**1. Introduction**

Additive manufacturing (AM) is widely used in aerospace, medicine, jewelry, and other industries because of its rapid fabrication [1,2], low cost, and the ability to print parts with complex geometries [3,4]. Today, many developed and developing countries regard AM technology as a fifth industrial revolution and make many efforts in the development of AM. The United States Department of Defense (DoD) released the Department of Defense Additive Manufacturing Strategy [5] to stimulate the development of AM applications in national defense. Meanwhile, the Office of the Under Secretary of Defense released the first policy paper, DoD 5000.93 Directive Use of Additive Manufacturing in the Department of Defense [6], which promoted the implementation of the AM strategy. The Ministry of

**Citation:** Li, Z.; Li, H.; Yin, J.; Li, Y.; Nie, Z.; Li, X.; You, D.; Guan, K.; Duan, W.; Cao, L.; et al. A Review of Spatter in Laser Powder Bed Fusion Additive Manufacturing: In Situ Detection, Generation, Effects, and Countermeasures. *Micromachines* **2022**, *13*, 1366. https://doi.org/ 10.3390/mi13081366

Academic Editor: Saulius Juodkazis

Received: 18 July 2022 Accepted: 15 August 2022 Published: 22 August 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Science and Technology of the People's Republic of China released the 2022 annual project application guide for the key projects of additive manufacturing and laser manufacturing under the 14th Five-Year National Key R&D Program [7] to establish a new standard system for AM that is consistent with international standards. Additionally, AM and laser manufacturing are two of the important tasks of the National Program for Medium-to-Long-Term Scientific and Technological Development and Made in China 2025. The EU began funding projects on AM technology as early as the first Framework Program for Research and Technological Development. Under these conditions, AM technology has advanced significantly and rapidly in developing standard systems, key technologies, and multi-industry applications. annual project application guide for the key projects of additive manufacturing and laser manufacturing under the 14th Five-Year National Key R&D Program [7] to establish a new standard system for AM that is consistent with international standards. Additionally, AM and laser manufacturing are two of the important tasks of the National Program for Medium-to-Long-Term Scientific and Technological Development and Made in China 2025. The EU began funding projects on AM technology as early as the first Framework Program for Research and Technological Development. Under these conditions, AM technology has advanced significantly and rapidly in developing standard systems, key technologies, and multi-industry applications. AM technology emerged in the 1990s and has been under development for approxi-

Additive manufacturing (AM) is widely used in aerospace, medicine, jewelry, and other industries because of its rapid fabrication [1,2], low cost, and the ability to print parts with complex geometries [3,4]. Today, many developed and developing countries regard AM technology as a fifth industrial revolution and make many efforts in the development of AM. The United States Department of Defense (DoD) released the Department of Defense Additive Manufacturing Strategy [5] to stimulate the development of AM applications in national defense. Meanwhile, the Office of the Under Secretary of Defense released the first policy paper, DoD 5000.93 Directive Use of Additive Manufacturing in the Department of Defense [6], which promoted the implementation of the AM strategy. The Ministry of Science and Technology of the People's Republic of China released the 2022

*Micromachines* **2022**, *13*, x FOR PEER REVIEW 2 of 42

**1. Introduction** 

AM technology emerged in the 1990s and has been under development for approximately three decades [8]. Unlike "subtractive manufacturing" (e.g., cutting, drilling, and milling) and "equal-material manufacturing" (e.g., welding, casting, and forging), AM is built on 3D models [9], relies on layer by layer printing-extrusion, sintering [10,11], melting, light curing, and jetting to form solids from metallic or non-metallic materials [12,13]. mately three decades [8]. Unlike "subtractive manufacturing" (e.g., cutting, drilling, and milling) and "equal-material manufacturing" (e.g., welding, casting, and forging), AM is built on 3D models [9], relies on layer by layer printing-extrusion, sintering [10,11], melting, light curing, and jetting to form solids from metallic or non-metallic materials [12,13]. Metal AM is one of the most difficult and cutting-edge AM technologies. As shown

Metal AM is one of the most difficult and cutting-edge AM technologies. As shown in Figure 1, metal AM technologies can be divided into two categories, direct energy deposition (DED) and powder bed fusion (PBF) [14,15]. PBF is one of the AM technologies used to fabricate metal objects from powder feedstocks with two kinds of input energy: laser and electron [16–18]. In the printing process, the metal powder bed is melted by the high energy source with a designed pattern using a layer by layer printing strategy [19–21]. in Figure 1, metal AM technologies can be divided into two categories, direct energy deposition (DED) and powder bed fusion (PBF) [14,15]. PBF is one of the AM technologies used to fabricate metal objects from powder feedstocks with two kinds of input energy: laser and electron [16–18]. In the printing process, the metal powder bed is melted by the high energy source with a designed pattern using a layer by layer printing strategy [19– 21].

**Figure 1.** Classification of metal manufacturing processes: equal-material manufacturing, additive manufacturing [22], subtractive manufacturing. **Figure 1.** Classification of metal manufacturing processes: equal-material manufacturing, additive manufacturing [22], subtractive manufacturing.

Figure 1 also illustrates the forming principle of laser powder bed fusion (L-PBF), which is widely used today to rapidly manufacture parts with complicated shapes, a fine grain size, high densities, and superior mechanical properties [23,24]. Although it can currently fabricate complicated metal parts [25,26], the reliability and stability of the printing process remain inadequate [27]. There are still defects in L-PBF processing that decrease the density and affect the mechanical characteristics of the part or even result in fabrication failure. The many unresolved problems with L-PBF become a barrier to the expansion of L-PBF applications.Spatter is generated in conventional laser welding and cutting, DED, and L-PBF. Spatters are the particles ejected from a melt pool during the laser–metal interaction [28]. In conventional laser welding and cutting, the laser scanning path is relatively simple, with few overlap regions between the scanning paths, and DED has a lower scanning velocity and a larger spot than L-PBF. However, L-PBF is a powder-bed-based technology, and the printing process is more complicated than that of the three technologies mentioned above, which results in a more complex spatter behavior. Furthermore, during multi-laser L-PBF, the thermal and stress cycling, melt pool characteristics, spatter behavior, and metal vapor evolution will be definitely different from that of the single-laser PBF. The detection of spatter under multi-laser L-PBF is more difficult. atively simple, with few overlap regions between the scanning paths, and DED has a lower scanning velocity and a larger spot than L-PBF. However, L-PBF is a powder-bedbased technology, and the printing process is more complicated than that of the three technologies mentioned above, which results in a more complex spatter behavior. Furthermore, during multi-laser L-PBF, the thermal and stress cycling, melt pool characteristics, spatter behavior, and metal vapor evolution will be definitely different from that of the single-laser PBF. The detection of spatter under multi-laser L-PBF is more difficult.

Figure 1 also illustrates the forming principle of laser powder bed fusion (L-PBF), which is widely used today to rapidly manufacture parts with complicated shapes, a fine grain size, high densities, and superior mechanical properties [23,24]. Although it can currently fabricate complicated metal parts [25,26], the reliability and stability of the printing process remain inadequate [27]. There are still defects in L-PBF processing that decrease the density and affect the mechanical characteristics of the part or even result in fabrication failure. The many unresolved problems with L-PBF become a barrier to the expansion of L-PBF applications.Spatter is generated in conventional laser welding and cutting, DED, and L-PBF. Spatters are the particles ejected from a melt pool during the laser–metal interaction [28]. In conventional laser welding and cutting, the laser scanning path is rel-

*Micromachines* **2022**, *13*, x FOR PEER REVIEW 3 of 42

For this reason, studies on L-PBF spatter are becoming very urgent. Spatter as a byproduct of L-PBF is unpreventable [29,30]. It is a detriment to the forming process, and the part and the redeposited spatters can destroy the original well-built powder layer, resulting in non-fusion defects [31,32]. Due to the uniqueness of L-PBF, the undesired effects of spatter are amplified during the layer-by-layer process. Spatter affects the subsequent re-coating and melting of the powder, resulting in internal defects in the produced part or the part failing to form. For this reason, studies on L-PBF spatter are becoming very urgent. Spatter as a byproduct of L-PBF is unpreventable [29,30]. It is a detriment to the forming process, and the part and the redeposited spatters can destroy the original well-built powder layer, resulting in non-fusion defects [31,32]. Due to the uniqueness of L-PBF, the undesired effects of spatter are amplified during the layer-by-layer process. Spatter affects the subsequent re-coating and melting of the powder, resulting in internal defects in the produced part or the part failing to form.

As spatter has a significant effect on L-PBF, it can be used to represent the L-PBF machining state. Spatter contains a plethora of information and can be used in various ways to analyze the manufacturing processing of L-PBF. By observing and quantifying the spatter, it is possible to establish an intrinsic correlation between spatter and the part quality, enabling a more comprehensive understanding of the L-PBF process to solve the problems of insufficient stability and reliability, allowing this technology to be popularized and applied more widely. As spatter has a significant effect on L-PBF, it can be used to represent the L-PBF machining state. Spatter contains a plethora of information and can be used in various ways to analyze the manufacturing processing of L-PBF. By observing and quantifying the spatter, it is possible to establish an intrinsic correlation between spatter and the part quality, enabling a more comprehensive understanding of the L-PBF process to solve the problems of insufficient stability and reliability, allowing this technology to be popularized and applied more widely.

Recently, the research concerning the spatter during L-PBF has received more and more extensive attention. In this work, we review academic publications concerning L-PBF spatter in the Web of Science database from 2015 to date (Topic: ["laser-powder bed fusion" and "spatter"] or ["selective laser melting" and "spatter"]). Figure 2 shows the trend in the number of articles on this topic over the last several years. Recently, the research concerning the spatter during L-PBF has received more and more extensive attention. In this work, we review academic publications concerning L-PBF spatter in the Web of Science database from 2015 to date (Topic: ["laser-powder bed fusion" and "spatter"] or ["selective laser melting" and "spatter"]). Figure 2 shows the trend in the number of articles on this topic over the last several years.

This article builds on previous research by reviewing a synthesis of in situ spatter detection systems, spatter detection equipment, the generation of spatter and its associated disadvantages, and current approaches for the suppression and removal of spatter. Finally, the future of research on L-PBF spatter is discussed.

#### **2. Laser Powder Bed Fusion Spatter In Situ Detection Device**

The L-PBF detection system can be categorized as: static detection (imaging of spreading powder and deformation) and dynamic detection (characterization of melt pool, spatter, and vapor plume).

The spatter generated by conventional laser welding, cutting, and DED is similar to that produced by L-PBF and is caused by the interaction between the laser and the metal material. However, L-PBF has a smaller spot (~10<sup>1</sup> to 10<sup>2</sup> µm), a smaller melt pool (up to 100 µm), a shorter lifetime (~10 ms), and a higher scanning velocity (~10<sup>2</sup> to 10<sup>3</sup> mm/s) compared to laser welding, cutting, and DED [33]. Furthermore, in L-PBF, the laser interacts with the powder bed and the metal part more than once, resulting in a greater number and variety of spatters and complicating in situ spatter detection. The spatter generated by conventional laser welding, cutting, and DED is similar to that produced by L-PBF and is caused by the interaction between the laser and the metal material. However, L-PBF has a smaller spot (~101 to 102 µm), a smaller melt pool (up to 100 µm), a shorter lifetime (~10 ms), and a higher scanning velocity (~102 to 103 mm/s) compared to laser welding, cutting, and DED [33]. Furthermore, in L-PBF, the laser interacts with the powder bed and the metal part more than once, resulting in a greater number and variety of spatters and complicating in situ spatter detection. The laser–powder bed interaction produces the melt pool, spatter, and vapor plume

This article builds on previous research by reviewing a synthesis of in situ spatter detection systems, spatter detection equipment, the generation of spatter and its associated disadvantages, and current approaches for the suppression and removal of spatter.

The L-PBF detection system can be categorized as: static detection (imaging of spreading powder and deformation) and dynamic detection (characterization of melt

The laser–powder bed interaction produces the melt pool, spatter, and vapor plume (even plasma). The trajectory of the melt pool is in the plane of the laser path and can be predicted according to the strategy path, whereas the motion of the spatter is in a 3D space, and its trajectory is complex and difficult to predict. So, the detection of spatter is more difficult. Spatter can be divided into hot droplet spatter (mainly from the instability of the melt pool) and cold powder spatter (mainly driven by the vapor-induced entrainment of the protective gas). Both of them can be detected with the visible-light camera equipped with an illumination source, and the relevant collected information can be used to analyze them. (even plasma). The trajectory of the melt pool is in the plane of the laser path and can be predicted according to the strategy path, whereas the motion of the spatter is in a 3D space, and its trajectory is complex and difficult to predict. So, the detection of spatter is more difficult. Spatter can be divided into hot droplet spatter (mainly from the instability of the melt pool) and cold powder spatter (mainly driven by the vapor-induced entrainment of the protective gas). Both of them can be detected with the visible-light camera equipped with an illumination source, and the relevant collected information can be used to analyze them.

According to various studies, the following methods are currently available for L-PBF spatter detection: (1) a visible-light high-speed camera, (2) X-ray video imaging, (3) infrared video imaging, and (4) schlieren video imaging. These detection techniques can detect different characteristics, as shown in Figure 3 and Table 1. According to various studies, the following methods are currently available for L-PBF spatter detection: (1) a visible-light high-speed camera, (2) X-ray video imaging, (3) infrared video imaging, and (4) schlieren video imaging. These detection techniques can detect different characteristics, as shown in Figure 3 and Table 1.

*Micromachines* **2022**, *13*, x FOR PEER REVIEW 4 of 42

Finally, the future of research on L-PBF spatter is discussed.

pool, spatter, and vapor plume).

**2. Laser Powder Bed Fusion Spatter In Situ Detection Device** 

**Figure 3.** Characteristics obtained from different in situ detection techniques: (**a1**–**a3**) time series snapshots taken by visible light high-speed camera (Reprinted with permission from Ref. [34]. Copyright 2019 Elsevier B.V.); (**b1**–**b3**) high-speed schlieren images during single track scans (Reprinted with permission from Ref. [35]. Copyright 2018 Springer Nature.); (**c1**–**c3**) dynamic X-ray images showing powder motion, A is the ejected powder (Reprinted with permission from Ref. [36]. Copyright 2018 Elsevier B.V.); (**d1**–**d3**) three consecutive frames of an infrared video acquired during L-PBF (Reprinted with permission from Ref. [37]. Copyright 2018 Elsevier B.V.). **Figure 3.** Characteristics obtained from different in situ detection techniques: (**a1**–**a3**) time series snapshots taken by visible light high-speed camera (Reprinted with permission from Ref. [34]. Copyright 2019 Elsevier B.V.); (**b1**–**b3**) high-speed schlieren images during single track scans (Reprinted with permission from Ref. [35]. Copyright 2018 Springer Nature.); (**c1**–**c3**) dynamic X-ray images showing powder motion, A is the ejected powder (Reprinted with permission from Ref. [36]. Copyright 2018 Elsevier B.V.); (**d1**–**d3**) three consecutive frames of an infrared video acquired during L-PBF (Reprinted with permission from Ref. [37]. Copyright 2018 Elsevier B.V.).

**Table 1.** Characteristics obtained from different in situ detection techniques.


#### *2.1. Visible-Light High-Speed Detector 2.1. Visible-Light High-Speed Detector*

There are two main methods for observing L-PBF with a high-speed visible-light camera: coaxial and off-axis. In Figure 4a, the camera shares the same optical path with the laser in a coaxial solution. In Figure 4b, the camera is placed at an angle to the optical path of the laser for viewing in Figure 4a, an off-axis solution. There are two main methods for observing L-PBF with a high-speed visible-light camera: coaxial and off-axis. In Figure 4a, the camera shares the same optical path with the laser in a coaxial solution. In Figure 4b, the camera is placed at an angle to the optical path of the laser for viewing in Figure 4a, an off-axis solution.

Schlieren video imaging Gas flow propagation and distribution

Flow behavior of melt inside the melt pool

Flow behavior of gas

*Micromachines* **2022**, *13*, x FOR PEER REVIEW 5 of 42

**Table 1.** Characteristics obtained from different in situ detection techniques.

**In Situ Detection Technology Obtained Characteristics**  Visible-light high-speed camera Surface characteristics X-ray video imaging Internal structure

Infrared video imaging Temperature distribution

**Figure 4.** L-PBF in situ detection system: (**a**) coaxial (sharing the optical path with the laser); (**b**) offaxial (at a slight angle to the laser optical path). **Figure 4.** L-PBF in situ detection system: (**a**) coaxial (sharing the optical path with the laser); (**b**) off-axial (at a slight angle to the laser optical path).

Coaxial in situ detection of a commercial L-PBF machine requires extensive modification of the machine, and it is still difficult to obtain clear images because of the distance between the optical path system and the powder bed in the L-PBF machine. Another hindrance is the small optical aperture of the scanner and F-theta lens, which results in low magnification. In addition, the low reflectivity of the scanner and the low transmittance of the F-theta lens also reduce the temporal and spatial resolution of imaging, and these two characters are vital for the analyzing the trajectory and behavior of spatters. To overcome the disadvantages of coaxial in situ detection, Zhang et al. [38] improved the optical path, built segmentation algorithms, and demonstrated the algorithms' efficiency in dealing with defocused and distorted spatter images. Coaxial in situ detection of a commercial L-PBF machine requires extensive modification of the machine, and it is still difficult to obtain clear images because of the distance between the optical path system and the powder bed in the L-PBF machine. Another hindrance is the small optical aperture of the scanner and F-theta lens, which results in low magnification. In addition, the low reflectivity of the scanner and the low transmittance of the F-theta lens also reduce the temporal and spatial resolution of imaging, and these two characters are vital for the analyzing the trajectory and behavior of spatters. To overcome the disadvantages of coaxial in situ detection, Zhang et al. [38] improved the optical path, built segmentation algorithms, and demonstrated the algorithms' efficiency in dealing with defocused and distorted spatter images.

Unlike the coaxial solution, the off-axis solution, which places the detection device at an angle to the powder bed, enables spatter detection without altering the existing L-PBF equipment, as shown in Figure 5. The system is more adaptable and simpler to alter, and because the detection system does not share the optical path of the laser, it is not constrained by the laser's original optical path and can be used to detect spatter at higher magnification and frame rates than those of the coaxial system. Due to these factors, offaxial in situ detection system is becoming increasingly popular. Unlike the coaxial solution, the off-axis solution, which places the detection device at an angle to the powder bed, enables spatter detection without altering the existing L-PBF equipment, as shown in Figure 5. The system is more adaptable and simpler to alter, and because the detection system does not share the optical path of the laser, it is not constrained by the laser's original optical path and can be used to detect spatter at higher magnification and frame rates than those of the coaxial system. Due to these factors, off-axial in situ detection system is becoming increasingly popular.

**Figure 5.** Off-axial detection system: (**a**) off-axial mounting of the high-speed camera outside the build chamber; (**b**) off-axial high-speed camera is at an angle of 45° to the plane of the observed powder bed. (Reprinted with permission from Ref. [39]. Copyright 2017 Elsevier B.V.). **Figure 5.** Off-axial detection system: (**a**) off-axial mounting of the high-speed camera outside the build chamber; (**b**) off-axial high-speed camera is at an angle of 45◦ to the plane of the observed powder bed. (Reprinted with permission from Ref. [39]. Copyright 2017 Elsevier B.V.).

In the case of coaxial detection, the detection equipment and external light source affect the final detection findings. Yang et al. installed a high-speed camera (pco. Dimax HS4, 3000 fps) outside the L-PBF machine at a 65° angle to the working platform to detect the spatter. Due to the little difference in brightness between the powder spatter and power bed within the view field of the high-speed camera, only the droplet spatters were detected, but not the nonmolten powder particles. Tan et al. [40] used a computational technique to analyze the obtained images, segmenting each block to extract the spatter. In the same year, Yin et al. [41] introduced an external light source (a CAVILUX® pulsed high-power diode laser light source) and a high-speed camera (Phantom V2012) to detect the spatter and obtain clearer images. After that, this in situ detection system was used to investigate the correlation between ex situ melt track properties and in situ high-speed, high-resolution characterizations [34]. In the case of coaxial detection, the detection equipment and external light source affect the final detection findings. Yang et al. installed a high-speed camera (pco. Dimax HS4, 3000 fps) outside the L-PBF machine at a 65◦ angle to the working platform to detect the spatter. Due to the little difference in brightness between the powder spatter and power bed within the view field of the high-speed camera, only the droplet spatters were detected, but not the nonmolten powder particles. Tan et al. [40] used a computational technique to analyze the obtained images, segmenting each block to extract the spatter. In the same year, Yin et al. [41] introduced an external light source (a CAVILUX® pulsed high-power diode laser light source) and a high-speed camera (Phantom V2012) to detect the spatter and obtain clearer images. After that, this in situ detection system was used to investigate the correlation between ex situ melt track properties and in situ high-speed, high-resolution characterizations [34].

The above studies were based on the monocular camera, and the picture information collected was in a 2D space. By combining multiple cameras and using image processing arithmetic, 3D information of spatter and its mobility can be gathered. Based on the use of monocular sensors, Luo et al. [42] innovatively proposed the use of acoustic signals combined with deep learning for spatter detection, demonstrating the feasibility of the acoustic signal detection of spatter behavior. Due to the dimensional limitation of the 2D image (acquired by the monocular sensor), it is difficult to accurately calculate the behavioral information of the spatter and obtain accurate spatter trajectory, velocity, and other information. A binocular stereo detector can obtain the spatter information in two viewing angles. By using the multi-directional information, its algorithm can present the 3D trajectory and velocity of a single spatter, and the obtained information is more accurate than those of a monocular sensor. Barret et al. [43] established a stereo vision spatter detection system for spatter tracking analysis at a cost of less than USD 1000 using two slow-motion cameras (FPS1000 by The Slow Motion Camera Company), as illustrated in Figure 6. Later, Eschner et al. [44] combined two ultra-high-speed cameras with algorithms to create a 3D tracking system for measuring spatter in L-PBF. Visible in situ detection systems for L-PBF in recent years are summarized in Table 2. The above studies were based on the monocular camera, and the picture information collected was in a 2D space. By combining multiple cameras and using image processing arithmetic, 3D information of spatter and its mobility can be gathered. Based on the use of monocular sensors, Luo et al. [42] innovatively proposed the use of acoustic signals combined with deep learning for spatter detection, demonstrating the feasibility of the acoustic signal detection of spatter behavior. Due to the dimensional limitation of the 2D image (acquired by the monocular sensor), it is difficult to accurately calculate the behavioral information of the spatter and obtain accurate spatter trajectory, velocity, and other information. A binocular stereo detector can obtain the spatter information in two viewing angles. By using the multi-directional information, its algorithm can present the 3D trajectory and velocity of a single spatter, and the obtained information is more accurate than those of a monocular sensor. Barret et al. [43] established a stereo vision spatter detection system for spatter tracking analysis at a cost of less than USD 1000 using two slow-motion cameras (FPS1000 by The Slow Motion Camera Company), as illustrated in Figure 6. Later, Eschner et al. [44] combined two ultra-high-speed cameras with algorithms to create a 3D tracking system for measuring spatter in L-PBF. Visible in situ detection systems for L-PBF in recent years are summarized in Table 2.

**Figure 6.** Spatter detection system for stereo vision: (**a**) the horizontal angle between the two sensors is 15°; (**b**) the vertical direction of the sensor is 12.5° from the observation object. (Reprinted with permission from Ref. [43]. Copyright 2018 University of Texas at Austin.). **Figure 6.** Spatter detection system for stereo vision: (**a**) the horizontal angle between the two sensors is 15◦ ; (**b**) the vertical direction of the sensor is 12.5◦ from the observation object. (Reprinted with permission from Ref. [43]. Copyright 2018 University of Texas at Austin).




#### **Table 2.** *Cont.*

\* Keyhole: also known as the depression zone, is wrapped by the gas–liquid interface and penetrates through the melt pool.

#### *2.2. Invisible-Light In Situ Detection*

For the invisible-light in situ detection of L-PBF, the imaging technologies mainly include X-ray imaging, schlieren video imaging, infrared imaging, and thermal imaging.

X-rays have a short wavelength, high energy, and high penetration ability. Highstrength X-rays can penetrate a certain thickness of metal with high temporal and spatial resolution, which is the preferred method for many L-PBF spatter studies [35]. A schematic diagram of an X-ray system is shown in Figure 7. As one of the most productive X-ray sources globally, the Advanced Photon Source (APS) in the Argonne National Laboratory provides experimental conditions for many researchers. More than 5500 researchers per year use X-rays produced by APS to do experiments. Many of those researchers use those X-rays to detect L-PBF spatter. For example, Zhao et al. [28] pioneered the use of high-speed X-rays (harmonic energy 24.4 keV) for in situ characterizations of L-PBF progress. Guo et al. [36] found transient spatter dynamics in L-PBF using a high-speed, high-resolution, and high-energy X-ray imaging technique. Ross Cunningham et al. quantified the keyhole in Ti-6Al-4V powder during laser melting based on X-ray image information [55]. Leung et al. raised the X-ray power (monochromatic X-ray power: 55 keV) and studied stainless steel (316L) and 13–93 bioactive glass. They found that melt pool wetting and vapor-driven powder entrainment are key track growth mechanisms for L-PBF [57]. A summary of X-ray in situ detection is shown in Table 2.

Due to the Schlieren video imaging and infrared imaging to picture previously invisible light or materials, these two technologies are also used for the in situ detection of spatter. Schlieren video imaging, used to detect the plume in L-PBF, can visualize the invisible substance by measuring its refractive index. Bidare et al. [58] used a combination of a high-speed camera and schlieren video imaging to capture images of the denuded region, laser plume, and argon atmosphere, and explained the relation between the powder-bed denuded region and spatter. An infrared camera can collect the light emitted by an infrared light source. Ye et al. [53] used infrared cameras to detect the properties of the original plume and spatter. Grasso et al. used the plume as the information source and examined it with an infrared camera to rapidly discover processing defects and unstable states [59,60].

X-ray

Phantom V2512 by

Argonne National Laboratory, USA

55 keV

Vision Research Inc. 1.5–11 8000

— 50,000

*2.2. Invisible-Light In Situ Detection* 

through the melt pool.

Lumencor SOLA SM white light source

—

— 45,259–135,776 Spatter Al-Si10-

monochromatic X-rays 6.6 5100 Melt pool Invar 36 Leung et al. (2019) [57]

summary of X-ray in situ detection is shown in Table 2.

1 54,310 Powder spatter 316L/Al-Si10-

Spatter and

Melt pool and

2 400,000 Keyhole \* Ti-6Al-4V Cunningham et al. (2019)

\* Keyhole: also known as the depression zone, is wrapped by the gas–liquid interface and penetrates

For the invisible-light in situ detection of L-PBF, the imaging technologies mainly include X-ray imaging, schlieren video imaging, infrared imaging, and thermal imaging. X-rays have a short wavelength, high energy, and high penetration ability. Highstrength X-rays can penetrate a certain thickness of metal with high temporal and spatial resolution, which is the preferred method for many L-PBF spatter studies [35]. A schematic diagram of an X-ray system is shown in Figure 7. As one of the most productive Xray sources globally, the Advanced Photon Source (APS) in the Argonne National Laboratory provides experimental conditions for many researchers. More than 5500 researchers per year use X-rays produced by APS to do experiments. Many of those researchers use those X-rays to detect L-PBF spatter. For example, Zhao et al. [28] pioneered the use of high-speed X-rays (harmonic energy 24.4 keV) for in situ characterizations of L-PBF progress. Guo et al. [36] found transient spatter dynamics in L-PBF using a high-speed, highresolution, and high-energy X-ray imaging technique. Ross Cunningham et al. quantified the keyhole in Ti-6Al-4V powder during laser melting based on X-ray image information [55]. Leung et al. raised the X-ray power (monochromatic X-ray power: 55 keV) and studied stainless steel (316L) and 13–93 bioactive glass. They found that melt pool wetting and vapor-driven powder entrainment are key track growth mechanisms for L-PBF [57]. A

denudation 316L Biadre et al. (2018) [35]

spatter Ti-6Al-4V Zhao et al. (2017) [28]

Mg Guo et al. (2018) [36]

Mg/Ti-6Al-4V Young et al. (2020) [56]

[55]

**Figure 7.** Schematic of the high-speed X-ray imaging and diffraction experiments on laser powder bed fusion process at the 32-ID-B beamline of the Advanced Photon Source. A pseudo-pink beam with a first harmonic energy of 24.4 keV (λ = 0.508Å) is generated by a short-period undulator. The **Figure 7.** Schematic of the high-speed X-ray imaging and diffraction experiments on laser powder bed fusion process at the 32-ID-B beamline of the Advanced Photon Source. A pseudo-pink beam with a first harmonic energy of 24.4 keV (λ = 0.508Å) is generated by a short-period undulator. The laser irradiates the micro-powder bed sample from the top, the X-rays penetrate from the side of the sample. The imaging and diffraction detectors are placed approximately 300 mm downstream from the sample. The inset surrounded by the dashed circle enlarges the view of the laser-sample and X-ray-sample interaction. The distance of each component from the source is labeled on top. (Reprinted with permission from Ref. [28]. Copyright 2017 Springer Nature).

#### *2.3. Data Processing during Spatter Detection*

Spatter image obtained from in situ detection requires post-processing to enable the extraction and analysis of spatter behaviors.
