2023 Volume 64 Issue 8 Pages 1946-1951
Binder jetting (BJT) holds enormous promise as an additive manufacturing technique due to its high throughput and low-cost equipment. The ability to fabricate geometrically complex parts from aluminum alloy via BJT would have a significant impact on the development of high-performance machine components and heat-dissipation devices. However, processing printed objects by sintering of aluminum alloy is difficult due to the surface oxide layer that covers the powder particles, which prevents the final parts from reaching a sufficiently high density. In order to investigate the sintering mechanism for controlling the quality of sintered parts, we directly observed the sintering behavior of pure Al and Al–10%Si–0.4%Mg (mass%) alloy using an in situ scanning electron microscopy system equipped with a heating stage. It is found that the surface roughness of the powder particles is reduced above their melting or solidus temperature. Subsequently, the liquid ruptures the oxide layer and forms necks between the particles.
This Paper was Originally Published in Japanese in J. Jpn. Soc. Powder Powder Metallurgy 68 (2021) 311–316.
Fig. 6 In situ SEM images of Al–10%Si–0.4%Mg powder at 800×: (a) before heating, (b) at 564°C (liquid fraction: 0 mass%), (c) at 569°C (liquid fraction: 7 mass%), (d) at 574°C (liquid fraction: 49 mass%), (e) at 594°C (liquid fraction: 100 mass%). (f) Temperature dependence of liquid mass fraction showing points (a)–(e).
Binder jetting (BJT) is an additive manufacturing (AM) technique that has been widely used to produce metal parts.1) In the BJT process, jetted ink containing a binder is deposited onto a powder bed, where it functions as glue for powder particles, forming green parts. Post heating burns out the binder in the green parts and then induces sintering of the particles. The BJT system can achieve high productivity using low-cost equipment. Although BJT has been applied to various metals such as stainless steel and Ni-based alloys, the production of Al parts by BJT has not been put into practical use because of the difficulty of sintering Al particles. The ability to manufacture geometrically complex parts from Al alloys via BJT would enable the mass production of high-performance machine components as well as heat-dissipation devices.
The common AM technique for fabricating Al parts is selective laser melting (SLM).2) In SLM, a laser beam selectively fuses powder particles within a powder layer to produce metallic components. The high reflectivity and thermal conductivity of Al alloy preclude the processing of Al parts with conventional SLM machines. In recent years, SLM machines equipped with a fiber laser, whose beam is effectively absorbed by Al, have achieved the fabrication of dense Al parts.3,4) Although SLM offers the ability to directly produce complex and fully dense metallic parts, it is not that suitable for mass production because of its low throughput and high equipment cost. BJT is promising for realizing the mass production of Al components, whose number generally tends to be large.
As previously mentioned, the difficulty associated with sintering Al is a substantial obstacle to the fabrication of Al products via BJT. Specifically, the stable surface oxide coating on Al powder particles inhibits solid-phase sintering, which prevents the green part from reaching a sufficiently high density. Increasing the green density by powder compaction5) and heating at high pressures by hot extrusion6) are effective methods to attain a sufficient sintered density. However, these methods, including the pressurization process, occasionally break complex green objects printed by BJT. A sintering technique without pressure is required to produce advanced components designed by AM.
Liquid-phase sintering has been proposed as a method to densify green parts without pressure.7) A certain amount of metal liquid forms inside the green part and promotes densification. Liquid-phase-sintering systems are classified into two types: (i) those that use a mixture of base and additive elemental powders and (ii) those that use prealloyed powder containing a base material and an additive (supersolidus liquid-phase sintering, SLPS8)). Si, Cu, and Mg are primary additives for the liquid-phase sintering of Al.9,10) A metal liquid that forms below the melting point of Al disrupts the oxide layer and facilitates neck growth between particles. However, some liquids are occasionally agglomerated inside a sintered part and others are exuded to the surface.7) This nonuniform distribution of the liquid phase causes an uneven density of the sintered part. Although the nonuniformity of a metal liquid depends on the wettability between the liquid and solid, details of the process remain unclear.
Several in situ observation techniques have been proposed to investigate the behavior during sintering. Lame et al. observed neck formation and particle rearrangement of Cu and steel by X-ray microtomography.11) However, X-ray microtomography cannot be used to investigate the change of a particle surface during liquid-phase sintering in detail because its spatial resolution is in the range of several to tens of micrometers; in addition, it uses cross-sectional images composed of transmission layers. The sintering behaviors of nanosized particles of materials such as ZrO2,12) Pt,13) and Ag14) have also been observed by transmission electron microscopy (TEM). Although the TEM technique can reveal even a change in the crystal structure inside a particle, it is unsuitable for observing micron-sized particles, which are commonly used for BJT. This limitation arises from the maximum thickness for TEM specimens, which is 100 nm.
Therefore, we here focus on using scanning electron microscopy (SEM) to observe the surfaces of micron-sized particles during sintering. In situ SEM has been used to observe the liquid-phase sintering of a powder compact composed of Al and Cu powders with particle sizes ranging from 700 to 900 µm.15) The authors reported that the Cu particles transform to a liquid and form a neck with neighboring Al particles after reaching 560°C, which is above the eutectic temperature of the Al–Cu system (548°C). For the liquid-phase sintering of a mixture composed of base metal and additive, the liquid transformed from the additive spreads into pores surrounded by solid grains, which leads to densification. By contrast, in SLPS, the liquid forms inside an alloy particle and spreads along grain boundaries, resulting in an interparticle bond.8) The literature contains no report of a direct observation of SLPS of an Al alloy. The behavior of SLPS is presumed to differ from that of liquid-phase sintering using a powder mixture.
In the present study, we directly observed the sintering process using in situ SEM to elucidate the mechanism of nonuniform liquid distribution during the SLPS of Al alloys. The sintering behavior is discussed on the basis of the SEM observations and the liquid fraction calculated under equilibrium conditions.
A scanning electron microscope (JSM-7200F, JEOL Ltd.) equipped with a heating stage was used to observe pure Al and Al alloy powders during heating in situ. A photograph of the heating stage and an SEM image of a powder sample pasted onto a stainless-steel plate are shown in Figs. 1 and 2, respectively. The powder sample was dispersed in alcohol and deposited onto the plate with a syringe; the sample adhered to the plate when the alcohol evaporated. An external thermocouple measured the temperature near the sample during heating. The SEM images were obtained with a lower detector and at an acceleration voltage of 5.0 kV. Several images were captured at 100× to observe the whole area of the sample. In situ images were recorded at a higher magnification, 800×, to investigate the sintering behavior in detail. The vacuum condition inside the sample chamber of the scanning electron microscope was in the range of 10−4 to 10−5 Pa.
Photograph of heating stage.
SEM image of powder sample pasted on stainless steel plate.
Two types of atomized powders, pure Al (A1070-30B, Toyo Aluminium K.K., median diameter: 34.6 µm) and Al–10%Si–0.4%Mg alloy (SI10MG-30B, Toyo Aluminium K.K., median diameter: 34.0 µm), were examined. The alloy composition is reported in mass percentage. The behavior of SLPS depends on the liquid fraction, which is derived from the alloy composition and temperature, in the sample during heating. The equilibrium liquid fraction was calculated using thermodynamic equilibrium calculation software (CaTCalc, Research Institute of Computational Thermodynamics, Inc.) to determine the heating procedures for observation. The temperature dependence of the liquid fraction is presented in Fig. 3. Pure Al undergoes a phase change from solid to liquid when the temperature reaches its melting point (660°C). By contrast, the liquid phase of Al–10%Si–0.4%Mg alloy begins to form when the temperature of the alloy exceeds its solidus temperature (567°C). The amount of liquid phase increases with increasing temperature. The whole alloy becomes liquid at temperatures above its liquidus temperature (593°C). Temperature profiles for observation were set according to the results of the thermodynamic calculations (Fig. 4). The pure Al powder was heated at a rate of 120°C/min to 600°C, where no liquid phase forms under equilibrium conditions. After the temperature reached 600°C, it was continuously increased and held for ∼1 min at every 10°C increment. The Al–10%Si–0.4%Mg alloy powder was heated at a rate of 180°C/min to 500°C, where no liquid phase forms under equilibrium conditions. Subsequently, the temperature was continuously increased and held for ∼1 min every 5–10°C. The temperature of both the heater and the sample were recorded during the observations. Hereinafter, the temperature refers to the sample temperature.
Temperature dependence of liquid mass fraction for pure Al and Al–10%Si–0.4%Mg determined by CatCalc calculation.
Temperature profiles for observation of (a) pure Al and (b) Al–10%Si–0.4%Mg.
Figure 5 shows in situ SEM images (magnification: 800×) of pure Al powder during heating. The temperature dependence of the liquid fraction in equilibrium states is presented in Fig. 5(d). The plots labeled a–c in Fig. 5(d) correspond to each liquid fraction and temperature at which the images in Fig. 5(a)–(c) were captured. As shown in Fig. 5(a), the particles before heating were spherical, with rough surfaces. Figure 5(b) shows that the particles apparently did not change shape until the specimen reached its melting point (660°C). The surface roughness was reduced above the melting point (Fig. 5(c)). Subsequently, the liquid ruptured the oxide layer and was distributed between the particles, which resulted in the formation of interparticle necks indicated by dashed lines in Fig. 5(c).
In situ SEM images of pure Al powder at 800×: (a) before heating, (b) at 659°C (liquid fraction: 0 mass%), (c) at 664°C (liquid fraction: 100 mass%). (d) Temperature dependence of liquid mass fraction showing points (a)–(c).
Figure 6 shows in situ SEM images of the Al–10%Si–0.4%Mg powder at 800× magnification. Figure 6(f) shows the relationship between temperature and the liquid mass fraction in the equilibrium states. The plots labeled a–e in Fig. 6(f) correspond to each liquid fraction and temperature when the images in Fig. 6(a)–(e) were captured. Figure 6(a) shows that the particle surface was uneven before heating. No significant change occurred until the temperature reached the solidus point (567°C), as shown in Fig. 6(b). Heating above the solidus temperature reduced the surface roughness of the particle and induced crack formation in the oxide layer, as marked with dashed lines in Fig. 6(c). In addition, necking connections formed among several particles, as indicated by solid lines in Fig. 6(c). When the specimen reached 594°C, where the liquid fraction was calculated to be 100 mass%, the liquid formed inside the particle marked with an arrow in Fig. 6(c) and absorbed neighboring semi-solid particles. As a result, the liquid was distributed nonuniformly, as indicated by dashed lines in Fig. 6(d). The sample remained almost unchanged when the temperature exceeded 594°C, as shown in Fig. 6(e). It should be noted that the particle indicated by an arrow in Fig. 6(e) moved rightward because of volumetric shrinkage resulting from the formation of the liquid phase. In addition, the particle was in the upper-left region of the field of view in Fig. 6(a)–(d) and then disappeared, as indicated by dashed lines in Fig. 6(e). The particle moved outside the visual field of the scanning electron microscope because of the volumetric shrinkage of the sample. The liquid ruptured the oxide layer and formed necks when the specimen reached the temperature at which the liquid phase forms, which is considered to be the same sintering behavior exhibited by pure Al.
In situ SEM images of Al–10%Si–0.4%Mg powder at 800×: (a) before heating, (b) at 564°C (liquid fraction: 0 mass%), (c) at 569°C (liquid fraction: 7 mass%), (d) at 574°C (liquid fraction: 49 mass%), (e) at 594°C (liquid fraction: 100 mass%). (f) Temperature dependence of liquid mass fraction showing points (a)–(e).
Figures 5 and 6 demonstrate that in situ SEM imaging with heating is an effective method to visualize the SLPS process of Al alloys.
Heating above the melting or solidus temperature reduces the surface roughness of pure Al or Al–10%Si–0.4%Mg particles and then causes cracks in the oxide films. This behavior appears to be due to the volumetric expansion of the particles during the solid-to-liquid phase change.
Figure 6 indicates that the liquid-phase sintering of Al–10%Si–0.4%Mg begins when its temperature exceeds its solidus temperature. Different from the results in Fig. 6, the liquid-phase sintering of Cu–10%Sn alloy has been reported to occur below the equilibrium temperature.8,16) Tandon et al. used a dilatometer to investigate the shrinkage response as a function of temperature for Cu–10%Sn alloy, revealing that the rapid shrinkage due to liquid-phase sintering occurs 25°C below the equilibrium temperature. They reported that a liquid might form at temperatures lower than indicated by the phase diagram because the atomization of particles is a nonequilibrium solidification process. Because the Al–10%Si–0.4%Mg alloy used for the observation is also an atomized powder, we expected a liquid phase to form below its solidus temperature. However, the temperature at which sintering progresses, as confirmed by SEM observation, coincides with the temperature at which the liquid phase is generated under equilibrium conditions. The discrepancy between the results of Tandon et al. and our SEM observations is attributed to an insufficient amount of liquid phase to rupture the oxide layer. The progression of sintering would require a certain amount of liquid, which leads to volumetric expansion and oxide disruption.
Figure 7 illustrates the models of SLPS of Cu–10%Sn8,16) and Al–10%Si–0.4%Mg powders based on the previous discussion. Cu–10%Sn particles have no oxide layers preventing sintering (Fig. 7(a)). The liquid spreads along grain boundaries when the powder is heated at temperatures below the solidus temperature and then locates among the particles, which leads to pendular necks in the initial stage of sintering. By contrast, Al–10%Si–0.4%Mg particles are covered by native oxides. The atomized powder has been reported to be covered by an oxide layer whose thickness varies from 50 to 150 Å.10) The Al powder used for the observation also has a surface oxide film because it was manufactured by an atomization method. Although the liquid occurs at temperatures below the solidus temperature, the volumetric expansion appears to be insufficient to rupture the surface oxides. The amount of liquid would increase at temperatures above the solidus temperature and induce sufficient volumetric expansion, rupturing the oxide layer and enhancing liquid-phase sintering.
Schematic illustrations of supersolidus liquid phase sintering of powder particles: (a) Cu–10%Sn and (b) Al–10%Si–0.4%Mg.
We here compare the sintering processes of pure Al and Al–10%Si–0.4%Mg on the basis of low-magnification observations. Figure 8 presents SEM images of the pure Al powder at 100×, showing the overall area of the sample. The dashed lines indicate the in situ observation area. Although the sample appeared unchanged at temperatures below the melting point, as shown in Fig. 8(a) and (b), most of the sample melted above this temperature, as shown by the arrow in Fig. 8(c). Figure 9 shows SEM images of the Al–10%Si–0.4%Mg powder at 100×, which is the same magnification as the images of pure Al. As shown in Fig. 9(b), the liquid began to distribute nonuniformly above the solidus temperature. At the temperature that gives 100 mass% liquid fractions, the melted area marked with arrows in the SEM image of Al–10%Si–0.4%Mg (Fig. 9(c)) is smaller than that in the image of pure Al (Fig. 8(c)). This difference is due to the difference in volumetric expansion between the two alloys. Specifically, pure Al undergoes a phase change from solid to liquid in a short period after its temperatures exceeds the melting point, whereas the liquid of Al–10%Si–0.4%Mg increases gradually with increasing temperature. The gradual increase of the liquid results in a low volumetric expansion per unit time, which is insufficient to disrupt the surface oxides. It should be noted that the high-magnification images in Figs. 5 and 6 show no significant difference in sintering behavior between the Al and the Al–10%Si–0.4%Mg alloy.
SEM images of pure Al powder at 100×: (a) before heating, (b) at 649°C (liquid fraction: 0 mass%), (c) at 664°C (liquid fraction: 100 mass%).
SEM images of Al–10%Si–0.4%Mg powder at 100×: (a) before heating, (b) at 574°C (liquid fraction: 49 mass%), (c) at 618°C (liquid fraction: 100 mass%).
Figures 8(c) and 9(c) indicate that the in situ observation area melted less than the other areas because long-term electron-beam irradiation contaminated the particle surface. The electron beam cleaves the C–H bonds of the hydrocarbon on the sample surface. The generated carbons polymerize on the surface, resulting in contamination.17) The contaminated layer on the particle surface prevents the liquid from rupturing the oxide, which causes the in situ observation area to melt less. Shortening the electron-beam irradiation time as well as cleaning the sample before observation would reduce the contamination level. In a future study, using such an improved procedure, we will investigate the sintering mechanism further by observing other Al alloys and powders with various oxide thicknesses.
We directly observed the SLPS of an Al alloy. Specifically, an in situ SEM technique with heating under vacuum conditions enabled us to investigate the sintering processes of pure Al and Al–10%Si–0.4%Mg alloy. The sintering behavior was discussed on the basis of the results and the liquid fraction calculated in equilibrium states. This methodology can be applied to further investigate the mechanism of liquid-phase sintering, potentially leading to improvements in the quality of sintered parts through precise control of the behavior of the liquid phase.
The authors thank Toyo Aluminium K.K. and Nikkeikin Aluminium Core Technology Company Ltd. for providing Al powder.
