spatter causes a loss of laser energy, moreover, spatter re-deposition and oxidation also have an effect on the quality and structure of parts. A coating of oxide is generated on the spatter surface after L-PBF and the oxide layer greatly reduces the humidity of the liquid metal, which induces spheroidization [88,89]. The particles with an oxidized surface require more energy for melting and incorporation in the melt pool and in the bulk material, resulting in a lack of fusion [82]. The seriously oxidized spatter particles redeposit into the high-temperature melt pool, reversing the Marangoni convection flow direction [90,91]. Additionally, the oxidized spatter particles in the melt pool induce holes and defects [88,92]. The oxide composition of Inconel 718 spatter particles was evaluated by SEM-EDS by Gasper et al., as shown in Figure 17. In order to determine the extent of the oxidation of the spatter particles, a particle with oxide spots and fully oxidized particles were also analyzed by SEM-EDS with an in situ Focused Ion Beam (FIB), as shown in Figure 18. also have an effect on the quality and structure of parts. A coating of oxide is generated on the spatter surface after L-PBF and the oxide layer greatly reduces the humidity of the liquid metal, which induces spheroidization [88,89]. The particles with an oxidized surface require more energy for melting and incorporation in the melt pool and in the bulk material, resulting in a lack of fusion [82]. The seriously oxidized spatter particles redeposit into the high-temperature melt pool, reversing the Marangoni convection flow direction [90,91]. Additionally, the oxidized spatter particles in the melt pool induce holes and defects [88,92]. The oxide composition of Inconel 718 spatter particles was evaluated by SEM-EDS by Gasper et al., as shown in Figure 17. In order to determine the extent of the oxidation of the spatter particles, a particle with oxide spots and fully oxidized particles were also analyzed by SEM-EDS with an in situ Focused Ion Beam (FIB), as shown in Figure 18.

**Figure 17.** Back-scattered electron micrograph, and electron X-ray dispersive spectroscopy mapping of elements of Inconel 718 spatter collected from the ReaLizer SLM50. (**a**–**c**) shows that the dark spots mostly contain Al and O, and that the larger dark spots also contain Ti. EDS quantification results indicated that the oxides were a combination of Al2O3 and TiO2. (Reprinted with permission from Ref. [66]. Copyright 2018 Elsevier B.V.). **Figure 17.** Back-scattered electron micrograph, and electron X-ray dispersive spectroscopy mapping of elements of Inconel 718 spatter collected from the ReaLizer SLM50. (**a**–**c**) shows that the dark spots mostly contain Al and O, and that the larger dark spots also contain Ti. EDS quantification results indicated that the oxides were a combination of Al2O<sup>3</sup> and TiO<sup>2</sup> . (Reprinted with permission from Ref. [66]. Copyright 2018 Elsevier B.V.).

permission from Ref. [66]. Copyright 2018 Elsevier B.V.).

*Micromachines* **2022**, *13*, x FOR PEER REVIEW 22 of 42

**Figure 18.** Back-scattered electron micrograph of particle with FIB sectioning to reveal microstructure and surface oxides (darker material) for particle (**a**) with oxide spots and (**b**) with oxide coating. The bright section is the sacrificial platinum strip deposited prior to ion milling. (Reprinted with permission from Ref. [66]. Copyright 2018 Elsevier B.V.). **Figure 18.** Back-scattered electron micrograph of particle with FIB sectioning to reveal microstructure and surface oxides (darker material) for particle (**a**) with oxide spots and (**b**) with oxide coating. The bright section is the sacrificial platinum strip deposited prior to ion milling. (Reprinted with permission from Ref. [66]. Copyright 2018 Elsevier B.V.). **Figure 18.** Back-scattered electron micrograph of particle with FIB sectioning to reveal microstructure and surface oxides (darker material) for particle (**a**) with oxide spots and (**b**) with oxide coating. The bright section is the sacrificial platinum strip deposited prior to ion milling. (Reprinted with

Schwerz et al. [82] investigated the effect of spatter on parts using destructive (metallographic analysis) and non-destructive (ultrasonic inspection) methods. It was discovered that the spatter redeposits zone included numerous internal defects. Based on the results of the redeposited spatters (Figure 19a,c), the cross-section metallography of samples with high and low rates of re-deposition spatters were analyzed. No obvious internal defects were found in the area with a low spatter re-deposition rate, as shown in Figure 19b. Numerous internal defects were found in the area with a high spatter deposition rate, as shown in Figure 19d. These internal defects are observed in conjunction with round particles with a dendritic structure, indicated by white arrows in Figure 19e,f, located with inter-melt pool boundaries, i.e., lack of fusion defects. Multiple internal defects larger than 500 µm were verified by the ultrasonic inspection as the layer thickness increased. Schwerz et al. [82] investigated the effect of spatter on parts using destructive (metallographic analysis) and non-destructive (ultrasonic inspection) methods. It was discovered that the spatter redeposits zone included numerous internal defects. Based on the results of the redeposited spatters (Figure 19a,c), the cross-section metallography of samples with high and low rates of re-deposition spatters were analyzed. No obvious internal defects were found in the area with a low spatter re-deposition rate, as shown in Figure 19b. Numerous internal defects were found in the area with a high spatter deposition rate, as shown in Figure 19d. These internal defects are observed in conjunction with round particles with a dendritic structure, indicated by white arrows in Figure 19e,f, located with inter-melt pool boundaries, i.e., lack of fusion defects. Multiple internal defects larger than 500 µm were verified by the ultrasonic inspection as the layer thickness increased. Schwerz et al. [82] investigated the effect of spatter on parts using destructive (metallographic analysis) and non-destructive (ultrasonic inspection) methods. It was discovered that the spatter redeposits zone included numerous internal defects. Based on the results of the redeposited spatters (Figure 19a,c), the cross-section metallography of samples with high and low rates of re-deposition spatters were analyzed. No obvious internal defects were found in the area with a low spatter re-deposition rate, as shown in Figure 19b. Numerous internal defects were found in the area with a high spatter deposition rate, as shown in Figure 19d. These internal defects are observed in conjunction with round particles with a dendritic structure, indicated by white arrows in Figure 19e,f, located with inter-melt pool boundaries, i.e., lack of fusion defects. Multiple internal defects larger than 500 µm were verified by the ultrasonic inspection as the layer thickness increased.

posits are detected in specimens manufactured in the proximity of the gas inlet. (**b**) Metallographic analysis of these specimens reveals no major internal defects. (**c**) Detections of spatter redeposits can be abundant in specimens manufactured in the proximity of the gas outlet, (**d**) and these specimens present large internal defects. (**e**,**f**) are round particles with dendritic structure neighbor and lack of fusion defects, indicated by white arrows. (Reprinted with permission from Ref. [82]. Copyright 2021 Elsevier B.V.). **Figure 19.** Cross section metallography of damaged testing [82]. (**a**) A low number of spatter redeposits are detected in specimens manufactured in the proximity of the gas inlet. (**b**) Metallographic analysis of these specimens reveals no major internal defects. (**c**) Detections of spatter redeposits can be abundant in specimens manufactured in the proximity of the gas outlet, (**d**) and these specimens present large internal defects. (**e**,**f**) are round particles with dendritic structure neighbor and lack of fusion defects, indicated by white arrows. (Reprinted with permission from Ref. [82]. Copyright 2021 Elsevier B.V.). **Figure 19.** Cross section metallography of damaged testing [82]. (**a**) A low number of spatter redeposits are detected in specimens manufactured in the proximity of the gas inlet. (**b**) Metallographic analysis of these specimens reveals no major internal defects. (**c**) Detections of spatter redeposits can be abundant in specimens manufactured in the proximity of the gas outlet, (**d**) and these specimens present large internal defects. (**e**,**f**) are round particles with dendritic structure neighbor and lack of fusion defects, indicated by white arrows. (Reprinted with permission from Ref. [82]. Copyright 2021 Elsevier B.V.).

**Figure 19.** Cross section metallography of damaged testing [82]. (**a**) A low number of spatter rede-

Spatter can cause a reduction in the tensile properties of the parts. Liu et al. [62] conducted tensile testing from fresh and contaminated 316L stainless steel powder, and the results showed that the mechanical properties of the specimens manufactured with contaminated powder are far inferior to those manufactured with fresh powder, as shown in Figure 20. Specimens with contaminated powder show considerably more voids in the fracture compared to specimens with fresh powder. These voids cause cracks and accelerate crack propagation during tensile testing, resulting in a dramatic reduction of mechanical properties in the specimens. Spatter can cause a reduction in the tensile properties of the parts. Liu et al. [62] conducted tensile testing from fresh and contaminated 316L stainless steel powder, and the results showed that the mechanical properties of the specimens manufactured with contaminated powder are far inferior to those manufactured with fresh powder, as shown in Figure 20. Specimens with contaminated powder show considerably more voids in the fracture compared to specimens with fresh powder. These voids cause cracks and accelerate crack propagation during tensile testing, resulting in a dramatic reduction of mechanical properties in the specimens.

**Figure 20.** The stress–strain curves of tensile test pieces (fabricated from fresh and contaminated). (Reprinted with permission from Ref. [62]. Copyright 2015 Elsevier B.V.). **Figure 20.** The stress–strain curves of tensile test pieces (fabricated from fresh and contaminated). (Reprinted with permission from Ref. [62]. Copyright 2015 Elsevier B.V.).

#### *4.3. Effect of Spatter on Powder Recycling 4.3. Effect of Spatter on Powder Recycling*

Only 2 wt.% to 3 wt.% of the powder is selected for laser melting to metal pieces during L-PBF. Therefore, powder recycling is an efficient method of extending powder use [93]. However, recycled powder contains L-PBF by-products, which causes difficulties in powder recycling. Spatter particles are distributed in various sizes, a sieving mesh can easily remove most of the particles, but a small percentage of spatters smaller than the size of the original powders still remain. The powder recycling shows a distinct impact on the L-PBF process for powders of different components. (1) The 316L stainless steel powder is unique with an inherent SiO2 oxide layer on its surface that prevents the variable valence of metallic elements. It can be used up to 15 times in L-PBF without much affecting the mechanical properties of parts, but the oxygen content of the print increases with the number of recycles, and the part density decreases after 5 to 6 recycles [94]. (2) Ti-6Al-4V also contains an oxide layer on the surface; the elemental content of the powder remains nearly the same after 31 recycles, and the tensile strength, yield strength, and elongation are also almost unchanged [95]. (3) The recyclability of Al-Si10-Mg is poor, and its oxygen content doubles after 6 recycles [96]. (4) The steel alloy 17-4 PH showed a narrowing of the particle size distribution and a loss of tensile strength after 5 recycles [97]. (5) Hastelloy X is easy to be oxidized because it contains oxygenophilic elements such as Si, Cr, and Ni. Due to the wettability of Hastelloy X powder, it produces more spatters, which affects the re-cycling of the powder. He et al. [98] found that after 6 cycles of Hastelloy, the average particle size increased by 22% and the oxygen content increased by 48%, and the part porosity increased, resulting in a reduced part quality. The following Table 7 summarizes the number of re-cycle times available for different powders. Only 2 wt.% to 3 wt.% of the powder is selected for laser melting to metal pieces during L-PBF. Therefore, powder recycling is an efficient method of extending powder use [93]. However, recycled powder contains L-PBF by-products, which causes difficulties in powder recycling. Spatter particles are distributed in various sizes, a sieving mesh can easily remove most of the particles, but a small percentage of spatters smaller than the size of the original powders still remain. The powder recycling shows a distinct impact on the L-PBF process for powders of different components. (1) The 316L stainless steel powder is unique with an inherent SiO<sup>2</sup> oxide layer on its surface that prevents the variable valence of metallic elements. It can be used up to 15 times in L-PBF without much affecting the mechanical properties of parts, but the oxygen content of the print increases with the number of recycles, and the part density decreases after 5 to 6 recycles [94]. (2) Ti-6Al-4V also contains an oxide layer on the surface; the elemental content of the powder remains nearly the same after 31 recycles, and the tensile strength, yield strength, and elongation are also almost unchanged [95]. (3) The recyclability of Al-Si10-Mg is poor, and its oxygen content doubles after 6 recycles [96]. (4) The steel alloy 17-4 PH showed a narrowing of the particle size distribution and a loss of tensile strength after 5 recycles [97]. (5) Hastelloy X is easy to be oxidized because it contains oxygenophilic elements such as Si, Cr, and Ni. Due to the wettability of Hastelloy X powder, it produces more spatters, which affects the re-cycling of the powder. He et al. [98] found that after 6 cycles of Hastelloy, the average particle size increased by 22% and the oxygen content increased by 48%, and the part porosity increased, resulting in a reduced part quality. The following Table 7 summarizes the number of re-cycle times available for different powders.


**Table 7.** Summary of the re-cycle times available for different powders. **Table 7.** Summary of the re-cycle times available for different powders. **Material Powder Parameters Re-Cycle Times References** 

*Micromachines* **2022**, *13*, x FOR PEER REVIEW 24 of 42

According to a study done by Marco Simonelli et al. [103], when powders are used for an extended period of time without sieving, numerous impurities mix with the powder and eventually become embedded in the surface of the manufactured part. Most of those impurities are spatter particles with the same composition as the slag produced during the conventional steel manufacturing process; the impurity consists primarily of SiO<sup>2</sup> and other oxides, which can lead to impurity in the composition of the powder. Even after sieving, some spatter particles remain, and printing using powders containing spatter particles easily results in defects inside the part. Wang et al. [104] discovered that during L-PBF formation of a porous structure, the spatter particles in the recycled powder became inclusions in the part, influencing the part quality. Santecchia et al. [105] found that the environmental conditions in the build chamber can lead to the rapid condensation of vaporized material, and large amounts of condensate and spatter deposited together on the powder bed can affect the reuse of the powder. High concentrations of condensate and condensate on spatter particles were found by Sutton et al. [90] by SEM imaging, as shown in Figure 21. According to a study done by Marco Simonelli et al. [103], when powders are used for an extended period of time without sieving, numerous impurities mix with the powder and eventually become embedded in the surface of the manufactured part. Most of those impurities are spatter particles with the same composition as the slag produced during the conventional steel manufacturing process; the impurity consists primarily of SiO2 and other oxides, which can lead to impurity in the composition of the powder. Even after sieving, some spatter particles remain, and printing using powders containing spatter particles easily results in defects inside the part. Wang et al. [104] discovered that during L-PBF formation of a porous structure, the spatter particles in the recycled powder became inclusions in the part, influencing the part quality. Santecchia et al. [105] found that the environmental conditions in the build chamber can lead to the rapid condensation of vaporized material, and large amounts of condensate and spatter deposited together on the powder bed can affect the reuse of the powder. High concentrations of condensate and condensate on spatter particles were found by Sutton et al. [90] by SEM imaging, as shown in Figure 21.

**Figure 21.** SEM images of condensate. (**a**) A heavy concentration of condensate. (**b**) Condensate on a captured laser spatter particle. (Reprinted with permission from Ref. [90]. Copyright 2019 Elsevier B.V.). **Figure 21.** SEM images of condensate. (**a**) A heavy concentration of condensate. (**b**) Condensate on a captured laser spatter particle. (Reprinted with permission from Ref. [90]. Copyright 2019 Elsevier B.V.).

The spatter has a negative effect on the whole process of L-PBF including the equipment (e.g., laser beam, scraper), current L-PBF manufacturing (e.g., structure and mechanical property), and subsequent L-PBF manufacturing (e.g., powder recycling). The generation of spatter will prevent the laser from directly irradiating on the powder bed, result-The spatter has a negative effect on the whole process of L-PBF including the equipment (e.g., laser beam, scraper), current L-PBF manufacturing (e.g., structure and mechanical property), and subsequent L-PBF manufacturing (e.g., powder recycling). The generation of spatter will prevent the laser from directly irradiating on the powder bed, resulting in the loss of laser energy. The redeposited spatters will damage the scraper and become inclusions in the parts, which will reduce the structure and mechanical properties

of the parts. Furthermore, spattering has an influence on the whole life cycle of powder. In current manufacturing, the spatters redeposit into the powder bed, and irregularly shaped spatter particles will become inclusions in the powder, increasing the powder's oxygen concentration. These powders can result in inferior quality parts in subsequent manufacturing, leading to a decrease in the amount of powder recycling. Metal powders are more expensive than ingot metal, therefore, increasing the number of recycles of the powder is critical to making it more efficient to utilize. Spatter reduces powder quality and re-cycle times, and its removal can effectively improve powder usage efficiency, thus it is essential to research spatter countermeasures. The disadvantages of spatter are summarized in Table 8.


**Table 8.** Summary of studies on the disadvantages of spatter.

#### **5. Spatter Countermeasures**

The disadvantages of spatter include the equipment, components, and powders. Effective spatter countermeasures would extend equipment life, improve the parts' quality, and enhance powder use. The full cycle of the spatter can be divided into three parts: generation, ejection, and re-deposition. In the generation stage, the generation of spatter can be suppressed by optimizing the laser volumetric energy density (VED), laser beam mode, and pressure of the building chamber. During the ejection and re-deposition stages, the protective gas flow is applied to remove the spatters which are in motion above the powder bed.
