Powder bed fusion (PBF) is an additive manufacturing (AM) technique that holds a great promise for alumina ceramic materials to be processed in a one step. To ensure an effective outcome, the powder material should be thoroughly tailored, and the process parameters should be appropriately investigated. These process parameters include laser power, scanning speed, hatching space, and scanning strategies. The alumina powder has been tailored and modified to be appropriately used for PBF using the spray-drying technique, and the process parameters have been predicted and selected using a developed numerical model. Different scanning speeds of 100, 200, 300, and 400 mm/s have been considered, and the other parameters have been numerically predicted. The results demonstrated that spray drying is an effective technique for tailoring the characteristics of alumina powder, such as particle shape, particle size distribution, flowability, and absorptivity, making it ideally suited for PBF processing. Furthermore, the developed numerical model demonstrated outstanding reliability in predicting the most effective laser power and hatching space for different scanning speeds, resulting in significant cost and time savings when compared to relying solely on experimental trials. Employing a scanning speed of 400 mm/s yielded a significant improvement in relative density and quality of the printed samples surpassing other scanning speeds. Moreover, this speed effectively addressed various challenges encountered by other scanning speeds. Following the optimization of process parameters, it was determined that a relative density of 94.5% could be achieved by utilizing a scanning speed of 400 mm/s, a laser power of 210 W, and a hatching space of 30 µm. However, the evaluation of mechanical performance revealed that while the microhardness of the printed alumina samples matched the values reported in the literature, the attained compressive strength fell significantly below the values reported in the literature.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
1 Introduction
Ceramics are regarded as one of the most important classes of materials on account of their robust physical and mechanical properties, which make them highly desirable in advanced markets such as aerospace, defense, automotive, and medicine [1]. Due to the recent manufacturing revolution and the emergence of Industry 4.0, conventional ceramic manufacturing techniques [2‐5] are unable to meet the changing demands, including the production of extraordinarily complex designs. Consequently, it became necessary to investigate alternative techniques for processing ceramic materials.
Additive manufacturing (AM), as advanced technology, can not only overcome the aforementioned obstacles but can also significantly improve ceramic production by providing a short and rapid manufacturing cycle as well as the flexibility to manufacture highly complex designs with increased reliability [6]. In addition, AM can ensure the homogeneity of structures as it permits a more efficient mixing of materials. AM has demonstrated its efficacy as a highly advanced technology due to the vast advantages it can offer [7, 8]. These advantages include design flexibility, rapid prototyping, a short manufacturing route, and cost-effectiveness, and they have allowed AM to quickly penetrate numerous high-tech industries, such as aerospace, defense, automotive, and biotechnology.
Anzeige
For the AM of ceramics, several techniques, including binder jetting (BJ), fused deposition modelling (FDM), stereolithography (SLA), powder bed fusion (PBF), and robocasting (direct ink writing), can be used. These techniques have their advantages and disadvantages when applied to the AM of ceramics based on their specifications. Moreover, the properties of printed parts are determined by the characteristics of the employed technique.
PBF is one of the earliest AM techniques and remains one of the most versatile, being suitable for polymers, metals, ceramics, and composites to a lesser extent. There are a growing variety of machines for fusing powders utilizing various energy sources. The most active area of development is laser-based metal PBF processes. As a method for direct manufacturing, PBF processes are of great interest in several of industries [9]. Figure 1 is a diagrammatic representation of the PBF technique.
Fig. 1
Schematic representation for the PBF technique [60]
×
PBF can be utilized in two ways for ceramic AM: directly and indirectly. Indirect PBF (In-PBF) uses a mixture of ceramic powder and other additives that act as a binder, followed by a laser system to selectively scan the powder, layer by layer, according to the 3D CAD model [10, 11]. The obtained part is called the “green part” because it contains ceramic powder with a binding material. Postprocessing includes debinding (to remove the binder) and sintering (to consolidate the ceramic particles) required to achieve the final shape. This method is fraught with difficulties, including a lengthy powder preparation process, posttreatment, shrinkage, and degradation [10, 12‐15]. In contrast, direct PBF (D-PBF) utilizes a ceramic powder, without any additives, as feedstock, and then, the laser scans the powder, layer by layer, based on the 3D CAD model. D-PBF has excellent potential for the additive manufacturing of ceramics, where final shapes can be obtained without postprocessing. However, D-PBF faces numerous obstacles, including cracking, balling phenomena, laser interaction with powder (absorptivity), and powder morphology [16‐18]. In addition, the process parameters (laser power, scanning speed, hatching distance, and scanning strategies) should be investigated in depth to determine their optimal values.
The cracking of ceramic materials is regarded as the greatest challenge for D-PBF of ceramics, as it alters the mechanical performance and prevents the use of ceramics in a variety of applications, such as structurally resistant applications. According to Zheng et al. [19], who studied the cracks developed during D-PBF of alumina, there are primarily two types of cracking that occur during AM of ceramics using the D-PBF technique: longitudinal (parallel to the laser path) and transverse (perpendicular to the laser path). The transverse cracks were caused by the high-temperature gradient along the scanning path, whereas the longitudinal cracks were caused by the solidification progress along the scanning path. Several previous studies addressing this obstacle, were conducted to overcome the problem of thermal shock and cracks. Hagedorn et al. [20] developed a preheating system to preheat the layer powder temperature with a CO2 laser before scanning with an Nd-YAG laser to reduce thermal shocks and cracks. They could print different alumina layers and discovered that cracks were significantly diminished but still present. The cracks were caused by the deposition of cold powder on a previously scanned heated layer. This demonstrates the necessity of preheating the powder in the feeder tank prior to deposition to avoid such complications. Wilkes et al. [21] examined the potential AM of zirconia and alumina mixtures using the preheating system described by Hagedorn et al. [20]. The outcomes demonstrated that preheating could virtually eliminate the occurrence of cracks in printed samples. Liu et al. [22] developed a preheating system for yttria-stabilized zirconia (YSZ) powder using the same concept as Hagedorn et al. [20]. Before the powder bed was scanned with a fiber laser, it was preheated with an Nd-YAG laser. They were able to print YSZ components with a relative density of 84% and fewer cracks. As a result of the preheating, however, the printed components were encased entirely in sintered powder.
Anzeige
Balling phenomenon is also a severe issue caused by factors such as improper process parameters, high surface tension, and viscosity of molten material. Qiu et al. [18] investigated alumina D-PBF balling phenomena and cracks. Due to the high surface tension and viscosity of molten alumina, they discovered that the top surface of the printed samples exhibited balling phenomena. The high surface tension of alumina prevents the molten particles from merging, and the high viscosity prevents the molten alumina from spreading through the surface. In addition, they demonstrated that the scanning speed and laser power play a significant role in regulating the balling phenomenon.
Ceramic-laser interaction, particularly in terms of absorptivity, presents a significant difficulty [23]. Oxide ceramics, for instance, exhibit exceptionally low absorption of Nd-YAG or fiber lasers, as low as 3%, whereas carbide ceramics exhibit high absorption [24, 25]. Since all commercially available PBF printers are equipped with Nd-YAG or fiber laser, the absorptivity of the oxide ceramics should be increased for the PBF technique to be successful. There are several techniques for increasing ceramic powder’s absorptivity, such as using absorptivity enhancers and calcination which employs a small amount of another material to increase absorptivity without altering the properties of the powder [26, 27]. Numerous prior studies have employed this technique to print ceramic powder using the PBF technique. For instance, Juste et al. [23] examined the AM of alumina using the D-PBF method. The printer utilized an Nd-YAG laser with a wavelength of 1.064 µm, and the absorption of this laser by alumina was only 3%. Combining alumina powder with a small amount of graphite (0.1 vol%) and spray drying the mixture increased the absorptivity to 50%. Liu et al. [17] investigated the D-PBF of alumina by enhancing the alumina absorptivity with boron carbide (B4C). The results demonstrated a significant improvement in alumina sintering, densification, and the window for process parameters was widened. However, increasing the amount of B4C in the printed alumina samples resulted in darker-colored samples. Although the technique developed by Liu et al. [17] yielded promising results, the low relative density and shrinkage were the most significant shortcomings.
Powder morphology is an important consideration that should be addressed in PBF of ceramics as it significantly influences flowability of the powder on the printer bed and, consequently, the density of the printed parts [28, 29]. It has been reported that the spherical powder shape is highly recommended in PBF due to its superior flowability compared to other shapes whose interaction between powder particles results in poor flowability [15]. Ceramic materials can be converted into spherical powder via spray drying or high-temperature plasma technology. Spray drying produces a fully spherical powder shape with a controlled particle size distribution, but the obtained density is relatively low due to the porous structure of the powder particles [30]. In contrast, plasma technology can produce a solid-spherical powder shape, significantly increasing the density of 3D-printed ceramic shapes; however, this technology is expensive [31, 32].
Process parameters in the PBF of ceramics play a crucial role and have to be investigated thoroughly. These parameters include laser power, scanning speed, hatching space, layer thickness, and scanning strategies. It is important to address these parameters in-depth, as they can potentially control and mitigate the previously mentioned challenges. However, it has not been exhaustively studied, and most studies have utilized random values for these parameters. Shishkovsky et al. [33] utilized PBF to investigate the direct AM of zirconia-aluminum composites (in some cases, alumina). They examined the effect of various process parameters on the obtained monolayer (one path) and discovered the importance of using the correct parameter values. This study is limited by its emphasis on AM of composite materials instead of ceramics. Fayed et al. [34] examined the D-PBF of monolayer alumina using a 50 MPa compaction die to form a 3 mm thick, 50 × 50 mm square layer in a separate study. After the layer was sintered, its surface roughness, thickness, deformation, density, porosity, and hardness were evaluated using various process parameters. Laser power and scanning speed were found to affect the quality of the sintered layer significantly. Moreover, Fan et al. [35] investigated the D-PBLSP of alumina experimentally and numerically, considering only a single path using a numerical model developed specifically for this study. The feedstock was 99.8% pure alpha-alumina powder with a d50 of 20 µm. They could obtain a process map detailing the laser power and scanning speeds necessary for continuous alumina melting. The flaw in this study is that it only considered a single track and not the entire layer or component, which is an entirely different process involving numerous obstacles such as heat accumulation, layer deposition, adhesion, thermal shock, and cracks.
Simulation models were used to investigate the D-PBF of ceramics and to predict the appropriate values for the process parameters in order to reduce experimental costs and efforts and to interpret the experimental findings. Zhang et al. [36] used simulation to study alumina’s thermal behavior and solidification during D-PBF by creating a finite element model (FEM). They discovered that the temperature of the printed component gradually rises over time because of heat accumulation during the printing process, which corresponds with the thermal camera measurements of the temperature history during scanning. In addition, the results demonstrated the significance of laser power and scanning speed on the obtained melt pool dimensions. Chen et al. [37] investigated the effect of different process parameters on temperature distribution, melt pool profiles, and bead shapes during D-PBF of alumina by developing a three-dimensional finite element thermomechanical model. They could obtain the transformation of the melt pool’s shape as a function of various process parameters; it is evident that increasing the scanning speed led to an unstable melt pool.
Upon reviewing the current literature on D-PBF of ceramics, it can be concluded that this technique has received insufficient attention, necessitating a more thorough investigation to overcome the previously described challenges. The present research investigated three challenges: ceramic-laser interaction, powder morphology, and process parameter investigation. The study considered the spray drying and provided a comprehensive explanation of how to tailor ceramic material powder for the D-PBF using this technique, thereby addressing the challenge of material laser interaction and powder morphology. For process parameter investigation, simulation and optimization techniques were utilized to conduct a comprehensive investigation and optimization of these parameters. As being a technical ceramic material with exceptional mechanical and physical properties, including high strength and hardness, electrical insulation, corrosion resistance, biocompatibility, and heat resistance [38], and widely used in advanced applications [39‐43], aluminum oxide (alumina) has been chosen as the subject of this study.
2 Materials and methods
2.1 Numerical method
The numerical model used in this study has been developed and thoroughly described in our previous publication [14, 44]. User-defined functions (UDF) were incorporated into the model to simulate the laser heat source and temperature-dependent material properties. Furthermore, the model accounts for the melting and solidification phenomena that occur during the scanning procedure. For additional information about the governing equations, boundary conditions, assumptions, and computational domain, please refer to our previous study [14, 44].
2.2 Experimental method
2.2.1 Alumina feedstock
The feedstock used in this study is alumina powder with purity of 99% (P172LSB, Alteo, France). Figure 2 depicts the PSD for alumina powder (Al-raw) where it can be observed that the Al-raw has a mean particle size distribution of 390.6 µm and was agglomerated in irregular large accumulations, which is completely unsuitable for powder bed AM techniques due to the fact that this high PSD and agglomeration may present difficulties in terms of its uniform spreading on the printer powder bed [15]. This is predominantly because agglomerated particles inhibit the flowability of particles during powder bed AM techniques [45]. Furthermore, the interparticle friction generated by these irregular particles affects the flowability. Powder bed AM techniques recommend a spherical powder shape and a PSD within the range of tens of microns to ensure low friction between particles and good flowability during the powder spreading on the powder bed. Furthermore, due to its limited absorptivity for the printer laser type (Nd:YAG), alumina powder is not appropriate for the D-PBF. Implementation of the spray-drying technique, which is explained in the following section, has effectively addressed all these challenges.
Fig. 2
As-received alumina powder. a Particle size distribution. b Morphology
×
2.2.2 Spray-drying technique
This study employed spray-drying technique to increase the absorptivity of alumina feedstock and alter the morphology for the D-PBF. Spray drying relies primarily on producing dry granules from a slurry as it rapidly dries droplets with hot gas and pressure. The entire process consists of four steps: preparation of the powder slurry, atomization of the feeding slurry, contact of the slurry droplet with air, droplet drying, and separation of the dried particles from the hot gas [46]. Typically, ceramic powder, water, and dispersant (with specified wt%) are combined to create powder slurry. Powder slurry was produced by mixing 1200 g alumina powder with 44.8 wt% water and 1.0 wt% carboxylic acid dispersant (Dolapix CE64, Zschimmer & Schwarz, Germany) relative to the alumina powder weight. The slurry was then ball milled with alumina balls for 12 h to disperse agglomerates. Figure 3a depicts the preparation of slurry graphically.
Fig. 3
Graphical representation for the alumina powder slurry preparation method (a) and for the spray dryer (Niro, GEA, Germany) used in this study
×
A graphite-based colloidal suspension was added to the powder slurry following the milling step and mixed uniformly to introduce additives into the powders to increase the powder absorptivity. The volume percentage of graphite (relative to the volume of ceramics) is determined to be 0.1 vol%. After preparing the powder slurry, the well-dispersed particle suspension is pumped into a nozzle, and the liquid feedstock is expelled as a spray of droplets. Droplets are dried in a chamber where the liquid phase evaporates, forming dried granules. Spray drying was conducted with a Niro machine (Niro, GEA, Germany) which is depicted graphically in Fig. 3b.
Several variables, including inlet temperature, solid loading of the suspension, feeding rate, and organic content, impact the quality of the spray-drying process and the properties of the final dried feedstock. In this study, the inlet temperature ranged from 210 to 225 °C, while the outlet temperature was maintained at approximately 95 °C. The intake air flow rate was 60 L/min, and the suspension input rate was 25 rpm.
Spray drying produces two primary particle categories, main fraction and cyclone fraction, from the drying chamber and the cyclone, respectively. After spray drying, both the main fractions were dried at 100 °C to remove any remnant moisture, as this is a crucial phase because the powder’s flowability is dependent on its moisture content. Moisture can create bridges between powder particles, reducing their flowability. After sieving at 100 µm, the main fraction was divided into two portions to ascertain the influence of granule size. Al-MF refers to the fine fraction, whereas Al-MC refers to the coarse fraction (powder sizes larger than 100 µm). Various characterization techniques, including SEM images, particle size distribution (PSD), and a flowability test, were used to evaluate the spray-dried alumina powder.
2.2.3 Powder bed fusion
The SLM 125 printer manufactured by Renishaw® (UK) was utilized for the D-PBF of alumina. Table 1 provides the specifications of this printer. Argon gas was utilized to prevent oxidation during the printing procedure. Regarding the process parameters, 100, 200, 300, and 400 mm/s were used as scanning speeds. For each investigated scanning speed, the developed numerical model was used to predict the appropriate laser power range and hatching space. Due to printer-related technical limitations, the layer thickness utilized in all cases is 100 µm and could not be altered. As demonstrated in our earlier study [44], the zigzag scanning strategy, rotated by 45°, as demonstrated in Fig. 4a, yielded promising results for alumina D-PBF. Consequently, this scanning strategy was considered in this study. To ensure appropriate adhesion of the first printed layers to the metallic baseplate of the printer, it is essential that the material of the baseplate be compatible with the printer metallic baseplate. Figure 4b depicts the attachment of an alumina-based material baseplate to the printer’s metallic baseplate for the alumina D-PBF process.
Table 1
Renishaw® PBF 125 printer specifications
No
Item
Value/description
1
Laser
Nd-YAG
2
Wavelength
1070 nm
4
Laser power
Up 200 W
5
Laser spot size
35:200 µm
6
Building volume
125 × 125 × 125 mm3
7
Inert gas
Argon
Fig. 4
Scanning strategy (a) and alumina-based material baseplate used for D-PBF of alumina (b)
×
2.2.4 Characterization techniques
Various characterization and evaluation techniques were utilized to evaluate the obtained feedstock and the printed samples. The powder’s particle size distribution was measured using the Mastersizer MS 2000 (Malvern Panalytical®), while the powder’s morphology was examined using a scanning electron microscope (SU500, HITACHI, Japan). Archimedes’ method was utilized to determine the relative density of the printed alumina samples, and the SEM (SU500, HITACHI, Japan) was utilized to examine the structure of the samples. Moreover, the surface quality and surface irregularity of the samples were evaluated using a 3D optical scanning microscope (VKX-250, Keyence, Japan). X-ray diffraction (XRD) was used to identify the phases present in the sample; BURUKER D8-Advance (D8-Advance, BURUKER, US) was used for this purpose. For the mechanical performance evaluation of the printed samples, Vickers indentation microhardness and compressive testing were utilized. The microhardness test was performed using a Vickers microhardness tester (HM-200, Mitutoyo, Japan), while the compressive tests were performed using a Z100 Universal testing machine (Zwick, Germany) with a 100 kN cell force, and the upper head of the machine was descended at a rate of 1 mm/min until failure occurred.
3 Results and discussion
3.1 Powder characterization
The PSD and powder morphology of the coarse fraction (Al-MC) and the fine fraction (Al-MF) were analyzed to evaluate each powder category. The PSD and scanning electron microscope (SEM) images of each powder fraction are shown in Fig. 5. Both the Al-MC and Al-MF powders exhibit a spherical powder morphology, which is extremely appropriate for the PBF process and verifies the efficacy of the spray-drying technique in producing spherical granules. However, the Al-MC exhibits a PSD distribution of approximately 109.2 µm, which is deemed unsuitable for the 100 µm layer thickness employed in this study. The majority of the material would be deposited in the printer’s excess tank, making its use in the D-PBF of alumina impractical. Consequently, the Al-MC particle category was not used as a feedstock in the D-PBF of alumina in this study. In contrast, the mean PSD of Al-MF powder is 35 µm, making it an ideal feedstock for PBF technique [15]. In addition, the powder morphology exhibits precisely spherical shapes, which facilitates free-flowing behavior during layer spreading over the printer powder bed. Therefore, Al-MF powder was chosen as the feedstock for the alumina D-PBF in this study.
Fig. 5
PSD and powder morphology for alumina powder obtained from the spray-drying technique a, c for the Al-MC fraction and b, d for the Al-MF fraction
×
The flowability of Al-MF powder was evaluated with a Hall Flowmeter in accordance with ASTM standard (B213-13) [47]. The Hall Flowmeter does not provide a specific flowability criterion, but it does enable the comparison and evaluation of various particles. It measures the amount of time required for 50 g of powder to pass through a standard funnel aperture, with the top and bottom of the hopper exposed to the air and equal pressure at both locations. This measurement estimates the typical mass flow rate [48, 49]. In the case of the Al-MF powder, the passage of 50 g of powder through the orifice of the Hall Flowmeter funnel was performed three times. The average duration was determined to be approximately 140.34 s, which is slightly longer than the time it takes for metallic powder to pass through the orifice, as detailed in the standard [47]. This prolonged flow time may present difficulties during the powder layer’s spreading on the powder substrate. Therefore, additional tests were conducted to evaluate the powder deposition on the printer substrate. Figure 6a demonstrates that the Al-MF powder was effectively and uniformly deposited without defects on the printer substrate. Despite the slightly prolonged flow time, the Al-MF powder demonstrated adequate spreadability and flowability during the powder deposition procedure. Regarding the effect of graphite on the absorptivity of the Al-MF powder, the absorptivity of the powder was measured with spectrometry, before and after the spray-drying procedure. As shown in Fig. 6b, the addition of graphite considerably increased the absorptivity of the powder, which reached nearly 50%.
Fig. 6
Spray-dried alumina (Al-MF). a Deposition on the printer baseplate. b Absorptivity enhancement after the spray-drying technique and graphite addition
×
3.2 Process parameter investigation
As previously stated, to obtain successful D-PBF outcome with ceramic materials, it is essential to investigate the process parameters and determine their optimal values. These parameters include laser power, scanning speed, hatching distance, layer thickness, and scanning strategies. The numerical model was used to predict the appropriate values for these process parameters to provide initial guidance for the experimental investigation.
3.2.1 Numerical investigation
To address the challenges caused by the low density of alumina material in D-PBF, low scanning speeds of 100, 200, 300, and 400 mm/s were considered and employed. The primary objective was to prevent the dispersion of powder particles, which can result from the increased inertia of the laser beam caused by high scanning speed.
Using the numerical model, the appropriate laser power range for each scanning speed was determined. The numerical model’s calculations led to the selection of laser power speeds that satisfied multiple criteria, including melting the entire layer thickness, assuring adhesion to the layer beneath, and maintaining a maximum temperature below the alumina evaporation limit. The laser power range that met the specified requirements for a scanning speed of 100 mm/s was determined to be between 50 and 65 W, as shown in Table 2, which presents the melting contour and temperature distribution at different locations. This power range could achieve melting of the layer and its adhesion to the layer beneath. Due to the dense layer (100 µm), the obtained temperature surpassed the alumina’s boiling point. However, given the rapidity of the laser scanning process, which happens in microseconds, the elevated temperature achieved may not be substantial enough to produce defects in the printed samples. According to the developed numerical model, the predicted laser power range was as follows: 95 to 105 W, 120 to 130 W, and 150 to 170 W for scanning speeds of 200 mm/s, 300 mm/s, and 400 mm/s, respectively. Tables I, II, and III in the supplementary material show the melting contour and temperature distribution at various locations for scanning speeds of 200 mm/s, 300 mm/s, and 400 mm/s, respectively.
Table 2
Melting contour and temperature distribution at different positions for scanning speed of 100 mm/s
The hatching space, which determines the adhesion and contact between adjacent scanning paths, has a significant impact on the mechanical and physical characteristics of printed samples. As a result, careful value investigation is required for this significant parameter. The developed numerical model was used in this work to evaluate several hatching spaces, specifically 100, 75, and 50 µm. For each scanning speed, the average value of the previously determined laser power range was used. Table 3 shows the melting path that was attained for each scanning speed.
Table 3
The upper layer melting contour at different hatching space
The hatching space investigation revealed that the width of the melting path decreases with increasing scanning speed. This is due to the fact that low scanning speeds allow the laser heat source to interface with more particles, allowing powder particles to absorb more heat from the laser beam. Higher scanning speeds, on the other hand, do not allow powder particles sufficient time to absorb more heat from the laser beam, resulting in a narrower molten path. Table 3 shows the effect of hatching space values for each scanning speed on the adhesion of adjacent paths. When a hatching space of 100 µm was utilized, the scanned paths were unadhered for all employed scanning speeds, suggesting that this value is not recommended. With a hatching spacing of 75 µm, the scanning paths only adhered together at a scanning speed of 100 mm/s, whereas other scanning speeds could not adhere the melting paths together. However, when a 50 µm hatching space was used, the melting paths were successfully adhered together at all investigated scanning speeds.
3.2.2 Experimental investigation
After evaluating the influence of various process parameters and determining the appropriate values for each parameter using the developed numerical model, it became necessary to conduct an experimental investigation for the D-PBF of alumina to evaluate each parameter efficiency value, especially the scanning speeds. Various scanning speeds, including 100, 200, 300, and 400 mm/s, as specified previously, were investigated, and alumina samples were printed using the obtained results from the numerical model with the linear 45° scanning strategy, as specified in our previous study [44]. Figure 7 demonstrates that alumina cubes (10 × 10 × 10 mm3) were successfully printed using different scanning speeds and laser powers, as determined by the numerical model. This demonstrates the ability of the numerical model to predict the appropriate process parameters. Two samples were printed at each laser power level to evaluate the printability and repeatability. The samples have a clearly defined cube shape with no discernible printing defects.
Fig. 7
Alumina samples printed using different scanning speeds and laser power as predicted from the numerical model
×
The relative density measurements were used to assess the quality of the printed samples at various scanning speeds and laser powers. The density of the samples was calculated using Archimedes’ technique, with three readings obtained for each sample and the average taken into account. Figure 8 depicts the relative density obtained at each condition where it is clear that the relative density increases as the laser power increases across all scanning speeds. This is mostly due to an increase in laser energy density (LED), which is defined as LED = P / (v × L × h), where P is laser power, v is the scanning speed, L is the layer thickness, and h is the hatching spacing. The increase in LED resulted in greater densification of the printed samples.
Fig. 8
Relative density of alumina samples at different scanning speeds: 100 mm/s (a), 200 mm/s (b), 300 mm/s (c), and 400 mm/s (d)
×
The highest relative density for scanning speeds of 100 mm/s, 200 mm/s, 300 mm/s, and 400 mm/s were 69.5%, 81.6%, 77.3%, and 84.2%, respectively. It is worth noting that, even though the laser energy density remained relatively constant, increasing the scanning speed resulted in increased relative density in the printed samples.
To further understand the effect of scanning speeds on relative density, several alumina samples were printed at scanning speed of 100 mm/s, 200 mm/s, 300 mm/s, and 400 mm/s, as shown in Fig. 9a. The laser power was tuned to maintain a consistent laser energy density of 90 J/mm3 and 100 J/mm3 at each scanning speed. When the laser energy density was held constant for each scanning speed, the relative density increases as the scanning speed increased. The maximum relative density was obtained at a scanning speed of 400 mm/s, as shown in Fig. 9b. This behavior is consistent with the findings of Juste et al. [23], who investigated the PBF of spray-dried alumina. However, several earlier research on metallic materials found that the relative density increases as the scanning speed decreased [50, 51]. This disparity between metallic materials and alumina needs additional examination. To acquire a better understanding of why the relative density of alumina increases with high scanning speeds, it is vital to attentively study the behavior of the powder bed while the laser scans the powder. This will assist in shedding further light on this finding and give an in-depth understanding of the link between scanning speeds and relative density.
Fig. 9
Alumina samples printed at different scanning speeds with fixed laser energy density of 90 and 100 J/mm.3 (a). Relative density of alumina samples at different scanning speeds and laser energy densities (b). Spray-dried alumina powder bed behavior at different scanning speeds (c)
×
Figure 9c shows a visual depiction of the powder bed after scanning the first layer, which was obtained by paused and opening the printer door to inspect the powder bed closely. The spray-dried alumina particles surrounding the scanning region were pulled towards the scanning area at high scanning speeds (400 mm/s), resulting in a greater level of densification, and this may be the fundamental reason for the increased relative density observed at high scanning speeds. However, no particle dragging was seen at moderate scanning speeds (100 mm/s), resulting in a reduced relative density. This observation is in line with the findings of Bidare et al. [52], who employed a high-speed camera to closely examine the behavior of powder particles while scanning at various speeds. They discovered that at high scanning speeds, aerodynamic drag drawn the powder particles towards the laser point, resulting in enhanced densification.
Figure 10a illustrates alumina samples produced with a laser energy density of 90 J/mm3 and different scanning speeds. A 3D laser scanning microscope (VKX-250, Keyence, Japan) was also used to examine the top surface’s 3D heights in detail. The sample’s build direction resolution is considerably diminished due to the 100 µm layer thickness. Notably, the top surface displayed non-flat characteristics for scanning speeds of 100, 200, and 300 mm/s, confirming the presence of the balling phenomena reported by Qiu et al. [18]. This effect is caused by the molten ceramic materials’ high viscosity and surface tension. However, at a scanning speed of 400 mm/s, the balling phenomena were missing due to the high scanning speed generating an enhanced draw of alumina powder particles surrounding the scanning region towards the laser scanning area, enhanced by aerodynamic drag as described previously. As a result, the alumina sample printed at 400 mm/s included more molten alumina material than samples printed at other speeds, resulting in a flat surface free of flaws exhibited at other scanning rates. Figure 10a also shows an interesting observation: the sample color darkens as the scanning speed increases. The sample printed at 100 mm/s has an off-white color, while the one produced at 200 mm/s has a significantly darker appearance. The darkness of the sample grows as the scanning speed increases up to 300 mm/s, ending in a completely dark appearance at 400 mm/s. Figure 10b shows a detailed view of the top surfaces of alumina samples printed using various scanning strategies obtained using the 3D laser scanning microscope (VKX-250, Keyence, Japan) while maintaining a consistent laser energy density of 90 and 100 J/mm3. Notably, regardless of the constant laser energy density, the darkness of the samples increases with increasing scanning speed.
Fig. 10
Alumina samples and the corresponding 3D surface height at different scanning speeds with a fixed laser energy density of 90 J/mm3 (a). Optical microscopic top surface of alumina samples printed at different scanning speeds at a fixed laser energy density of 90 and 100 J/mm.3 (b)
×
The laser energy density is often used as a straightforward comparison and design parameter in PBF. It takes into account aspects such as laser power, scanning speed, layer thickness, and hatching space, resulting in a single metric that may be connected with different PBF process output characteristics such as relative density, porosity, surface roughness, and flaws [53]. Numerous previous investigations on metallic PBF have found a significant relationship between laser energy density and process output characteristics as density, porosity, surface roughness, defect development, and sample color [54, 55]. However, based on prior study findings, it is clear that the relative density, surface texture, and sample color of alumina samples vary despite the laser energy density being constant. The intricate physics involved in the melt pool during the PBF technique, including phenomena such as Marangoni flow, hydrodynamic effects, and recoil pressure, impose limitations on the efficacy of laser energy density as a comprehensive design parameter for accurately describing the PBF process, according to Bertoli et al. [53]. This constraint was also seen in the current work employing alumina PBF, where the laser energy density failed to capture the process’s output efficiently, necessitating the development of a new parameter to address this intrinsic shortcoming.
Figure 11a shows SEM images of alumina samples scanned at various speeds while keeping a constant laser energy density of 90 J/mm3. Due to the poor connection between the melted particles, the surfaces of the samples printed at scanning rates of 100 and 200 mm/s were difficult to polish for SEM imaging. As a result, only samples printed at scanning speeds of 300 and 400 mm/s were effectively polished. SEM investigation of samples printed at scanning speeds of 100, 200, and 300 mm/s showed noticeable particle balling, as well as high porosity and the presence of cracks. However, the microstructure of samples printed at 400 mm/s demonstrates superior characteristics when compared to other scanning speeds, indicating a lack of balling while still presenting obvious cracks similar to the other speeds. These cracks are caused by thermal shocks and developed residuals stress during the PBF process and might potentially be minimized by using the preheating systems outlined in the introduction section.
Fig. 11
SEM images for the top surface of alumina samples printed at a fixed laser energy density of 90 J/mm3 (a) and the side surface of alumina sample (along the build direction) printed with 400 mm/s scanning speed at a fixed laser energy density of 90 J/mm.3 (b)
×
Figure 11b shows SEM images of the side surface of an alumina sample printed at a scanning speed of 400 mm/s. The layers attach strongly with no visible separation. However, cyclic deterioration along the building direction of the sample is visible, resembling a brick wall pattern with a 500-µm frequency. This cyclic damage happens after every five layers of printing and is most likely caused by the movement mechanism of the building’s baseplate. This mechanism lowers the baseplate after each layer printing to allow for the deposition of the next layer.
To investigate the influence of the graphite addition on the spray-dried process, XRD analysis was utilized to identify the phases inside the alumina-printed sample, and Rietveld analysis was used to quantify these phases. Figure 12 depicts the XRD spectra for alumina raw powder (Alumina-P172LSB), alumina spray-dried powder, and an alumina sample printed using the D-PBF (400 mm/s, 180W, and 50 µm). Neither the spray-drying approach nor the D-PBF technique changed the XRD spectra due to the production of additional phases. Rietveld analysis was performed to determine the amount of each phase inside each case by refining the XRD spectra for each case. Table 4 summarizes the results of the Rietveld analysis, which utilized alpha-alumina and graphite phases for the refining. In all situations (alumina raw powder, spray-dried alumina powder, and alumina-printed sample), only the alumina phase was present, which explains why the XRD spectra in Fig. 12 remained unchanged. Therefore, the dark color of the D-PBF-printed alumina sample is primarily attributable to the chamber’s high oxygen content (5000 ppm) which may burn with graphite at high scanning speed (400 mm/s) and laser power (180 W).
Fig. 12
XRD spectra for alumina raw powder, alumina spray-dried powder, and alumina sample printed using D-PBF
Table 4
Alumina quantitative phase analysis
Item
Alpha-Al2O3
Carbon
Rwp (%)
Rexp (%)
Alumina-P172LSB, as received
100% ± 0.0
0.0 ± 0.0
10.99
4.62
Spray-dried alumina
100% ± 5.68e-8
0.0 ± 0.0
26.33
4.59
Alumina-printed sample
100 ± 8.14e-8
0.0 ± 0.0
30.37
6.04
×
From the scanning speed investigation, it can be concluded that the 400 mm/s scanning speed was able to overcome all the challenges encountered with the other investigated scanning speeds and that employing the 400 mm/s scanning speed for the D-PBF of alumina results in a higher building rate. Therefore, the scanning speed of 400 mm/s was considered for the process parameter optimization of alumina D-PBF in the coming section.
3.3 Process parameter optimization
After thoroughly understanding the effect of the process parameters, it is essential to optimize these process parameters to obtain the optimal values for achieving the desired properties, such as high relative density. Using a scanning speed of 400 mm/s, a linear 45° scanning strategy, laser power calculated from the numerical model, and hatching space of 50 µm or less are promising process parameter values for obtaining good-quality alumina samples using the D-PBF technique.
The Taguchi optimization technique can provide an effective and efficient procedure for determining the optimal parameters for D-PBF of alumina in order to attain the highest relative density and good surface quality. The Taguchi optimization technique consists of three distinct stages. The first stage is the design phase, which entails executing the system within which the experiments will be conducted, identifying all the factors (process parameters) that influence the process, determining the range of each factor (levels) included in the experiments, and identifying the response factors. The second stage is the excitation stage, which consists of two steps. In the first step, the orthogonal array (OA) is constructed based on the factors and level of each factor. The design of the OA should incorporate all feasible treatments that address all factors and levels under consideration. The second step involves conducting experiments following the OA and determining the response factors for each treatment. The second stage is the most crucial and should be executed with care. All possible factors should be considered during the design phase, as this greatly aids in the early identification of ineffective factors. The final stage is the analysis and optimization phase, which consists of response factor analysis (data analysis), determining the optimal factors value and conducting a confirmation test using the value of the optimal factors.
3.3.1 Taguchi optimization—first stage
The factors that influence the PBF of alumina have been identified and their levels based on the previous numerical and experimental investigation for the D-PBF of alumina, as obtained in Sect. 3.2. These factors include laser power and hatching space. In addition, the hatching space, as recommended by the numerical model, should be 50 µm. Since the scanning speed of 400 mm/s yielded promising results for the PBF of alumina, it was considered during the optimization treatments. The levels for each factor are shown in Table 5, where three levels were considered for each factor.
Table 5
Factors and levels used in the D-PBF optimization of alumina
Factor
Symbol
Level 1
Level 2
Level 3
Laser power (W)
A
180
190
200
Hatching distance (µm)
B
50
60
70
This optimization aims to successfully print alumina samples with the highest possible relative density and good surface conditions. Consequently, the relative density and the surface roughness were considered as response functions (factors) during the optimization of alumina PBF.
3.3.2 Taguchi optimization–second stage
Using a full factorial design of experiments resulted in 27 treatments, which is a time- and cost-intensive process, whereas the Taguchi optimization method can reduce the number of treatments in a way that guarantees to capture the optimal level for each factor. The standard orthogonal array (OA) L9 (33) was used to construct the treatments in this study. This array consists of 9 treatments (instead of 27 if the full factorial was used), including the 2 factors at different levels as described in Table 5. Table 6 shows the nine treatments (experiments) included in this optimization, combining different level conditions for each factor (E1–E9). The treatments considered a different level of laser energy density: high energy density (E1, E4, E7), medium energy density (E2, E5, E8), and low energy density (E3, E6, E9).
Table 6
Factor levels for each treatment in PBF of alumina
Treatment
A
B
E1
1
1
E2
1
2
E3
1
3
E4
2
1
E5
2
2
E6
2
3
E7
3
1
E8
3
2
E9
3
3
After obtaining the orthogonal array, the second stage consisted of conducting the experiments (treatments) outlined in Table 6. Figure 13 depicts the printed alumina samples. Two samples were printed for each treatment to obtain more accurate measurements and check the printability and repeatability. For each treatment, the relative density was measured using Archimedes’ method (the density was measured three times, and the average was considered), and the top surface texture and roughness were evaluated using the 3D laser scanning microscope (VKX-250, Keyence, Japan). Since the 3D laser scanning microscope employed accurate surface roughness detection via multiple line measurements and the average value was calculated and considered, as shown in Fig. 13b, only one measurement was considered for the surface roughness of each treatment. The number of lines used for the surface roughness measurement is 25 lines with 0.15 mm spacing between each two lines. The 3D surface height for each sample is shown in Fig. 13c. Table 7 summarizes the relative density and surface roughness measurements for every treatment. It can be observed that relative density increases as the laser power increases and hatching space decreases.
Fig. 13
Alumina samples printed according to the treatments described in Table 6 (a), surface roughness measurement through multiple lines as used in the optical microscope (b), and 3D top surface height of alumina sample at each treatment (c)
Table 7
Relative density and surface roughness of alumina sample at each treatment
Treatment
Levels of input factors
Measured response factor
A
B
Relative density
Surface roughness (Ra) (µm)
First reading
Second reading
Third reading
Average
Average
E1
1
1
0.87
0.86
0.87
0.87
54
E2
1
2
0.85
0.83
0.85
0.84
108
E3
1
3
0.77
0.75
0.78
0.77
112
E4
2
1
0.90
0.89
0.90
0.90
151
E5
2
2
0.84
0.85
0.85
0.85
146
E6
2
3
0.81
0.81
0.81
0.81
145
E7
3
1
0.87
0.92
0.92
0.92
132
E8
3
2
0.87
0.86
0.87
0.87
171
E9
3
3
0.79
0.79
0.79
0.79
180
×
3.3.3 Taguchi optimization—final stage
The final phase in the Taguchi optimization is devoted to response factor analysis (data analysis), determining the optimal factor value and conducting a confirmation test using the optimal factor value. To analyze the obtained data, the Taguchi optimization technique uses the signal-to-noise (S/N) response analysis to evaluate the quality of each treatment instead of the standard deviation. This is mainly because the standard deviation decreases or increases according to the mean. The S/N ratio mainly focuses on measuring the response factor’s variation to the nominal or target value under different noise conditions (uncontrollable factors). There are four categories for the S/N ratio calculation based on the desired output quality: smaller is the better, larger is the better, nominal is the better, and nominal is the best. For the D-PBF of alumina, the larger is the better was used for the relative density, as described by Eq. (1), while for surface roughness, smaller is the better was used, as described by Eq. (2) [56].
where \({Y}_{i}\) represents the individual measured relative density (first, second, and third readings for the relative density and the calculated surface roughness, as described in Table 7), while n represent the number of the reading (n = 3 for relative density and n = 1 for surface roughness). Table 8 summarizes the S/N ratio for each treatment’s relative density and SiC percentage.
Table 8
Calculated S/N ratio for each treatment used in alumina-PBF optimization
Treatment
Levels of input factors
Calculated S/N ratio
A
B
Density
Surface roughness
E1
1
1
− 1.25404
− 34.6479
E2
1
2
− 1.48772
− 40.6685
E3
1
3
− 2.36043
− 40.9844
E4
2
1
− 0.97368
− 43.5795
E5
2
2
− 1.45583
− 43.2871
E6
2
3
− 1.84450
− 43.2274
E7
3
1
− 0.76754
− 42.4115
E8
3
2
− 1.28461
− 44.6599
E9
3
3
− 2.06087
− 45.1055
In Taguchi optimization technique, the largest S/N ratio would represent the optimal response for the desired output (maximum relative density and low surface roughness). As described earlier, the S/N category used in this study for relative density is “the larger is the better” and for surface roughness is “the smaller is the better” which means that the largest relative density content and the smallest surface roughness value are desired. The S/N ratio response graphs for the relative density and surface roughness are shown in Fig. 14 The optimal combination of the process parameters can be determined from these graphs. It can be seen that the hatching space (factor B) is the most significant parameter, followed by the laser power (factor A), and to achieve the maximum relative density, the smallest hatching space value of 50 µm (B1) and the highest laser power (A3) should be used; i.e., treatment E7 should be used.
Fig. 14
S/N ratio response graph of relative density (a) and surface roughness (b)—PBLSP optimization of alumina
×
The optimal combination of the process parameter values can be determined from Fig. 14b for achieving a good top surface roughness. It can be seen that the laser power (factor A) is the most significant parameters followed by the hatching space (factor B), and the optimal process parameter values to achieve good surface conditions are a laser power of 180W (A1) and hatching space of 50 µm (B1); i.e., treatment E1 should be used.
3.3.4 Pareto ANOVA: an alternative technique
Pareto ANOVA is a technique used to analyze data for process parameter optimization, and it can also provide the percentage contribution of each parameter to the response functions straightforwardly [81,82]. The S/N response data for each response function is used to construct the Pareto ANOVA analysis. The S/N response data can be calculated by taking the sum of all S/N ratio values (as described in Table 8) at the same level as the input parameter. Table 9 summarizes the S/N response data values for relative density and surface roughness.
Table 9
S/N response data of the relative density and surface roughness—PBLSP optimization of alumina
Relative density (%)
Surface roughness
Levels
A
B
Levels
A
B
1
− 5.10
− 3.36
1
− 116.30
− 120.64
2
− 4.27
− 4.22
2
− 130.09
− 128.62
3
− 4.48
− 6.26
3
− 132.18
− 129.32
After calculating the S/N response data for each input parameter, the summation of squares of differences was calculated for each input parameter using the following equation:
where \({S}_{d}^{A}\) represents the squares of difference for the input parameter (laser power) A; similarly, the square of differences can be calculated for the hatching space. The percentage contribution for each input parameter was calculated by considering the percentage summation of the squares of differences to the total summation of the squares of differences for all input parameters. The Pareto diagram was plotted considering the obtained percentage for each input parameter. The input parameters were organized in Pareto diagram in a way that the parameter with highest contribution comes first and then is followed by other parameters based on their contributions. Table IV and Table V in the supplementary material summarize the Pareto ANOVA analysis for relative density and surface roughness, respectively.
Table IV shows that the hatching space is the most influential parameter for relative density, contributing 90.64%. The laser power’s contribution is then 9.36%. The optimal combination of input process parameters to achieve the maximum relative density is A3–B1. This is primarily since decreasing the hatching distance and increasing the laser power allow more alumina powder particles to be well consolidated and densified. For surface roughness, as described in Table V, the laser power contributed the most, 76.21%, followed by the hatching space, which contributed 23.79%. A1–B3 is the optimal combination of input process parameters for producing a surface with favorable characteristics. A low laser energy density was indeed required to achieve desirable surface conditions (high hatching space and low laser power).
From the optimization of the process parameters for PBF of alumina, it is evident that achieving a high relative density value will have an adverse effect on the sample surface quality. Consequently, the effectiveness of the process parameters to achieve higher density values and good surface roughness is limited, and alternative solutions should be considered. Clearly, powder improvement can play a significant role in achieving this objective. As described in Fig. 5c, the spray-dried alumina powder has a porous structure (the powder particle is hollow and not completely solid), which may be the primary cause of the poor surface conditions obtained with a high laser energy density used to maximize the density. When melted by the laser, this porous structure causes a large drop in layer level in specific locations, and the surface roughness increases. This can be compensated by using solid-spherical alumina powder produced by high-temperature plasma technology [15]. In addition, some PBF printers are equipped with a compaction cylinder that is integrated with the recoater (as available in the Phenix printer used for D-PBF of SiC [12‐14]). This compaction cylinder can be used to compensate for the porous structure of spray-dried powder, thereby allowing the falling of the layering level at certain locations to be significantly mitigated and the surface quality to be greatly enhanced. This is an important research topic that should be addressed in future investigations.
3.3.5 Confirmation test
Since the recommended optimal values of the process parameters, as determined by the Taguchi optimization method and Pareto ANOVA analysis, have already been considered, it is unnecessary to conduct a confirmation test for both the relative density and the surface roughness. However, as shown by the optimization of process parameters, the relative density increased with increased laser power and decreased hatching space. Therefore, examining the impact of increasing the laser power and decreasing the hatching space on the relative density is worthwhile. As depicted in Fig. 15a, alumina samples were printed with varying laser power, including 200 and 210 W (the laser power could not be increased above 210 W, as this is the highest value permitted by the printer) and hatching space, including 50, 40, and 30 µm. The alumina samples were successfully printed but with a wavy top surface. Table VI in the supplementary material shows the printed alumina samples, their top surface under an optical microscope, 3D top surface heights, and surface roughness value measured by optical microscope (VKX-250, Keyence, Japan) at 210 W and various hatching spaces. The top surfaces were clearly unlevel, wavy, and filled with defects. In addition, it can be observed that the surface roughness value decreased as the hatching space decreased, which corresponded to the conclusion provided in Sect. 3.2. Figure 15b depicts the variation in relative density with respect to laser power and hatching spacing for the alumina samples depicted in Fig. 15a. Observably, the relative density increased with an increase in laser power and a decrease in the hatching space, with a maximum relative density of 94.7% attained using 210 W of laser power and a hatching spacing of 30 µm.
Fig. 15
Alumina samples printed with 200 and 210 W laser power at different hatching spaces (a) and the measured relative density (b)
×
Figure I, in the supplementary material, shows SEM images of alumina samples printed with 210 W of laser power and 50, 40, and 30 µm of hatching space. The samples contain defects in the form of small holes and notches. In addition, there are numerous cracks dispersed over the surface because of thermal shock experienced during D-PBF. As previously explained, these cracks can be eliminated by preheating the powder layer prior to scanning.
Different alumina lattice structures, including gyroid, diamond, and primitive, were printed to demonstrate the capabilities of the D-PBF technology to print extremely complicated geometries, as shown in Fig. 16. The removal of the lattice from the baseplate led to the appearance of some defects in the samples’ base. In addition, the layers are visible along the building direction, primarily due to the 100 µm layer thickness used.
Fig. 16
Alumina lattice structures printed for demonstration purposes
×
3.4 Mechanical performance evaluation
The mechanical performance of alumina samples printed using the PBF process was evaluated utilizing microhardness and compression tests. The findings of each test were compared to those that had previously been reported in the literature. The Vickers indentation microhardness test was used to evaluate the hardness of PBF-printed alumina sample. To get a reliable microhardness measurement, three readings were recorded and the average measured microhardness was 2180 HV/10. The literature reports that the microhardness of alumina is roughly 2000 HV [57, 58], nearly identical to the hardness determined using the PBF approach in this investigation.
The compressive test was used to evaluate the compressive strength of printed alumina sample, and for this purpose, cylindrical alumina samples (10 mm in diameter and 25 mm in length) were printed with optimized process parameters (210 W of laser power, 400 mm/s of scanning speed, 30 µm of hatching distance, and 100 µm of layer thickness). Figure II(a), in the supplementary material, depicts the printed compressive samples before being unloaded from the printer baseplate. To prepare the alumina samples for the compressive test, the two opposite cross-sectional areas were hand-polished until they became flat and parallel, as shown in Figure II(b) in the supplementary material. The length of the samples was adjusted to 20 mm (L = 2D, where L is the length of the sample and D is its diameter). Tests were conducted on ten samples, and the average of all measurements was calculated. The obtained compressive strength was 140.8 ± 11.6 MPa, which is considered low compared to the compressive strength of alumina manufactured with conventional techniques, reaching 3000 MPa [59]. This is primarily due to the cyclic damage experienced along the printing direction, as described in Sect. 3.2 (Fig. 11) and the presence of pores and cracks in the printed sample. The maximum compressive strength reported for alumina ceramic material processed by AM techniques is about 1000 MPa. Figure 17a shows the stress–strain curve for one of the alumina-printed samples, and Fig. 17b depicts alumina samples that have been crushed after reaching their compressive strength limit.
Fig. 17
Stress–strain curve for alumina sample under compression test. Alumina sample failure behavior under the compressive test
×
4 Conclusions and future study
This study investigates experimentally and numerically the PBF of alumina. First, the alumina powder was tailored by using the spray-drying technique. The alumina powder was effectively modified through spray drying to attain the desired, spherical powder morphology, good flowability, and enhanced absorptivity. These modifications have considerably improved the powder’s compatibility for use in the D-PBF process, demonstrating the effectiveness of spray drying in optimizing powder properties for additive manufacturing applications. Then, the developed numerical model was employed to predict the initial values of the laser power, scanning, and hatching space at a different scanning speed of 100, 200, 300, and 400 mm/s, and the experimental procedure was employed to print alumina samples using the predicted values. Concerning the effects of scanning speeds, the results demonstrated that increasing the scanning speed positively impacted the quality of the printed samples. Using a scanning speed of 400 mm/s, 85% relative density alumina samples were obtained. The process parameters were then optimized using the Taguchi optimization technique, with laser power, scanning speed, and hatching space. The optimization identified the optimal parameters for the PBF of alumina, which enabled the printing of alumina with a relative density of 94.5%. Compressive and microhardness tests were used to evaluate the mechanical performance of the printed samples. For the microhardness test, the printed samples exhibited a hardness value comparable to that of alumina processed using conventional techniques. For the compressive test, the printed samples demonstrated a low compressive strength compared to conventional techniques. As depicted in the SEM images, the low performance obtained during the compressive test was due to cyclic damage along the building direction of the sample. Future research should consider the porous structure of the alumina powder obtained via the spray-drying technique in order to overcome the top layer’s irregular surface. In addition, an efficient preheating system should be developed in order overcome the cracking problem with ceramic D-PBF.
Declarations
Ethics approval
All applicable international, national, and/or institutional guidelines for the care and use of research and materials were followed.
Consent to participate
Not applicable.
Consent for publication
The authors give the publisher the permission to publish the work.
Conflict of interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Die im Laufe eines Jahres in der „adhäsion“ veröffentlichten Marktübersichten helfen Anwendern verschiedenster Branchen, sich einen gezielten Überblick über Lieferantenangebote zu verschaffen.
