2024 Volume 65 Issue 6 Pages 672-676
Additive manufacturing for open-cell aluminum foam is attracting attention because the cell morphology determined by the characteristics of the aluminum foam can be accurately controlled according to the application. However, there are few reports on the additive manufacturing of closed-cell aluminum foams. Wire arc additive manufacturing (WAAM) is a low-cost additive manufacturing method and has potential for manufacturing closed-cell aluminum foams from a wire precursor. In this study, WAAM was used to produce closed-cell aluminum foam. The precursor round bar was obtained by hot-pressing a mixture of pure aluminum powder, 3 mass% TiH2 powder, and 1 mass% Al2O3 powder. After the precursor round bar with a diameter of 15 mm was inserted into an A5052 aluminum alloy tube with an outer diameter of 19 mm and inner diameter of 15.5 mm, the assemblage was swaged to a precursor wire measuring ϕ1.6 × 100 mm. The WAAM equipment was based on the torch of a TIG arc welding machine that moved under numerical control. Foam formation tests were carried out with the precursor wire at a torch travel speed range from 120 mm/min to 620 mm/min. An additive foaming test was also carried out using a precursor wire measuring ϕ1.6 × 100 mm to investigate welding of the second layer to the first layer while the second layer was foaming. The density of the aluminum foam was measured by Archimedes’ method to calculate the porosity. The aluminum foam samples were cut perpendicular to the long axis, and cross-sections were observed with an optical microscope. For the single-layer aluminum foam formed with a TIG arc, the maximum porosity of 60% was obtained at a travel speed of 320 mm/min. A two-layer aluminum foam with a porosity of 18% and mean pore diameter of approximately 86 µm was obtained in the additive experiment. The second and first layers were bonded without forming a boundary. These results indicate that closed-cell aluminum foam was successfully produced using WAAM. However, the viscosity of the base aluminum of the precursor wire must be high to obtain a high-porosity aluminum foam with WAAM.
Aluminum foams are attractive structural and functional materials due to their light weight [1]. The characteristics of aluminum foams, such as their strength [2], thermal insulation [3], sound absorption [4], vibration resistance [5], and impact absorption [6], depend on the cell morphology. Many processes for manufacturing metal foams exist, such as the powder metallurgy and casting routs [7]. Recently, additive manufacturing processes have been attracting attention as methods to manufacture aluminum foams [8]. This is because the cell characteristics such as geometric structure and cell and edge sizes can be easily controlled, thereby facilitating the manufacturing of aluminum foams with appropriate cell structures for target applications. Although many reports on additive manufacturing of aluminum foam have addressed open-cell aluminum foams [9], few reports on additive manufacturing have addressed closed-cell aluminum foams [10]. If closed-cell aluminum foams can be formed with additive manufacturing by foaming a precursor, it is possible to manufacture closed-cell aluminum foams quickly and to use aluminum foam as a coating [11]. A 3D metal printer with wire arc additive manufacturing (WAAM) [12] can make products inexpensively with various types of 3D metal printer [13]. In this study, closed-cell aluminum foam was made with additive manufacturing. The precursor wire was manufactured by swaging a precursor bar obtained by powder metallurgy. The closed-cell aluminum foam was manufactured with the precursor wire as media by heating it and forming a foam with a TIG arc welding machine in which movement of the torch was numerically controlled.
Aluminum powder with a purity of 99% or higher and particle diameters of less than 180 µm (Kojundo Chemical Lab. Co., Ltd.) was prepared as the matrix material for aluminum foam. Titanium hydride (TiH2) powder with particle diameters of less than 45 µm and alumina (Al2O3) powder with particle diameters of about 1 µm were prepared as foaming and thickening agents, respectively (Kojundo Chemical Lab. Co., Ltd.). The aluminum powder, 3 mass% TiH2 powder, and 1 mass% Al2O3 powder were mixed and placed in a stainless steel mold. A lab-made machine was used to hot-press the mixed powder at 798 K and 150 MPa for 180 min to obtain an aluminum precursor with a diameter of 15 mm and length of 50 mm. The precursor was inserted into an A5052 aluminum alloy tube with outer diameter of 19 mm, inner diameter of 15.5 mm, and length of 50 mm. A5052 aluminum alloy has more strength and weldability than pure aluminum, which are advantageous for the additive foaming process and the strength of the precursor wire. A composite round bar with a diameter of 3 mm was obtained by swaging the aluminum precursor and A5052 tube with a swaging machine (USD-10000, Yoshida Kinen Co., Ltd.) at room temperature. The precursor wire with a diameter of 1.6 mm was obtained by swaging the composite bar again with a second swaging machine (USD-500, Yoshida Kinen Co., Ltd.) that can process a fine wire of less than 3 mm. The content ratio of the TiH2 in the precursor wire was about 2 mass% and the precursor wire contained enough TiH2 particles to produce foaming [14]. The diameter of 1.6 mm was necessary to load the wire into the wire feed device of the TIG welding machine. The precursor wire was cut and its cross-section was observed by scanning electron microscopy (SEM).
2.2 Foaming testA commercial small computer numerical control (CNC) router (Genmitsu 3018-PRO, Sain Smart Co., Ltd.) was modified to build a wire arc additive manufacturing machine. The main shaft was removed from the CNC machine and the torch of a DC TIG welding machine (TIG Mini 200PII inverter, DAIHEN Co., Ltd.) was installed in the CNC machine instead of the main shaft. The modified CNC machine was used as an automatic welding machine for WAAM with a wire feeding device.
The experimental setup of the foaming test is shown in Fig. 1. The precursor wire was cut into specimens 100 mm long. For processing, first a specimen was placed on an aluminum plate stage. The torch was set where the tungsten electrode with a diameter of 2.4 mm created an arc length of 2.5 mm above one end of specimen. The TIG arc was started at a current of 4 A and the torch was carried along the specimen end to end to foam it. Argon gas with a purity of 99.99% was used as the shield gas. The foaming tests were carried out using different travel speeds VTS from 120 mm/min to 620 mm/min, and the foaming behavior was observed with a high-speed microscope (MCHU30RS-CDC, Shodensha Co., Ltd.). The foaming tests were carried out three times at each travel speed. The density of the foamed specimens was measured by Archimedes’ method to calculate the porosity. The porosity p was calculated as
| \begin{equation} p = 1 - \frac{\rho_{\text{f}}}{\rho_{\text{Al}}} \end{equation} | (1) |
where ρAl and ρf are the densities of the pure aluminum and the foamed specimen, respectively. The foamed specimen was cut perpendicular to the long axis with rotating whetstones, the cross-section was observed with an optical microscope (ECLIPSE MA100, Nikon Co., Ltd.), and the pore diameter was evaluated as the equivalent circle diameter by image analysis with ImageJ software (Ver. 1.52v) [15]. The equivalent circle diameter dm was expressed as
| \begin{equation} d_{\text{m}} = \sqrt{\frac{4}{n\pi} \sum_{i = 1}^{n} A_{i}} \end{equation} | (2) |
where n is the number of pores and A is the pore cross-section area. The local porosity pIA was also calculated with image analysis as
| \begin{equation} p_{\text{IA}} = \frac{1}{A_{\text{f}}} \sum_{i = 1}^{n} A_{i} \end{equation} | (3) |
where Af is the cross-section area of the foamed specimen.
Schematic illustration of the experimental setup for the foaming test.
An additive foaming test was also carried out for manufacturing an aluminum foam by WAAM. The precursor wire was cut into short specimens 10 mm long. The first layer was foamed as shown in Fig. 1 with a travel speed of 220 mm/min and an arc current of 6 A. The density of the short foamed specimen was measured to calculate its porosity. Another short specimen was stacked on the top of the foamed specimen (first layer) and both ends of the short specimens were fixed with metal clips as shown in Fig. 2. The upper short specimen was heated with the TIG arc to foam and was added while moving the torch end to end with a travel speed of 220 mm/min and arc current of 6 A. The density of the two-layer foamed specimens was measured to calculate their porosity. The two-layer foamed specimens were cut into halves perpendicular to the long axis with a rotating whetstone, and the cross-sections were observed with the optical microscope.
Schematic illustration of the experimental setup for the additive foaming test.
A precursor wire with a diameter of 1.6 mm and length of about 3 m was obtained by powder metallurgy and swaging. The precursor wire cut into a 100 mm length is shown in Fig. 3(a). A metallic luster was observed on the surface of the precursor wire. A SEM micrograph of the cross-section of the precursor wire is shown in Fig. 3(b). The bonding boundary between the precursor and A5052 tube was clearly observed, but this boundary was not a disadvantage for foaming. No cracks were identified in either cross-section of the precursor and A5052 tube. The TiH2 particles were dispersed in the precursor. Al2O3 particles were not clearly observed due to their small diameter (Fig. 3(c)). The relative density of the precursor wire was over 99%.
(a) Photograph of the precursor wire; (b) and (c) SEM micrographs of the cross-section of the precursor wire.
The appearances of single-layer foamed specimens are shown in Fig. 4. The single-layer specimen formed with a travel speed of 120 mm/min had a relatively flat shape. This is because the specimen was foamed by the H2 gas generated from the TiH2 particles, after which the gas was released from the aluminum matrix into the atmosphere due to the long heating time at a travel speed 120 mm/min. In contrast, humping [16, 17] was observed in the foamed specimens formed with travel speeds from 220 mm/min to 620 mm/min. These results show that there are well-expanded and almost-unexpanded zones. This is due to an arc jump caused by expansion of the specimen surface toward the tungsten electrode tip as shown in Fig. 5. Generally, humping occurs when the travel speed of the torch is high compared to the heat input (welding current). In addition, the humping in this study was caused by interruption of foaming. Thus, humping can be prevented by increasing the viscosity of the base aluminum precursor, increasing the arc current, and decreasing the travel speed of the torch. The porosity of the single-layer foamed specimen is plotted against the travel speed in Fig. 6. The porosity increased as the travel speed increased to 320 mm/min and then decreased. The maximum porosity pmax was about 60% at 320 mm/min. The porosity of a single-layer foamed specimen can be increased by preventing the humping and drainage.
Appearance of the foamed specimens.
Photos obtained at (a) t = 20.748 s and (b) t = 20.909 s from a high-speed microscopy movie at VTS = 320 mm/min.
Porosity plotted against the travel speed of the welding torch.
Micrographs of cross-sections of the expanded region of the single-layer foamed specimen are shown in Fig. 7. Several pores were observed in all specimens. A dense zone was clearly observed in the specimens formed at travel speeds of 120 mm/min and 620 mm/min. The dense zone of the specimen formed at 120 mm/min was caused by drainage [18]. In contrast, the dense zone of the specimen formed at 620 mm/min was an unfoamed region caused by the short heating time. The local porosity pIA of the single-layer foamed specimen calculated by the image analysis is also shown in Fig. 7. The local porosity by image analysis is larger than that by the Archimedes’ method because the pIA value was obtained from the optical micrographs of cross-sections of the foamed zone. The maximum local porosity by image analysis was 90% obtained at 320 mm/min. This result indicates that aluminum foams with high porosity, over 80%, can be obtained by optimizing the composition and manufacturing conditions of the precursor wire and the heating conditions.
Optical micrographs of the foamed specimens.
The mean pore diameters of the single-layer foamed specimens are also shown in Fig. 7. The mean pore diameter almost generally corresponds to the porosity. The maximum mean pore diameter, 560 µm, was obtained at a travel speed of 320 mm/min. Generally, the mean pore diameter of the aluminum foam manufactured by the powder metallurgy and foaming are several millimeters [14, 15, 18–20]. The maximum mean pore diameter of single-layer foamed specimens in this study is smaller than that of the aluminum foam manufactured by conventional powder metallurgy and foaming. This result indicates that WAAM has the potential to produce foams with high porosity and fine pore diameter.
Additive tests were carried out to evaluate whether WAAM can produce multi-layer aluminum foams. The second layer was added after polishing a single layer with a porosity of 36%. A two-layer foamed specimen could be obtained with a travel speed of 220 mm/min and arc current of 6 A. The porosity of the two-layer foamed specimen was 18%, a decrease from that of a single-layer foam. The viscosity of the precursor wire and the heating conditions need to be optimized to obtain a high-porosity aluminum foam. The cross-section of the two-layer foamed specimen is shown in Fig. 8. The first and second layers were bonded, and no bonding boundary between the layers was observed, even though aluminum forms a thick oxide film in air. The mean pore diameters of the first layer, second layer, and both layers were 49 µm, 134 µm, and 86 µm, respectively. The pores of the first layer were collapsed because the viscosity of the specimen was low. The pores of the first layer were not combined with the pores of second layer and were not coarsened. The results show that additive foaming can be achieved and aluminum foams can be manufactured by WAAM. However, the viscosity of the base aluminum of precursor wire must be high to obtain aluminum foam products with high porosity with WAAM.
Optical micrographs of (a) the whole specimen and (b) around the bonding region of the laminated aluminum foam.
The efficiency of WAAM for fabricating the complex shape of aluminum foam products was investigated. The following conclusions were obtained:
This work was supported by the Iketani Science and Technology Foundation under Grant No. 0331048-A and the Japan Society for the Promotion of Science under KAKENHI Grant No. 20K0516300. This work was also supported by the Japan Science and Technology Agency Mirai Program under Grant No. JPMJMI19E5.
