**3. Mechanism of Spatter Generation**

Under the interaction with a high-energy laser in L-PBF, metal powders are melted to form a melt pool when the temperature attains the melting point, then vaporized to form metal vapor or even a plasma plume when the surface temperature of the melt pool surpasses the boiling point. The different phases (solid, liquid, and vapor) significantly interact with each other during L-PBF process, among which the vapor–solid interaction and vapor–liquid interaction are the main mechanism of spatter generation. Therefore, it is necessary to investigate the mechanism of spatter generation.

#### *3.1. Spatter Classification Micromachines* **2022**, *13*, x FOR PEER REVIEW 11 of 42

The spatters generated in L-PBF are in a different morphology, and a variety of parameters affect spatter generation. Until now, there has been no common definition of spatter categorization. Liu et al. [62] performed L-PBF single-pass scanning experiments with 316L stainless steel powder, reflecting the dynamic behavior of spatter perpendicular to the single-track scanning direction by the high-speed imaging technology. They divided the spatter into two categories: droplet spatter and powder spatter. It is known that the spatter formation mechanism can be demonstrated as the hot spatter ejection, mainly driven by the instability of the melt pool due to the vapor-induced recoil pressure, and the cold spatter ejection, mainly driven by the vapor-induced entrainment of the protective gas. Wang et al. [63] used a high-speed camera to record the dynamic spattering process of Co– Cr alloys during L-PBF manufacturing and investigated the spatter generation mechanism in further detail. As shown in Figure 8, they recognized three major sources of spattering: recoil pressure, the Marangoni effect, and the heat effect in the melt pool. These three different sources of spattering led to three types of spattering morphologies. spatter categorization. Liu et al. [62] performed L-PBF single-pass scanning experiments with 316L stainless steel powder, reflecting the dynamic behavior of spatter perpendicular to the single-track scanning direction by the high-speed imaging technology. They divided the spatter into two categories: droplet spatter and powder spatter. It is known that the spatter formation mechanism can be demonstrated as the hot spatter ejection, mainly driven by the instability of the melt pool due to the vapor-induced recoil pressure, and the cold spatter ejection, mainly driven by the vapor-induced entrainment of the protective gas. Wang et al. [63] used a high-speed camera to record the dynamic spattering process of Co–Cr alloys during L-PBF manufacturing and investigated the spatter generation mechanism in further detail. As shown in Figure 8, they recognized three major sources of spattering: recoil pressure, the Marangoni effect, and the heat effect in the melt pool. These three different sources of spattering led to three types of spattering morphologies.

**Figure 8.** Formation mechanisms of different types of spatter during L-PBF: (**a**) morphology of spherical spatter (Type-I spatter); (**b**) morphology of coarse spherical morphology (Type-II spatter); (**c**) morphology of irregular spatter (Type-III spatter). (Reprinted with permission from Ref. [63]. Copyright 2016 Elsevier B.V.). **Figure 8.** Formation mechanisms of different types of spatter during L-PBF: (**a**) morphology of spherical spatter (Type-I spatter); (**b**) morphology of coarse spherical morphology (Type-II spatter); (**c**) morphology of irregular spatter (Type-III spatter). (Reprinted with permission from Ref. [63]. Copyright 2016 Elsevier B.V.).

According to Ref. [63], there are three types of spatters: (ⅰ) The Type I spatters are associated with the extreme expansion of the gas phase. The spontaneous metal liquid flowing will occur from the high-temperature bottom of the excavation to the low-temperature sidewall and edge at the back under the Marangoni effect. (ⅱ) Then, in this process, the recoil pressure can induce the jet of low-viscosity metal liquid, and this jetted liquid metal will divide into small drops in the flight process to minimize the surface tension; therefore, the Type II spatter is formed. (ⅲ) In the printing process, some metal liquid accumulates near the spot laser, and it can be easily squeezed by the blast wave and then interrupt these non-melted particles in the front-end area; then the Type III spatter will occur at the front of the melt pool. According to Ref. [63], there are three types of spatters: (i) The Type I spatters are associated with the extreme expansion of the gas phase. The spontaneous metal liquid flowing will occur from the high-temperature bottom of the excavation to the low-temperature sidewall and edge at the back under the Marangoni effect. (ii) Then, in this process, the recoil pressure can induce the jet of low-viscosity metal liquid, and this jetted liquid metal will divide into small drops in the flight process to minimize the surface tension; therefore, the Type II spatter is formed. (iii) In the printing process, some metal liquid accumulates near the spot laser, and it can be easily squeezed by the blast wave and then interrupt these non-melted particles in the front-end area; then the Type III spatter will occur at the front of the melt pool.

Ly et al. [64] used a high-speed camera to explore the influence of gas flow entrainment on spatter during L-PBF. They described the entrainment phenomena of 316L stainless steel powder and Ti-6Al-4V powder layers and divided spatter into three categories. As shown in Figure 9, 60% of the ejection was due to hot entrainment ejection at velocities Ly et al. [64] used a high-speed camera to explore the influence of gas flow entrainment on spatter during L-PBF. They described the entrainment phenomena of 316L stainless steel powder and Ti-6Al-4V powder layers and divided spatter into three categories. As shown in Figure 9, 60% of the ejection was due to hot entrainment ejection at velocities ranging

ranging from 6 m/s to 20 m/s: 25% was cold entrainment ejection, which occurred at a velocity of 2 m/s to 4 m/s, and 15% was droplet breakup ejection from the melt pool as a result of the recoil pressure applied at a velocity of 3 to 8 m/s. Raza et al. [65] also found

from 6 m/s to 20 m/s: 25% was cold entrainment ejection, which occurred at a velocity of 2 m/s to 4 m/s, and 15% was droplet breakup ejection from the melt pool as a result of the recoil pressure applied at a velocity of 3 to 8 m/s. Raza et al. [65] also found that spatter from the melt pool was less than that due to vapor-induced entrainment. *Micromachines* **2022**, *13*, x FOR PEER REVIEW 12 of 42

**Figure 9.** Schematic representation of the entrainment effect of metal-vapor-driven airflow on fine particles in the presence of a fixed laser beam. A vapor jet creates a zone of low pressure, which results in three different trajectories of entrained particles: (1) Particles with low vertical momentum are swept into the melt pool; (2) Particles with higher vertical momentum but originating > 2 melt pool widths away are swept into the trailing portion of the vapor jet, and ejected as cold particles; (3) particles with roughly the same vertical momentum as (2) but originating closer to the point of laser irradiation (<2 melt pool widths) are swept into or near the laser beam itself rapidly heat, and are ejected as incandescent, hot particles. (Reprinted with permission from Ref. [64]. Copyright 2017 Springer Nature.). **Figure 9.** Schematic representation of the entrainment effect of metal-vapor-driven airflow on fine particles in the presence of a fixed laser beam. A vapor jet creates a zone of low pressure, which results in three different trajectories of entrained particles: (1) Particles with low vertical momentum are swept into the melt pool; (2) Particles with higher vertical momentum but originating > 2 melt pool widths away are swept into the trailing portion of the vapor jet, and ejected as cold particles; (3) particles with roughly the same vertical momentum as (2) but originating closer to the point of laser irradiation (<2 melt pool widths) are swept into or near the laser beam itself rapidly heat, and are ejected as incandescent, hot particles. (Reprinted with permission from Ref. [64]. Copyright 2017 Springer Nature).

Young et al. [56] showed the characteristics and generation mechanisms of five unique types of spatter during L-PBF by in situ high-speed, high-energy X-ray video imaging: solid spatter, metallic ejected spatter, agglomeration spatter, entrainment melting spatter, and defect-induced spatter. They quantified the speed, size, and direction of metallic ejected spatter, powder agglomeration spatter, and entrainment melting spatter. The results showed that the metallic ejected spatter speed was the highest, and the size of the powder agglomeration spatter was the largest. The spatter direction was highly dependent on the characteristics of the depression zone, which was impacted directly by the metal vapor recoil pressure. Young et al. [56] showed the characteristics and generation mechanisms of five unique types of spatter during L-PBF by in situ high-speed, high-energy X-ray video imaging: solid spatter, metallic ejected spatter, agglomeration spatter, entrainment melting spatter, and defect-induced spatter. They quantified the speed, size, and direction of metallic ejected spatter, powder agglomeration spatter, and entrainment melting spatter. The results showed that the metallic ejected spatter speed was the highest, and the size of the powder agglomeration spatter was the largest. The spatter direction was highly dependent on the characteristics of the depression zone, which was impacted directly by the metal vapor recoil pressure.

Whereas the above researches had classified spatter using an in-process analysis, the following is a study that classified spatter using a post-mortem analysis. Gasper et al. [66] divided the spatter into seven categories according to the size, morphology, and other descriptors, such as oxides and agglomeration derived from SEM analysis, namely: (1) particles similar to virgin gas-atomized particles, (2) particles morphologically different from those gas-atomized, (3) larger singular particles with different morphologies, (4) particles with oxide spots, (5) particles covered with oxide, (6) small particles, and (7) agglomerates. Yang et al. [67] studied the influence of the L-PBF parameters on the pore characteristics and mechanical properties of Al-Si10-Mg parts. Three distinct types of solidified droplets were detected: hollow droplets, semi-hollow droplets, and solid droplets. Hollow droplets and semi-hollow droplets were a major source of pores inside the sample. Table 3 summarizes current studies on the categorization of spatter generation during L-PBF. Whereas the above researches had classified spatter using an in-process analysis, the following is a study that classified spatter using a post-mortem analysis. Gasper et al. [66] divided the spatter into seven categories according to the size, morphology, and other descriptors, such as oxides and agglomeration derived from SEM analysis, namely: (1) particles similar to virgin gas-atomized particles, (2) particles morphologically different from those gas-atomized, (3) larger singular particles with different morphologies, (4) particles with oxide spots, (5) particles covered with oxide, (6) small particles, and (7) agglomerates. Yang et al. [67] studied the influence of the L-PBF parameters on the pore characteristics and mechanical properties of Al-Si10-Mg parts. Three distinct types of solidified droplets were detected: hollow droplets, semi-hollow droplets, and solid droplets. Hollow droplets and semi-hollow droplets were a major source of pores inside the sample. Table 3 summarizes current studies on the categorization of spatter generation during L-PBF.


#### **Table 3.** Summary of spatter classification studies.
