- Email: [email protected]
Bond formation and fiber embedment during ultrasonic consolidation
Journal of Materials Processing Technology 209 (2009) 4915–4924
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Bond formation and fiber embedment during ultrasonic consolidation Y. Yang, G.D. Janaki Ram, B.E. Stucker ∗ Department of Mechanical and Aerospace Engineering, Utah State University, Logan, UT 84322-4130, USA
a r t i c l e
i n f o
Article history: Received 15 July 2008 Received in revised form 8 January 2009 Accepted 17 January 2009 Keywords: Additive manufacturing Ultrasonic consolidation Ultrasonic metal welding Interfacial microstructures Bonding mechanism Fiber embedment Metal matrix composites
a b s t r a c t The quality of ultrasonically consolidated parts critically depends on the bond quality between individual metal foils. This necessitates a detailed understanding of interface microstructures and ultrasonic bond formation mechanisms. In this work, the interface microstructures of a variety of ultrasonically consolidated similar and dissimilar metal samples were investigated. Samples with embedded SiC fibers were also investigated. Based on detailed microstructural studies, the mechanisms of foil bonding and fiber embedment in ultrasonic consolidation have been discussed. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Ultrasonic consolidation (UC) is a novel additive manufacturing process implementing ultrasonic metal welding for fabrication of complex three-dimensional (3D) structures from metal foils. The Solidica FormationTM UC machine (commercially introduced by Solidica Inc., USA, in 2000) incorporates an ultrasonic metal welding head, a 3-axis CNC milling head, a foil feeding apparatus, and a software program to automatically generate tool paths for material deposition and machining. In the UC process, a CAD model of the component to be fabricated is initially generated. The model is converted to an STL file format and systematically sliced into a number of horizontal layers using a customized computer program. Each horizontal layer corresponds to the thickness of the metal foil used for part fabrication. Fig. 1 illustrates the basic UC process. A rotating ultrasonic welding head, a.k.a. sonotrode, travels along the length of a thin metal foil placed on a base plate. The foil is held closely in contact with the base plate by applying a normal force via the rotating sonotrode. The sonotrode vibrates transversely to the direction of travel at a frequency of 20 kHz and at a user-set oscillation amplitude, while traveling over the metal foil to bond it to the base plate. After depositing a strip of foil, another foil could be deposited adjacent to it depending on the geometry of the part. This operation repeats until a complete layer is placed. After placing a layer, a CNC milling head subtractively shapes the layer to its
∗ Corresponding author. Tel.: +1 435 797 8173; fax: +1 435 797 2417. E-mail address: [email protected] (B.E. Stucker). 0924-0136/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2009.01.014
slice contour. This milling can occur after each layer or after several deposited layers. After shaping a layer, machining chips are blown away using compressed air and foil deposition starts for the next layer. These additive and subtractive operations repeat until the component is finished. Research on the practical applications of UC has been extensive. Previous work has demonstrated the unique capabilities of UC for fabrication of multi-functional 3D structures with high dimensional accuracy and desirable surface finish. George (2006) has demonstrated honeycomb panels with complex internal features, Janaki Ram et al. (2007a) have made multiple material structures, and Siggard (2007) has fabricated objects with integrated wiring and electronics. Another interesting application of UC is embedding fibers within metal matrices for producing metal matrix composites (MMCs) and for other purposes. A variety of fibers have been successfully embedded in aluminum matrices. Yang et al. (2007) have successfully embedded silicon carbide (SiC) fibers within Al 3003, and Kong (2005) reported similar successes on shape memory alloy and optical fibers using UC. Although the application potential of UC has been widely investigated, the mechanism of bond formation between metal foils during UC has not been clearly understood. Since the additive operation during UC is essentially a seam ultrasonic metal welding (UMW) application, an improved understanding of bond formation in UC can be obtained from the available information on bond formation in UMW. As discussed by Jones and Powers (1956) it is generally agreed that ultrasonic metal welding is a solid-state joining process. Apart from pure solid-state bonding, which is caused
4916
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924 Table 2 Process parameters used for SiC fiber embedment. Parameter
Value Al 3003/SiC/Al 3003
Vibrational amplitude (m) Normal force (kPa) Welding speed (mm/s)
Al 3003/SiC/Cu
10 275 28
12 275 28
the machine, tapes were manually placed onto the base plate and secured with an adhesive at the ends, followed by welding. 2.2. Ultrasonic consolidation of dissimilar metal foils
Fig. 1. Schematic of ultrasonic consolidation process (not to scale).
by atomic-level forces across nascent metal contact points, bond formation during UMW has been suggested by Joshi (1971) to be due to: (i) mechanical interlocking, (ii) interfacial melting, and (iii) metal diffusion. While all these mechanisms can contribute to bond formation, the operating or the dominant mechanism is expected to change with the materials in question and with the processing conditions. Therefore, it is important to know which mechanisms dominate under standard UC processing conditions. The current lack of consensus on foil bonding mechanisms in UC limits further exploitation of its processing capabilities. In view of the above, an attempt is made in this work to investigate bond formation during UC using a variety of ultrasonically consolidated similar and dissimilar metal samples. Further, an attempt is made to understand the mechanism of ceramic fiber embedment during UC. 2. Experimental details 2.1. Ultrasonic consolidation of Al 3003 and Ni 201 Metal tapes used in the current study are listed in Table 1. No special foil cleaning procedures were adopted to facilitate bonding. UC experiments were conducted using a Solidica FormationTM machine with a sonotrode diameter of 147 mm. Metal foils were deposited onto an Al 3003 base plate firmly bolted to the heat plate of the UC machine. The procedure for depositing Al 3003 tapes involved first depositing a layer of Al 3003 on the aluminum base plate. Subsequently, another layer of Al 3003 was deposited on the top of the previously deposited Al 3003 layer. The Al 3003 tapes were automatically fed through the tape feeding mechanism of the UC machine. The processing parameters used for Al 3003 foil consolidation were: oscillation amplitude, 16 m; welding speed, 28 mm/s; normal force, 1750 N; and base plate temperature, 149 ◦ C. These parameters were found by Janaki Ram et al. (2007b) to produce a high level of linear welding density for Al 3003. A similar procedure and processing parameters were implemented for depositing Ni 201 metal tapes. Since Ni 201 tapes were not available in standard coils suitable for automatic feeding by
Dual-material samples were fabricated from Al 3003 and Ni 201 tapes. Here again, deposition experiments were conducted using the Solidica FormationTM machine on an Al 3003 base plate. After depositing a few layers of Al 3003 one over another, a layer of Ni 201 was welded to the top most Al 3003 layer. Subsequently, another layer of Ni 201 was welded to the previously deposited Ni 201 layer. This layer arrangement was chosen to facilitate microstructural examination of Ni–Al and Ni–Ni interfaces. As before, Ni tapes were manually placed onto the substrate and secured with adhesive while Al 3003 layers were automatically fed. The process parameters were identical to those used for ultrasonic consolidation of Al 3003. 2.3. SiC fiber embedment The fiber embedment experiments were conducted using an ultrasonic seam welder at Loughborough University, UK, with a sonotrode diameter of 50 mm. This machine is as capable as the commercial UC machine in terms of joining thin metal foils. Manually fed 100 m thick foils were used and no base plate preheating was employed in all the experiments conducted using this machine. Silicon carbide fibers of 100 m diameter were embedded between Al 3003 foils along their length, following the embedding procedure described by Yang et al. (2007), who successfully embedded SiC fibers within Al 3003 matrices using this procedure. In order to understand the effects of foil material properties on fiber embedment, SiC fibers were embedded between Al 3003 foil and high-purity Cu (99.9%) foil of 100 m thickness. The process parameters used for these experiments are listed in Table 2. 2.4. Microstructural characterization Ultrasonically consolidated similar and dissimilar metal samples as well as fiber embedded samples were metallographically investigated. Transverse sections (perpendicular to the foil length direction) cut from the various samples were prepared for microstructural examination following standard metallographic practices. Some of the Al 3003 samples and fiber embedded samples were etched with Keller’s solution (HF-1%, HCl-1.5%, HNO3 -2.5%, and H2 O-95%). Ni 201 samples were etched with a mixture of 1 part 10% aqueous solution of CaCN and 1 part 10% aqueous solution of (NH4 )2 S2 O8 . Metal/metal and fiber/metal interface microstructures were examined using a scanning electron microscope (SEM). X-ray energy dispersive spectroscopy (EDS) was utilized for micro-
Table 1 Materials used for UC experiments. Material Al alloy 3003 (H18 condition) Ni alloy 201 (annealed condition)
Nominal composition (wt.%)
Dimensions
Al–1.2Mn–0.12Cu Ni–0.02C–0.35Mn–0.25Si–0.25Fe–0.15Cu
25 mm wide, 150 m thick foil 25 mm wide, 75 m thick foil
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924 Table 3 Process parameter used for temperature measurement experiments. Combination
Oscillation amplitude (m)
Normal force (kPa)
1 2 3 4
10 10 12 12
206 275 206 275
chemical characterization. Orientation imaging microscopy (OIM) was utilized for studying plastic deformation at the interfaces. 2.5. Temperature measurements Temperature measurements at weld interfaces were conducted using the ultrasonic seam welder at Loughborough University. Because of its constructional simplicity, this machine was considered more suitable for conducting these experiments. Measurements were made at various combinations of oscillation amplitude and normal force, listed in Table 3. The temperature measurement procedure consisted of the following: (i) depositing a metal tape onto a base plate (maintained at room temperature), (ii) placing a small K-type thermocouple (50 m diameter) on the top surface of the deposited tape, and (iii) depositing a second metal tape on the top of the previously deposited tape, with the small thermocouple sandwiched in
Fig. 2. Unbonded regions in ultrasonically consolidated Al 3003: (a) lowmagnification secondary electron image; (b) high-magnification back-scattered electron image (note the thin layer with a different contrast all around the defect (shown by arrow)).
4917
between. During the depositing of the second layer, the interface temperature was measured at a sampling rate of 1000 samples/s. 3. Results 3.1. Microstructures Fig. 2 shows the microstructure of an ultrasonically consolidated Al 3003 sample. The dark regions seen along the layer interfaces are the unbonded regions (defects). Examination of several unbonded regions using back-scattered electron imaging at higher magnifications indicated that these defects were covered with a thin differently contrasted layer (Fig. 2b). EDS spot analysis showed significantly higher oxygen content in this layer (Fig. 3a) than in the regions adjacent to it (Fig. 3b). Fig. 4 shows the microstructures of Al 3003/Ni 201 dissimilar metal samples. Generally, Ni 201 seemed to bond well to itself and to Al 3003 (Fig. 4a and b). As can be seen, the thickness of Ni foils remained practically unchanged. As shown in Fig. 4c, there were a few unbonded regions along the Ni/Ni interface. There was no evidence of metal melting or recrystallization at the weld interface. The weld interface appeared flat and mechanical interlocking did not seem to take place. Fig. 5a shows an image of the Al 3003/Ni 201 interface at a higher magnification, indicating no obvious evidence of intermetallic formation. The results of EDS line scans (for two elements, Al and Ni) performed across the Ni–Al interface (along the scan line shown in Fig. 5a (100 points, 1 m spot spacing, from Ni to Al side)) are shown in Fig. 5b. As can be seen, compositions seem to change sharply across the interface. While some amount of diffusion on a very fine scale cannot be ruled out, the results suggest that there was no bulk diffusion of Ni into Al and vice versa. Fig. 6 shows an inverse pole figure of a well-bonded Ni/Ni interface (generated from several OIM scans of contiguous areas), which
Fig. 3. (a) EDS spectra showing a distinct oxygen peak in the thin layer with a different contrast. (b) The oxygen peak is absent in the regions adjacent to, but outside the thin layer.
4918
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
Fig. 5. (a) Ni 201/Al 3003 interface, (b) EDS line scan results across the Ni 201/Al 3003 interface (along the scan line shown in (a), scan started on the Ni side).
Fig. 6. An image of several inverse pole figures of contiguous areas along a wellbonded Ni–Ni interface stitched together. The grains in the image are color coded to reflect their orientation. The line across the center of the image defines the Ni–Ni weld interface. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)
Fig. 4. SEM micrographs of the Al 3003/Ni 201 dual-material deposit: (a) shows a well-bonded Ni/Ni region, (b) shows a well-bonded Al 3003/Ni 201 region, and (c) shows a few Ni/Ni unbonded regions.
is color coded to indicate the crystallographic orientation ({h k l} direction parallel to the section normal) of each grain within the sample. Grains that have been plastically deformed typically show a smooth intra-grain color transition. Adams et al. (2008) suggested that the intra-grain color transition indicated rotations of the crystal lattice. Such smooth color transitions are evident in the picture, indicating that the foil interfaces plastically deform during the bonding process. However, there was no evidence of recrystallization at the weld interface. Interestingly, OIM examination of the unbonded regions along the Ni/Ni interface revealed a thin layer of extremely fine grains all around the unbonded region (Fig. 7),
Fig. 7. Inverse pole figure of an unconsolidated portion of the Ni–Ni interface. Note the extremely fine grains that are present along the defect boundaries.
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
Fig. 8. SiC fiber embedded between Al 3003 foils.
indicating that the crystals have been dynamically deformed into nano-sized grains during processing or that the foil is covered with an oxide layer and/or some kind of contamination at these locations. Such fine grains were not observed along the well-bonded portions of the interface (Fig. 6). Fig. 8 shows a SiC fiber embedded between Al 3003 foils. The fiber was not deformed during the process, retaining its circular external geometry. The center of the fiber was found to be located at the interface between top and bottom Al 3003 foils. Metal foils in the vicinity of the fiber were found to bond well, in particular. The cavity created by the placement of a SiC fiber between two foils was fully filled by the matrix metal, indicating significant plastic deformation of the matrix material during UC processing. Elemental mapping studies across the fiber/matrix interface showed no evidence of interdiffusion (Fig. 9). Fig. 10 shows a SiC fiber embedded between Al 3003 and Cu foils. In contrast to fiber embedment between two Al 3003 foils (Fig. 8), the center of the SiC fiber was found to be displaced into the bottom Al foil. Because the bottom aluminum foil is softer than the top copper foil, the SiC fiber is pressed into the Al foil during UC. It was also observed that the cavity created by the SiC fiber was filled only by Al 3003, and the geometry of the bottom surface of the copper foil indicated indentation by the SiC. An EDS line scan was conducted along the line shown in Fig. 11a (0.5 m point spacing, two elements (Al and Cu), started on the Cu side). As seen in Fig. 11b, there appears to be some amount of diffusion across the interface. However, the width of the diffusion zone is very small (less than 2 m)—too small to for conclusions with any degree of confidence, considering the resolution limit of the EDS technique. More importantly, however, our results show that there is no bulk diffusion of elements across the interface.
4919
Fig. 10. SiC fiber embedded between Al 3003/Cu foils. Note the aluminum flow around the fiber into the Cu side (shown by arrows).
3.2. Temperature measurements The results of temperature measurements at the weld interface are shown in Fig. 12. The peak temperature at the weld interface was found to vary from 68 to 98 ◦ C as a function of process parameters. 4. Discussion 4.1. Bond formation during UC 4.1.1. Mechanical interlocking Bonding due to mechanical interlocking can occur during ultrasonic metal welding. In a study of ultrasonic welding between Al and Au, Joshi (1971) observed a liquid-like metal flow of Au into the Al side at the weld interface and concluded that mechanical interlocking was responsible for bonding between Al and Au. However, in the present study, ultrasonically consolidated similar (Al/Al and Ni/Ni) and dissimilar (Al/Ni and Al/Cu) metal samples showed flat interfaces, suggesting absence of mechanical interlocking (Figs. 4 and 10). While mechanical interlocking during ultrasonic welding of similar metal foils appears to be very unlikely, it can occur in dissimilar welds, especially when the materials in question have a large difference in hardness (as in the case of Al and Au). 4.1.2. Metal melting Melting phenomena are related to the temperature rise at the metal/metal welding interface. During ultrasonic metal welding, heat is generated primarily due to dynamic frictional effects at the interface. Various methods have been applied to identify the actual temperature at the weld interface during UMW.
Fig. 9. Element mapping results of Al 3003/SiC/Al 3003 sample.
4920
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
Fig. 11. (a) Al 3003/Cu interface, (b) EDS line scan results across the Al 3003/Cu interface (along the scan line shown in (a), scan started on the Cu side).
These methods include: (1) embedding small thermocouples at the interface, (2) measuring the thermoelectric electromotive forces (e.m.f.) between workpieces during welding, and (3) observing the interface during welding using an infrared camera. Joshi (1971) measured the interface temperature during ultrasonic welding of a thin gold wire to a gold substrate using the thermocouple method, and the temperature was found to be no more than 120 ◦ C. Jones and Powers (1956) measured an interface temperature of around 204 ◦ C (400 ◦ F) during ultrasonic welding between pure aluminum sheets using the thermocouple method and 538 ◦ C (1000 ◦ F) during ultrasonic welding between Monel and aluminum sheets using
Fig. 12. Results of temperature measurements at the weld interface.
the e.m.f. method. Using the e.m.f. method, Weare et al. (1960) reported an interface temperature of 232 ◦ C (450 ◦ F) during ultrasonic welding between copper and Monel. Using an infrared camera with an accuracy of ±10 ◦ C, De Vries (2004) measured a temperature of 314 ◦ C (597 ◦ F) during ultrasonic metal welding of Al 6061. In all these cases, no substrate preheating was employed. While there is large variation in reported interface temperatures in ultrasonic metal welding, the measured temperatures are, in most cases, less than the melting point of the welded metal. A drawback of all the methods listed above is that the measured temperatures were all volume averaged temperatures of a certain volume of material at the metal/metal interface. It is possible that at some localized spots the temperature may reach or exceed the melting point of the metal, even if the averaged temperature at the weld interface is below the melting point. Consequently, microscopic examination was employed by several researchers to investigate the possibility of metal melting. In the investigations by Jones and Powers (1956), no metal melting was observed in ultrasonic welding between aluminum sheets. Daniels (1965) successfully achieved ultrasonic welds between copper and aluminum sheets without presence of metal melting. Joshi (1971) reported that no cast microstructures were observed during ultrasonic welding of Al, Cu, and Au. However, Kreye (1977) found some evidence of metal melting during ultrasonic welding of Cu2 Co. In their studies on ultrasonic welding between Al 1100 H19 and pure zinc sheets, Gunduz et al. (2005) reported localized melting of Al–Zn and solid solution formed at the welding interface. A preheat of 300 ◦ C was employed in this study. Thus, while most reports confirm that ultrasonic welding is a solidstate joining process, there is at least some evidence that localized melting can occur during ultrasonic welding depending on material combinations and processing conditions. In the present study, no evidence of melting was observed along the foil interfaces and the measured peak processing temperatures were significantly below the melting point of the metals. 4.1.3. Diffusion Ultrasonic welding results in only a modest increase in temperature at the weld interface. Moreover, the times available for diffusion are extremely short (owing to the short residence time of the sonotrode). Thus, from this standpoint, significant diffusion appears unlikely. However, there are indications that diffusivities can increase significantly under the conditions of ultrasonic welding. Ultrasonic welding can produce plastic deformation at very high strain rates, as reported by Gunduz et al. (2005). Consequently, dislocation densities and vacancy concentrations in the bond region can be significantly higher, which enhance diffusion. For example, Gunduz et al. (2005), in their studies on ultrasonic welding between Al and Zn, found that a strain rate up to 103 s−1 can be produced during welding, which increased the vacancy concentration to around 10−1 . As a consequence, they observed significant Zn diffusion into Al, nearly five orders of magnitude higher than the calculated diffusivity of Zn under normal conditions at the measured interface temperature. Therefore, it appears that ultrasonic welding can provide conditions for significant diffusion. Diffusion is further facilitated when high preheat temperature are employed. In the present study, bulk diffusion of elements across the interface was not observed (Fig. 5b for Al 3003/Ni 201 and Fig. 11b for Al 3003/Cu). Therefore, bond formation during UC does not seem to depend heavily on diffusion. However, there were indications of a very narrow diffusion zone (less than 2 m) at the weld interface (especially, in the case of Al 3003/Cu). While such a small-scale diffusion cannot be held completely responsible for bonding during UC, it can positively contribute to bonding. Diffusion allows for mass transfer across the interface. Further, any chemical interactions between component metals consequent to diffusion can
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
4921
lead to a stronger interface. However, it should be noted that diffusion is not a beneficial phenomenon in all cases, especially when dealing with metallurgically incompatible dissimilar material combinations. Diffusion often leads to chemical reactions, which can be detrimental to the bonding process and part mechanical properties. For example, formation of brittle intermetallics or low-melting eutectics is not generally desirable. 4.1.4. Atomic forces across nascent metal contact points While mechanical interlocking, interfacial melting, and diffusion all can occur during ultrasonic metal welding, they do not seem to have universal presence; rather, they seem to be specific to certain material combinations or processing conditions. Bonding in ultrasonic consolidation thus appears primarily to be solid-state, caused by atomic-level forces across the nascent metal contact points. As in the case of other solid-state welding processes, two conditions must be fulfilled for bond formation during ultrasonic welding: (i) generation of atomically clean surfaces, and (ii) intimate contact between clean metal surfaces. Surface oxide layers exist on all metallic materials, which prevent intimate contact between nascent metal surfaces. Hence removal of surface oxide layers is critical for achieving metallurgical bonding. During ultrasonic metal welding, as discussed by Daniels (1965), it is generally believed that frictional effects at the mating metal surfaces due to ultrasonic vibrations result in break-up and removal of surface oxide layers. In the current study, microstructural studies confirmed the presence of oxide layers along the defect boundaries (Fig. 2b). Oxide layers were, however, absent in the fully bonded or consolidated regions (Figs. 2a and 6). This confirms that ultrasonic vibrations (and consequent frictional effects) at the weld interface help remove surface oxide layers. However, this occurs only at the locations wherever there is surface contact. If the mating surfaces are not in contact, there cannot be any friction to break-up the surface oxide layers. Thus, the presence of oxide layers along the defect boundaries shown in Fig. 2b indicates that the mating surfaces across these defects had not come into contact. These noncontact regions with unremoved surface oxide layers show up as unbonded regions along the foil interfaces in the final deposit. It should be noted that unbonded regions along the foil interfaces can also be caused by cracking-related effects subsequent to bonding, especially under conditions of excessive energy input. Kong (2005) noticed some bond degradation due to excess energy input during ultrasonic consolidation of Al 3003. However, such unbonded regions would have nanometer-sized oxide layers, as the original oxide layers are removed in the process of bonding prior to cracking. Therefore, it can be deduced that the unbonded regions observed in the current study along the foil interfaces are not due to some cracking-related phenomena, but are due to a lack of complete surface contact between the mating foil surfaces as a result of surface roughness prior to bonding. The second necessary condition for solid-state welding is intimate nascent metal contact. Atomic-level forces can come into play only when the surfaces are brought close enough together. An impediment to this is surface roughness, which precludes complete contact between mating surfaces. In solid-state welding, this is normally overcome by plastic deformation at weld interfaces. The occurrence of plastic deformation was confirmed in the OIM studies (Fig. 6). Interfacial plastic deformation plays an important role in UC. First, it facilitates removal of the surface oxides. During ultrasonic welding, simultaneous application of normal and oscillating shear forces results in generation of dynamic interfacial stresses between the two mating surfaces at the contact points. As a consequence, cracks are initiated in the surface oxide layer. Kong et al. (2004) discussed how these interfacial stresses also induce plastic deformation in a thin layer of metal (∼20 m) just beneath the oxide
Fig. 13. Schematic of the mating surfaces at the beginning of ultrasonic consolidation.
layer. The nascent metal from beneath then extrudes through the cracks, resulting in the break-up of the oxide layers. These broken oxides are removed from the bond region by metal flow and are dispersed in the vicinity of the weld zone. Enjio (1986) identified such broken oxide fragments (0.05–0.2 m sizes) in an Al alloy diffusion weld subjected to ultrasonic vibrations. However, as noted by Kong et al. (2003) the dispersed oxide pieces may not be noticeable in all cases, as oxide layers are often only a few nanometers thick. Secondly, plastic deformation is crucial for producing a weld with satisfactory linear weld density (LWD), which is the length of well-bonded interface length divided by the total interface length, expressed in percentage. Roughness on the metal surfaces precludes 100% surface contact and, consequently, 100% bonding along the interface. Metal foils bond only at the initial surface contact points if plastic deformation does not occur. However, several researchers, such as Janaki Ram et al. (2007b), have reported nearly 100% LWD, which indirectly confirm the occurrence of metal plastic deformation. The situation at the mating surfaces at the beginning of ultrasonic welding can be visualized as shown in Fig. 13. As can be seen, contact between mating surfaces occurs only at surface asperities, leaving numerous no-contact regions along the interface. Bonding across these no-contact regions will not occur unless there is a mechanism to close these voids and to bring the mating surfaces into intimate contact. This is where plastic flow is believed to play a major role. Initially, bonds are established at the existing surface contact points. As the process progresses, these bonded regions grow in size, aided by plastic deformation. Plastic deformation at the bonded regions results in squeezing of metal into the voids and the mating surfaces across the void regions approach. As this happens, new points come into contact, leading to friction, oxide layer removal and bonding. These newly bonded regions also grow during subsequent deformation cycles, generating more contact points. This process can result in sound metallurgical bonding with relatively high linear weld density levels. The bonding process in ultrasonic welding can thus be recognized as repeated occurrence of two distinct stages, contact and bond stages, with each leading to the other. During the contact stage, new contact points are generated by plastic deformation of existing bonds. During the bond stage, oxide layers are removed and metallurgical bonds are formed at the contact points generated in the contact stage. Plastic deformation in ultrasonic metal welding significantly differs from that in several other well-known solid-state welding processes, such as roll bonding and friction welding. The latter involve bulk plastic deformation. In cold roll bonding, for example, typically a large rolling load is applied to produce solid-state bonds between materials. Materials during roll bonding undergo bulk plastic deformation and, consequently, significant reduction in section thickness. In a recently publication, Li et al. (2008) reported that a reduction in section thickness up to 60–70% is required for achieving satisfactory bonding with cold roll bonding. In the UC process, on the other hand, the forces applied are very low and plastic deformation is more or less confined to the interface. As can be seen in Fig. 4, the thickness of Ni 201 foils remained practically unchanged after consolidation.
4922
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
Interfacial plastic deformation during ultrasonic welding is expected to be facilitated by the “Blaha effect” or “acoustic softening” effect, as described by Langenecker (1966). Langenecker found that, in the presence of ultrasonic energy, metallic materials are known to experience significant softening, which is not connected to any rise in temperature resulting from being subjected to an ultrasonic field. In fact, ultrasonic energy was shown to be more effective than thermal energy in reducing the flow stress of metallic materials. For example, Eaves et al. (1975) reported that bulk heating can reduce stresses by 45% while ultrasonic vibration reduces it by 75%. In terms of energy density, Langenecker (1966) reported that it takes approximately 1022 eV/cm3 of thermal energy density to produce a zero stress in aluminum without ultrasonic superimposition, while only about 1015 eV/cm3 using ultrasonic energy. He explained this difference as “acoustic energy is assumed to be absorbed only at those regions in the metal lattice which are known to carry out the mechanisms of plastic deformation. Heat, on the other hand, is distributed rather homogeneously among all the atoms of the crystal including those which do not participate in the mechanisms of plastic deformation.” In addition to acoustic softening, thermal softening can also occur at the weld interface due to frictional heating, further contributing to a reduction in flow stress. Thus, although the forces involved in typical ultrasonic welding are generally modest, there seem to be conditions enough for causing plastic deformation and metal flow, when considering the combined acoustic and thermal softening effects. Plastic deformation and temperature rise at the interface can provide the necessary driving force for recrystallization during ultrasonic welding, although whether there is adequate time for recrystallization or not is still unknown. Recrystallization is desirable as it brings in necessary readjustments to the grain structure at the interface and replaces the strained grain structure on both sides of the interface with a set of freshly formed strainfree, fine grains. It facilitates grain continuity across the interface and a smooth transition of weld zone microstructure from the rest of the material. However, no conclusive evidence exists on whether recrystallization takes place during ultrasonic welding. Kenik and Jahn (2003) reported recrystallized grain structures in ultrasonic joint of Al alloys, and Allameh et al. (2005) claimed that recrystallization of grains does not occur due to insufficient thermal energy generated during relatively short welding duration although satisfactory bonding was demonstrated in all cases. In the current work, no recrystallization was observed in Ni 201 samples. It appears that specific processing conditions and the material systems in question govern the recrystallization process. 4.1.5. Effects of foil surface condition Foil surface condition is a critical factor in ultrasonic consolidation. Thick adherent surface oxide scales and large-scale surface roughness are undesirable, as surface oxide layers prevent intimate nascent metal contact. The ease with which oxide layers can be removed during ultrasonic welding depends on the ratio of metal oxide hardness to nascent metal hardness—higher ratios facilitate easier removal. This is the reason why Al alloys, with a very high oxide-to-metal hardness ratio, are one of the best-suited materials for ultrasonic welding. Noble metals such as gold, which do not have surface oxide layers, are quite amenable for ultrasonic welding as well. Materials that present difficulties with oxide layer removal have been reported to be problematic for ultrasonically welding. For example, Kong et al. (2003) found that Al–Mg–Si alloy 6061 was difficult to ultrasonically consolidate, which was attributed to difficulties with oxide layer removal, thought to be due to the presence of MgO in surface oxide layers of these alloys. Satisfactory bonding between Al 6061 foils was achieved by acid etching the
foil surfaces just prior to welding for removing the surface oxide layers. Due to surface roughness, when two metal surfaces are put together, 100% surface contact is not possible and the no-contact areas are inevitable. Generally, metal surfaces with a low-level of surface roughness, which produce fewer and smaller no-contact areas, are desirable for bond formation during ultrasonic consolidation. When the consolidation parameters are right, these no-contact areas are eliminated due to adequate plastic deformation, leading to nearly 100% linear weld densities. The effect of surface roughness on bond formation was investigated by Janaki Ram et al. (2007b). Examining the surfaces of ultrasonically consolidated foils, the authors noticed significant roughening of foil top surfaces as a result of sonotrode motion. This roughness was shown to be the major source of defects for unbonded regions in ultrasonically consolidated parts. Further, a simple technique, called surface machining, was proposed by the authors, which involved removal of a thin layer (∼30 m) of metal from consolidated foil surfaces prior to subsequent layer deposition. With this technique, nearly 100% linear weld densities were achieved even under processing conditions that would have otherwise resulted in a large number of unbonded regions along foil interfaces. It was claimed that, besides smoothening the substrate surface, surface machining mechanically removes the substrate surface oxide layers, which facilitates bond formation. 4.2. Fiber embedment during UC Microstructural examination produced no evidence of significant diffusion or chemical reactions between fibers and matrix materials during fiber embedment using UC. Therefore, it appears that fibers are mechanically trapped between the top and bottom foils during UC, with fiber embedment made possible by plastic deformation of matrix materials. When a fiber is placed between two metal foils, a large gap is created between the metal foils in the vicinity of the fiber. During consolidation, the fiber acts as a stress riser leading to larger stresses and more extensive matrix material plastic deformation in the vicinity of the fiber than at regions away from the fiber. Consequently, metal flow occurs into the gap around the fiber resulting in sound fiber embedment. Similar phenomena were noticed by Janaki Ram et al. (2007a) during stainless steel mesh embedment between Al 3003 foils using UC. If the fiber is sandwiched between metal foils of similar hardness, both metal foils deform to the same extent, resulting in a fiber being embedded with its center aligned with the metal/metal interface (as noticed in the present study in the case of Al 3003/SiC/Al 3003, Fig. 8). When dissimilar metal foils with a large difference in hardness are used, the softer metal tends to deforms more than the harder metal, resulting in the fiber being embedded more into the softer metal side (as noticed in the current study in the case of Al 3003/SiC/Cu, Fig. 10). To further confirm the suggestion of mechanical entrapment of fibers, tensile fractured Al 3003 specimens containing multiple SiC fibers were examined under SEM. Studies revealed a few instances of fiber pull-out (Fig. 14a). Examination of a pulled-out fiber at higher magnifications showed the fiber surface morphology (Fig. 14b and c). The surfaces of the holes created in the matrix by fiber pull-out showed a precisely matching surface morphology (Fig. 14d), indicating that the matrix tightly gripped the fiber and, as a consequence, took the negative image of the fiber surface. Clearly, such features would not show up if there had been any chemical bonding between the fiber and matrix. These observations thus provide strong evidence that fibers are mechanically entrapped in the matrix material during UC.
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
4923
Fig. 14. (a) Tensile fractured Al 3003 specimen containing SiC fibers. (b and c) Surface morphology of the pulled-out fiber. (d) Surface morphology of the “matrix hole” created by fiber pull-out.
5. Conclusions
Acknowledgements
The interface microstructures of various ultrasonically consolidated similar (Al 3003/Al 3003 and Ni 201/Ni 201) and dissimilar (Al 3003/Ni 201, Al 3003/Cu) metal samples were reported. None of the samples showed evidence of mechanical interlocking, localized metal melting, significant diffusion, or recrystallization at the weld interface. Bond formation between metal foils during ultrasonic consolidation under standard processing conditions appears to be essentially solid-state, which is caused by atomiclevel forces across the nascent metal contact points. Two crucial conditions for bond formation during ultrasonic consolidation are: (i) removal of surface oxide layers, and (ii) intimate contact between clean metal surfaces. Frictional effects at the weld interface ensure removal of surface oxide layers. Intimate nascent metal contact is facilitated by interfacial plastic deformation. The process of ultrasonic consolidation seems to involve repeated occurrence of contact and bond stages, with each leading to the other. During the contact stage, new contact points are generated by plastic deformation of existing bonds. During the bond stage, oxide layers are removed and metallurgical bonds are formed at the contact points generated in the preceding contact stage. Al 3003/SiC/Al 3003 and Al 3003/SiC/Cu fiber embedded samples showed no evidence of diffusion or chemical reactions between fiber and matrix materials. Fibers are mechanically entrapped in the matrix. Matrix material plastic deformation facilitates sound fiber embedment during ultrasonic consolidation.
The authors thank the National Science Foundation (DMI 0522908) and the Office of Naval Research (under Grant no. N000140710633) for supporting this work. The National Science Foundation provided travel support (International Research & Education in Engineering) facilitating the experiments at Loughborough University in the UK. The authors would like to thank Dr. Rupert Soar and Dr. Dezhi Li at Loughborough University for their constructive thoughts and suggestions. The authors would also like to thank Dr. Brent Adams and other personnel at Brigham Young University, Provo, UT for their assistance in OIM studies. References Adams, B.L., Nylander, C., Aydelotte, B., Ahmadi, S., Landon, C., Stucker, B.E., Janaki Ram, G.D., 2008. Accessing the elastic–plastic properties closure by rotation and lamination. Acta Materialia 56, 128–139. Allameh, S.M., Mercer, C., Popoola, D., Soboyejo, W.O., 2005. Microstructural characterization of ultrasonically welded aluminum. Journal of Engineering Materials and Technology 127, 65–74. Daniels, H.P.C., 1965. Ultrasonic welding. Ultrasonics 3, 190–196. De Vries, E., 2004. Mechanics and Mechanisms of Ultrasonic Metal Welding. PhD Thesis. Ohio State University. Enjio, T., 1986. Effects of ultrasonic vibration on diffusion welding of aluminum. Transactions of JWRI 15, 289–296. Eaves, A.E., Smith, A.W., Waterhouse, W.J., Sansome, D.H., 1975. Review of the application of ultrasonic vibrations to deforming metals. Ultrasonics 13, 162–170. George, J.L., 2006. Utilization of Ultrasonic Consolidation in Fabricating Satellite Decking. Masters Thesis. Utah State University, Logan, UT.
4924
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
Gunduz, I.E., Ando, T., Shattuck, E., Wong, P., Doumanidis, C., 2005. Enhanced diffusion and phase transformations during ultrasonic welding of zinc and aluminum. Scripta Materialia 52, 939–943. Janaki Ram, G.D., Robinson, C., Yang, Y., Stucker, B., 2007a. Use of ultrasonic consolidation for fabrication of multi-material structures. Rapid Prototyping Journal 13, 226–235. Janaki Ram, G.D., Yang, Y., Stucker, B., 2007b. Effect of process parameters on bond formation during ultrasonic consolidation of aluminum alloy 3003. Journal of Manufacturing Systems 25, 221–238. Jones, J.B., Powers, J.J., 1956. Ultrasonic welding. Welding Journal 35, 761–766. Joshi, K.C., 1971. The formation of ultrasonic bonds between metals. Welding Journal 50, 840–848. Kenik, E., Jahn, R., 2003. Microstructure of ultrasonic welded aluminum by orientation imaging microscopy. Microscopy and Microanalysis 9, 720–721. Kong, C.Y., Soar, R.C., Dickens, P.M., 2003. Characterization of aluminium alloy 6061 for the ultrasonic consolidation process. Materials Science and Engineering A A363, 99–106. Kong, C.Y., Soar, R.C., Dickens, P.M., 2004. Optimum process parameters for ultrasonic consolidation of 3003 aluminium. Journal of Materials Processing Technology 146, 181–187.
Kong, C.Y., 2005. Investigation of ultrasonic Consolidation for Embedding Active/Passive Fibers in Aluminum Matrices. Doctoral Thesis. Loughborough University, Loughborough, United Kingdom. Kreye, H., 1977. Melting phenomena in solid state welding processes. Welding Journal 56, 154s–158s. Langenecker, B., 1966. Effects of ultrasound on deformation characteristics of metals. IEEE Transactions on Sonic and Ultrasonic 13, 1–8. Li, L., Nagai, K., Yin, F., 2008. Progress in cold roll bonding of metals. Science and Technology of Advanced Materials 9, 1–11. Siggard, E.J., 2007. Investigative Research into the Structure Embedding of Electrical and Mechanical Systems Using Ultrasonic Consolidation (UC). Masters Thesis. Utah State University, Logan, UT. Weare, N.E., Antonevich, J.N., Monroe, R.E., 1960. Fundamental studies of ultrasonic welding. Welding Journal 39, 331s–341s. Yang, Y., Janaki Ram, G.D., Stucker, B., 2007. An experimental determination of optimum processing parameters for Al/SiC metal matrix composites made using ultrasonic consolidation. Journal of Engineering Materials and Technology 129, 538–549.
Contents lists available at ScienceDirect
Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec
Bond formation and fiber embedment during ultrasonic consolidation Y. Yang, G.D. Janaki Ram, B.E. Stucker ∗ Department of Mechanical and Aerospace Engineering, Utah State University, Logan, UT 84322-4130, USA
a r t i c l e
i n f o
Article history: Received 15 July 2008 Received in revised form 8 January 2009 Accepted 17 January 2009 Keywords: Additive manufacturing Ultrasonic consolidation Ultrasonic metal welding Interfacial microstructures Bonding mechanism Fiber embedment Metal matrix composites
a b s t r a c t The quality of ultrasonically consolidated parts critically depends on the bond quality between individual metal foils. This necessitates a detailed understanding of interface microstructures and ultrasonic bond formation mechanisms. In this work, the interface microstructures of a variety of ultrasonically consolidated similar and dissimilar metal samples were investigated. Samples with embedded SiC fibers were also investigated. Based on detailed microstructural studies, the mechanisms of foil bonding and fiber embedment in ultrasonic consolidation have been discussed. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Ultrasonic consolidation (UC) is a novel additive manufacturing process implementing ultrasonic metal welding for fabrication of complex three-dimensional (3D) structures from metal foils. The Solidica FormationTM UC machine (commercially introduced by Solidica Inc., USA, in 2000) incorporates an ultrasonic metal welding head, a 3-axis CNC milling head, a foil feeding apparatus, and a software program to automatically generate tool paths for material deposition and machining. In the UC process, a CAD model of the component to be fabricated is initially generated. The model is converted to an STL file format and systematically sliced into a number of horizontal layers using a customized computer program. Each horizontal layer corresponds to the thickness of the metal foil used for part fabrication. Fig. 1 illustrates the basic UC process. A rotating ultrasonic welding head, a.k.a. sonotrode, travels along the length of a thin metal foil placed on a base plate. The foil is held closely in contact with the base plate by applying a normal force via the rotating sonotrode. The sonotrode vibrates transversely to the direction of travel at a frequency of 20 kHz and at a user-set oscillation amplitude, while traveling over the metal foil to bond it to the base plate. After depositing a strip of foil, another foil could be deposited adjacent to it depending on the geometry of the part. This operation repeats until a complete layer is placed. After placing a layer, a CNC milling head subtractively shapes the layer to its
∗ Corresponding author. Tel.: +1 435 797 8173; fax: +1 435 797 2417. E-mail address: [email protected] (B.E. Stucker). 0924-0136/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2009.01.014
slice contour. This milling can occur after each layer or after several deposited layers. After shaping a layer, machining chips are blown away using compressed air and foil deposition starts for the next layer. These additive and subtractive operations repeat until the component is finished. Research on the practical applications of UC has been extensive. Previous work has demonstrated the unique capabilities of UC for fabrication of multi-functional 3D structures with high dimensional accuracy and desirable surface finish. George (2006) has demonstrated honeycomb panels with complex internal features, Janaki Ram et al. (2007a) have made multiple material structures, and Siggard (2007) has fabricated objects with integrated wiring and electronics. Another interesting application of UC is embedding fibers within metal matrices for producing metal matrix composites (MMCs) and for other purposes. A variety of fibers have been successfully embedded in aluminum matrices. Yang et al. (2007) have successfully embedded silicon carbide (SiC) fibers within Al 3003, and Kong (2005) reported similar successes on shape memory alloy and optical fibers using UC. Although the application potential of UC has been widely investigated, the mechanism of bond formation between metal foils during UC has not been clearly understood. Since the additive operation during UC is essentially a seam ultrasonic metal welding (UMW) application, an improved understanding of bond formation in UC can be obtained from the available information on bond formation in UMW. As discussed by Jones and Powers (1956) it is generally agreed that ultrasonic metal welding is a solid-state joining process. Apart from pure solid-state bonding, which is caused
4916
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924 Table 2 Process parameters used for SiC fiber embedment. Parameter
Value Al 3003/SiC/Al 3003
Vibrational amplitude (m) Normal force (kPa) Welding speed (mm/s)
Al 3003/SiC/Cu
10 275 28
12 275 28
the machine, tapes were manually placed onto the base plate and secured with an adhesive at the ends, followed by welding. 2.2. Ultrasonic consolidation of dissimilar metal foils
Fig. 1. Schematic of ultrasonic consolidation process (not to scale).
by atomic-level forces across nascent metal contact points, bond formation during UMW has been suggested by Joshi (1971) to be due to: (i) mechanical interlocking, (ii) interfacial melting, and (iii) metal diffusion. While all these mechanisms can contribute to bond formation, the operating or the dominant mechanism is expected to change with the materials in question and with the processing conditions. Therefore, it is important to know which mechanisms dominate under standard UC processing conditions. The current lack of consensus on foil bonding mechanisms in UC limits further exploitation of its processing capabilities. In view of the above, an attempt is made in this work to investigate bond formation during UC using a variety of ultrasonically consolidated similar and dissimilar metal samples. Further, an attempt is made to understand the mechanism of ceramic fiber embedment during UC. 2. Experimental details 2.1. Ultrasonic consolidation of Al 3003 and Ni 201 Metal tapes used in the current study are listed in Table 1. No special foil cleaning procedures were adopted to facilitate bonding. UC experiments were conducted using a Solidica FormationTM machine with a sonotrode diameter of 147 mm. Metal foils were deposited onto an Al 3003 base plate firmly bolted to the heat plate of the UC machine. The procedure for depositing Al 3003 tapes involved first depositing a layer of Al 3003 on the aluminum base plate. Subsequently, another layer of Al 3003 was deposited on the top of the previously deposited Al 3003 layer. The Al 3003 tapes were automatically fed through the tape feeding mechanism of the UC machine. The processing parameters used for Al 3003 foil consolidation were: oscillation amplitude, 16 m; welding speed, 28 mm/s; normal force, 1750 N; and base plate temperature, 149 ◦ C. These parameters were found by Janaki Ram et al. (2007b) to produce a high level of linear welding density for Al 3003. A similar procedure and processing parameters were implemented for depositing Ni 201 metal tapes. Since Ni 201 tapes were not available in standard coils suitable for automatic feeding by
Dual-material samples were fabricated from Al 3003 and Ni 201 tapes. Here again, deposition experiments were conducted using the Solidica FormationTM machine on an Al 3003 base plate. After depositing a few layers of Al 3003 one over another, a layer of Ni 201 was welded to the top most Al 3003 layer. Subsequently, another layer of Ni 201 was welded to the previously deposited Ni 201 layer. This layer arrangement was chosen to facilitate microstructural examination of Ni–Al and Ni–Ni interfaces. As before, Ni tapes were manually placed onto the substrate and secured with adhesive while Al 3003 layers were automatically fed. The process parameters were identical to those used for ultrasonic consolidation of Al 3003. 2.3. SiC fiber embedment The fiber embedment experiments were conducted using an ultrasonic seam welder at Loughborough University, UK, with a sonotrode diameter of 50 mm. This machine is as capable as the commercial UC machine in terms of joining thin metal foils. Manually fed 100 m thick foils were used and no base plate preheating was employed in all the experiments conducted using this machine. Silicon carbide fibers of 100 m diameter were embedded between Al 3003 foils along their length, following the embedding procedure described by Yang et al. (2007), who successfully embedded SiC fibers within Al 3003 matrices using this procedure. In order to understand the effects of foil material properties on fiber embedment, SiC fibers were embedded between Al 3003 foil and high-purity Cu (99.9%) foil of 100 m thickness. The process parameters used for these experiments are listed in Table 2. 2.4. Microstructural characterization Ultrasonically consolidated similar and dissimilar metal samples as well as fiber embedded samples were metallographically investigated. Transverse sections (perpendicular to the foil length direction) cut from the various samples were prepared for microstructural examination following standard metallographic practices. Some of the Al 3003 samples and fiber embedded samples were etched with Keller’s solution (HF-1%, HCl-1.5%, HNO3 -2.5%, and H2 O-95%). Ni 201 samples were etched with a mixture of 1 part 10% aqueous solution of CaCN and 1 part 10% aqueous solution of (NH4 )2 S2 O8 . Metal/metal and fiber/metal interface microstructures were examined using a scanning electron microscope (SEM). X-ray energy dispersive spectroscopy (EDS) was utilized for micro-
Table 1 Materials used for UC experiments. Material Al alloy 3003 (H18 condition) Ni alloy 201 (annealed condition)
Nominal composition (wt.%)
Dimensions
Al–1.2Mn–0.12Cu Ni–0.02C–0.35Mn–0.25Si–0.25Fe–0.15Cu
25 mm wide, 150 m thick foil 25 mm wide, 75 m thick foil
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924 Table 3 Process parameter used for temperature measurement experiments. Combination
Oscillation amplitude (m)
Normal force (kPa)
1 2 3 4
10 10 12 12
206 275 206 275
chemical characterization. Orientation imaging microscopy (OIM) was utilized for studying plastic deformation at the interfaces. 2.5. Temperature measurements Temperature measurements at weld interfaces were conducted using the ultrasonic seam welder at Loughborough University. Because of its constructional simplicity, this machine was considered more suitable for conducting these experiments. Measurements were made at various combinations of oscillation amplitude and normal force, listed in Table 3. The temperature measurement procedure consisted of the following: (i) depositing a metal tape onto a base plate (maintained at room temperature), (ii) placing a small K-type thermocouple (50 m diameter) on the top surface of the deposited tape, and (iii) depositing a second metal tape on the top of the previously deposited tape, with the small thermocouple sandwiched in
Fig. 2. Unbonded regions in ultrasonically consolidated Al 3003: (a) lowmagnification secondary electron image; (b) high-magnification back-scattered electron image (note the thin layer with a different contrast all around the defect (shown by arrow)).
4917
between. During the depositing of the second layer, the interface temperature was measured at a sampling rate of 1000 samples/s. 3. Results 3.1. Microstructures Fig. 2 shows the microstructure of an ultrasonically consolidated Al 3003 sample. The dark regions seen along the layer interfaces are the unbonded regions (defects). Examination of several unbonded regions using back-scattered electron imaging at higher magnifications indicated that these defects were covered with a thin differently contrasted layer (Fig. 2b). EDS spot analysis showed significantly higher oxygen content in this layer (Fig. 3a) than in the regions adjacent to it (Fig. 3b). Fig. 4 shows the microstructures of Al 3003/Ni 201 dissimilar metal samples. Generally, Ni 201 seemed to bond well to itself and to Al 3003 (Fig. 4a and b). As can be seen, the thickness of Ni foils remained practically unchanged. As shown in Fig. 4c, there were a few unbonded regions along the Ni/Ni interface. There was no evidence of metal melting or recrystallization at the weld interface. The weld interface appeared flat and mechanical interlocking did not seem to take place. Fig. 5a shows an image of the Al 3003/Ni 201 interface at a higher magnification, indicating no obvious evidence of intermetallic formation. The results of EDS line scans (for two elements, Al and Ni) performed across the Ni–Al interface (along the scan line shown in Fig. 5a (100 points, 1 m spot spacing, from Ni to Al side)) are shown in Fig. 5b. As can be seen, compositions seem to change sharply across the interface. While some amount of diffusion on a very fine scale cannot be ruled out, the results suggest that there was no bulk diffusion of Ni into Al and vice versa. Fig. 6 shows an inverse pole figure of a well-bonded Ni/Ni interface (generated from several OIM scans of contiguous areas), which
Fig. 3. (a) EDS spectra showing a distinct oxygen peak in the thin layer with a different contrast. (b) The oxygen peak is absent in the regions adjacent to, but outside the thin layer.
4918
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
Fig. 5. (a) Ni 201/Al 3003 interface, (b) EDS line scan results across the Ni 201/Al 3003 interface (along the scan line shown in (a), scan started on the Ni side).
Fig. 6. An image of several inverse pole figures of contiguous areas along a wellbonded Ni–Ni interface stitched together. The grains in the image are color coded to reflect their orientation. The line across the center of the image defines the Ni–Ni weld interface. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)
Fig. 4. SEM micrographs of the Al 3003/Ni 201 dual-material deposit: (a) shows a well-bonded Ni/Ni region, (b) shows a well-bonded Al 3003/Ni 201 region, and (c) shows a few Ni/Ni unbonded regions.
is color coded to indicate the crystallographic orientation ({h k l} direction parallel to the section normal) of each grain within the sample. Grains that have been plastically deformed typically show a smooth intra-grain color transition. Adams et al. (2008) suggested that the intra-grain color transition indicated rotations of the crystal lattice. Such smooth color transitions are evident in the picture, indicating that the foil interfaces plastically deform during the bonding process. However, there was no evidence of recrystallization at the weld interface. Interestingly, OIM examination of the unbonded regions along the Ni/Ni interface revealed a thin layer of extremely fine grains all around the unbonded region (Fig. 7),
Fig. 7. Inverse pole figure of an unconsolidated portion of the Ni–Ni interface. Note the extremely fine grains that are present along the defect boundaries.
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
Fig. 8. SiC fiber embedded between Al 3003 foils.
indicating that the crystals have been dynamically deformed into nano-sized grains during processing or that the foil is covered with an oxide layer and/or some kind of contamination at these locations. Such fine grains were not observed along the well-bonded portions of the interface (Fig. 6). Fig. 8 shows a SiC fiber embedded between Al 3003 foils. The fiber was not deformed during the process, retaining its circular external geometry. The center of the fiber was found to be located at the interface between top and bottom Al 3003 foils. Metal foils in the vicinity of the fiber were found to bond well, in particular. The cavity created by the placement of a SiC fiber between two foils was fully filled by the matrix metal, indicating significant plastic deformation of the matrix material during UC processing. Elemental mapping studies across the fiber/matrix interface showed no evidence of interdiffusion (Fig. 9). Fig. 10 shows a SiC fiber embedded between Al 3003 and Cu foils. In contrast to fiber embedment between two Al 3003 foils (Fig. 8), the center of the SiC fiber was found to be displaced into the bottom Al foil. Because the bottom aluminum foil is softer than the top copper foil, the SiC fiber is pressed into the Al foil during UC. It was also observed that the cavity created by the SiC fiber was filled only by Al 3003, and the geometry of the bottom surface of the copper foil indicated indentation by the SiC. An EDS line scan was conducted along the line shown in Fig. 11a (0.5 m point spacing, two elements (Al and Cu), started on the Cu side). As seen in Fig. 11b, there appears to be some amount of diffusion across the interface. However, the width of the diffusion zone is very small (less than 2 m)—too small to for conclusions with any degree of confidence, considering the resolution limit of the EDS technique. More importantly, however, our results show that there is no bulk diffusion of elements across the interface.
4919
Fig. 10. SiC fiber embedded between Al 3003/Cu foils. Note the aluminum flow around the fiber into the Cu side (shown by arrows).
3.2. Temperature measurements The results of temperature measurements at the weld interface are shown in Fig. 12. The peak temperature at the weld interface was found to vary from 68 to 98 ◦ C as a function of process parameters. 4. Discussion 4.1. Bond formation during UC 4.1.1. Mechanical interlocking Bonding due to mechanical interlocking can occur during ultrasonic metal welding. In a study of ultrasonic welding between Al and Au, Joshi (1971) observed a liquid-like metal flow of Au into the Al side at the weld interface and concluded that mechanical interlocking was responsible for bonding between Al and Au. However, in the present study, ultrasonically consolidated similar (Al/Al and Ni/Ni) and dissimilar (Al/Ni and Al/Cu) metal samples showed flat interfaces, suggesting absence of mechanical interlocking (Figs. 4 and 10). While mechanical interlocking during ultrasonic welding of similar metal foils appears to be very unlikely, it can occur in dissimilar welds, especially when the materials in question have a large difference in hardness (as in the case of Al and Au). 4.1.2. Metal melting Melting phenomena are related to the temperature rise at the metal/metal welding interface. During ultrasonic metal welding, heat is generated primarily due to dynamic frictional effects at the interface. Various methods have been applied to identify the actual temperature at the weld interface during UMW.
Fig. 9. Element mapping results of Al 3003/SiC/Al 3003 sample.
4920
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
Fig. 11. (a) Al 3003/Cu interface, (b) EDS line scan results across the Al 3003/Cu interface (along the scan line shown in (a), scan started on the Cu side).
These methods include: (1) embedding small thermocouples at the interface, (2) measuring the thermoelectric electromotive forces (e.m.f.) between workpieces during welding, and (3) observing the interface during welding using an infrared camera. Joshi (1971) measured the interface temperature during ultrasonic welding of a thin gold wire to a gold substrate using the thermocouple method, and the temperature was found to be no more than 120 ◦ C. Jones and Powers (1956) measured an interface temperature of around 204 ◦ C (400 ◦ F) during ultrasonic welding between pure aluminum sheets using the thermocouple method and 538 ◦ C (1000 ◦ F) during ultrasonic welding between Monel and aluminum sheets using
Fig. 12. Results of temperature measurements at the weld interface.
the e.m.f. method. Using the e.m.f. method, Weare et al. (1960) reported an interface temperature of 232 ◦ C (450 ◦ F) during ultrasonic welding between copper and Monel. Using an infrared camera with an accuracy of ±10 ◦ C, De Vries (2004) measured a temperature of 314 ◦ C (597 ◦ F) during ultrasonic metal welding of Al 6061. In all these cases, no substrate preheating was employed. While there is large variation in reported interface temperatures in ultrasonic metal welding, the measured temperatures are, in most cases, less than the melting point of the welded metal. A drawback of all the methods listed above is that the measured temperatures were all volume averaged temperatures of a certain volume of material at the metal/metal interface. It is possible that at some localized spots the temperature may reach or exceed the melting point of the metal, even if the averaged temperature at the weld interface is below the melting point. Consequently, microscopic examination was employed by several researchers to investigate the possibility of metal melting. In the investigations by Jones and Powers (1956), no metal melting was observed in ultrasonic welding between aluminum sheets. Daniels (1965) successfully achieved ultrasonic welds between copper and aluminum sheets without presence of metal melting. Joshi (1971) reported that no cast microstructures were observed during ultrasonic welding of Al, Cu, and Au. However, Kreye (1977) found some evidence of metal melting during ultrasonic welding of Cu2 Co. In their studies on ultrasonic welding between Al 1100 H19 and pure zinc sheets, Gunduz et al. (2005) reported localized melting of Al–Zn and solid solution formed at the welding interface. A preheat of 300 ◦ C was employed in this study. Thus, while most reports confirm that ultrasonic welding is a solidstate joining process, there is at least some evidence that localized melting can occur during ultrasonic welding depending on material combinations and processing conditions. In the present study, no evidence of melting was observed along the foil interfaces and the measured peak processing temperatures were significantly below the melting point of the metals. 4.1.3. Diffusion Ultrasonic welding results in only a modest increase in temperature at the weld interface. Moreover, the times available for diffusion are extremely short (owing to the short residence time of the sonotrode). Thus, from this standpoint, significant diffusion appears unlikely. However, there are indications that diffusivities can increase significantly under the conditions of ultrasonic welding. Ultrasonic welding can produce plastic deformation at very high strain rates, as reported by Gunduz et al. (2005). Consequently, dislocation densities and vacancy concentrations in the bond region can be significantly higher, which enhance diffusion. For example, Gunduz et al. (2005), in their studies on ultrasonic welding between Al and Zn, found that a strain rate up to 103 s−1 can be produced during welding, which increased the vacancy concentration to around 10−1 . As a consequence, they observed significant Zn diffusion into Al, nearly five orders of magnitude higher than the calculated diffusivity of Zn under normal conditions at the measured interface temperature. Therefore, it appears that ultrasonic welding can provide conditions for significant diffusion. Diffusion is further facilitated when high preheat temperature are employed. In the present study, bulk diffusion of elements across the interface was not observed (Fig. 5b for Al 3003/Ni 201 and Fig. 11b for Al 3003/Cu). Therefore, bond formation during UC does not seem to depend heavily on diffusion. However, there were indications of a very narrow diffusion zone (less than 2 m) at the weld interface (especially, in the case of Al 3003/Cu). While such a small-scale diffusion cannot be held completely responsible for bonding during UC, it can positively contribute to bonding. Diffusion allows for mass transfer across the interface. Further, any chemical interactions between component metals consequent to diffusion can
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
4921
lead to a stronger interface. However, it should be noted that diffusion is not a beneficial phenomenon in all cases, especially when dealing with metallurgically incompatible dissimilar material combinations. Diffusion often leads to chemical reactions, which can be detrimental to the bonding process and part mechanical properties. For example, formation of brittle intermetallics or low-melting eutectics is not generally desirable. 4.1.4. Atomic forces across nascent metal contact points While mechanical interlocking, interfacial melting, and diffusion all can occur during ultrasonic metal welding, they do not seem to have universal presence; rather, they seem to be specific to certain material combinations or processing conditions. Bonding in ultrasonic consolidation thus appears primarily to be solid-state, caused by atomic-level forces across the nascent metal contact points. As in the case of other solid-state welding processes, two conditions must be fulfilled for bond formation during ultrasonic welding: (i) generation of atomically clean surfaces, and (ii) intimate contact between clean metal surfaces. Surface oxide layers exist on all metallic materials, which prevent intimate contact between nascent metal surfaces. Hence removal of surface oxide layers is critical for achieving metallurgical bonding. During ultrasonic metal welding, as discussed by Daniels (1965), it is generally believed that frictional effects at the mating metal surfaces due to ultrasonic vibrations result in break-up and removal of surface oxide layers. In the current study, microstructural studies confirmed the presence of oxide layers along the defect boundaries (Fig. 2b). Oxide layers were, however, absent in the fully bonded or consolidated regions (Figs. 2a and 6). This confirms that ultrasonic vibrations (and consequent frictional effects) at the weld interface help remove surface oxide layers. However, this occurs only at the locations wherever there is surface contact. If the mating surfaces are not in contact, there cannot be any friction to break-up the surface oxide layers. Thus, the presence of oxide layers along the defect boundaries shown in Fig. 2b indicates that the mating surfaces across these defects had not come into contact. These noncontact regions with unremoved surface oxide layers show up as unbonded regions along the foil interfaces in the final deposit. It should be noted that unbonded regions along the foil interfaces can also be caused by cracking-related effects subsequent to bonding, especially under conditions of excessive energy input. Kong (2005) noticed some bond degradation due to excess energy input during ultrasonic consolidation of Al 3003. However, such unbonded regions would have nanometer-sized oxide layers, as the original oxide layers are removed in the process of bonding prior to cracking. Therefore, it can be deduced that the unbonded regions observed in the current study along the foil interfaces are not due to some cracking-related phenomena, but are due to a lack of complete surface contact between the mating foil surfaces as a result of surface roughness prior to bonding. The second necessary condition for solid-state welding is intimate nascent metal contact. Atomic-level forces can come into play only when the surfaces are brought close enough together. An impediment to this is surface roughness, which precludes complete contact between mating surfaces. In solid-state welding, this is normally overcome by plastic deformation at weld interfaces. The occurrence of plastic deformation was confirmed in the OIM studies (Fig. 6). Interfacial plastic deformation plays an important role in UC. First, it facilitates removal of the surface oxides. During ultrasonic welding, simultaneous application of normal and oscillating shear forces results in generation of dynamic interfacial stresses between the two mating surfaces at the contact points. As a consequence, cracks are initiated in the surface oxide layer. Kong et al. (2004) discussed how these interfacial stresses also induce plastic deformation in a thin layer of metal (∼20 m) just beneath the oxide
Fig. 13. Schematic of the mating surfaces at the beginning of ultrasonic consolidation.
layer. The nascent metal from beneath then extrudes through the cracks, resulting in the break-up of the oxide layers. These broken oxides are removed from the bond region by metal flow and are dispersed in the vicinity of the weld zone. Enjio (1986) identified such broken oxide fragments (0.05–0.2 m sizes) in an Al alloy diffusion weld subjected to ultrasonic vibrations. However, as noted by Kong et al. (2003) the dispersed oxide pieces may not be noticeable in all cases, as oxide layers are often only a few nanometers thick. Secondly, plastic deformation is crucial for producing a weld with satisfactory linear weld density (LWD), which is the length of well-bonded interface length divided by the total interface length, expressed in percentage. Roughness on the metal surfaces precludes 100% surface contact and, consequently, 100% bonding along the interface. Metal foils bond only at the initial surface contact points if plastic deformation does not occur. However, several researchers, such as Janaki Ram et al. (2007b), have reported nearly 100% LWD, which indirectly confirm the occurrence of metal plastic deformation. The situation at the mating surfaces at the beginning of ultrasonic welding can be visualized as shown in Fig. 13. As can be seen, contact between mating surfaces occurs only at surface asperities, leaving numerous no-contact regions along the interface. Bonding across these no-contact regions will not occur unless there is a mechanism to close these voids and to bring the mating surfaces into intimate contact. This is where plastic flow is believed to play a major role. Initially, bonds are established at the existing surface contact points. As the process progresses, these bonded regions grow in size, aided by plastic deformation. Plastic deformation at the bonded regions results in squeezing of metal into the voids and the mating surfaces across the void regions approach. As this happens, new points come into contact, leading to friction, oxide layer removal and bonding. These newly bonded regions also grow during subsequent deformation cycles, generating more contact points. This process can result in sound metallurgical bonding with relatively high linear weld density levels. The bonding process in ultrasonic welding can thus be recognized as repeated occurrence of two distinct stages, contact and bond stages, with each leading to the other. During the contact stage, new contact points are generated by plastic deformation of existing bonds. During the bond stage, oxide layers are removed and metallurgical bonds are formed at the contact points generated in the contact stage. Plastic deformation in ultrasonic metal welding significantly differs from that in several other well-known solid-state welding processes, such as roll bonding and friction welding. The latter involve bulk plastic deformation. In cold roll bonding, for example, typically a large rolling load is applied to produce solid-state bonds between materials. Materials during roll bonding undergo bulk plastic deformation and, consequently, significant reduction in section thickness. In a recently publication, Li et al. (2008) reported that a reduction in section thickness up to 60–70% is required for achieving satisfactory bonding with cold roll bonding. In the UC process, on the other hand, the forces applied are very low and plastic deformation is more or less confined to the interface. As can be seen in Fig. 4, the thickness of Ni 201 foils remained practically unchanged after consolidation.
4922
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
Interfacial plastic deformation during ultrasonic welding is expected to be facilitated by the “Blaha effect” or “acoustic softening” effect, as described by Langenecker (1966). Langenecker found that, in the presence of ultrasonic energy, metallic materials are known to experience significant softening, which is not connected to any rise in temperature resulting from being subjected to an ultrasonic field. In fact, ultrasonic energy was shown to be more effective than thermal energy in reducing the flow stress of metallic materials. For example, Eaves et al. (1975) reported that bulk heating can reduce stresses by 45% while ultrasonic vibration reduces it by 75%. In terms of energy density, Langenecker (1966) reported that it takes approximately 1022 eV/cm3 of thermal energy density to produce a zero stress in aluminum without ultrasonic superimposition, while only about 1015 eV/cm3 using ultrasonic energy. He explained this difference as “acoustic energy is assumed to be absorbed only at those regions in the metal lattice which are known to carry out the mechanisms of plastic deformation. Heat, on the other hand, is distributed rather homogeneously among all the atoms of the crystal including those which do not participate in the mechanisms of plastic deformation.” In addition to acoustic softening, thermal softening can also occur at the weld interface due to frictional heating, further contributing to a reduction in flow stress. Thus, although the forces involved in typical ultrasonic welding are generally modest, there seem to be conditions enough for causing plastic deformation and metal flow, when considering the combined acoustic and thermal softening effects. Plastic deformation and temperature rise at the interface can provide the necessary driving force for recrystallization during ultrasonic welding, although whether there is adequate time for recrystallization or not is still unknown. Recrystallization is desirable as it brings in necessary readjustments to the grain structure at the interface and replaces the strained grain structure on both sides of the interface with a set of freshly formed strainfree, fine grains. It facilitates grain continuity across the interface and a smooth transition of weld zone microstructure from the rest of the material. However, no conclusive evidence exists on whether recrystallization takes place during ultrasonic welding. Kenik and Jahn (2003) reported recrystallized grain structures in ultrasonic joint of Al alloys, and Allameh et al. (2005) claimed that recrystallization of grains does not occur due to insufficient thermal energy generated during relatively short welding duration although satisfactory bonding was demonstrated in all cases. In the current work, no recrystallization was observed in Ni 201 samples. It appears that specific processing conditions and the material systems in question govern the recrystallization process. 4.1.5. Effects of foil surface condition Foil surface condition is a critical factor in ultrasonic consolidation. Thick adherent surface oxide scales and large-scale surface roughness are undesirable, as surface oxide layers prevent intimate nascent metal contact. The ease with which oxide layers can be removed during ultrasonic welding depends on the ratio of metal oxide hardness to nascent metal hardness—higher ratios facilitate easier removal. This is the reason why Al alloys, with a very high oxide-to-metal hardness ratio, are one of the best-suited materials for ultrasonic welding. Noble metals such as gold, which do not have surface oxide layers, are quite amenable for ultrasonic welding as well. Materials that present difficulties with oxide layer removal have been reported to be problematic for ultrasonically welding. For example, Kong et al. (2003) found that Al–Mg–Si alloy 6061 was difficult to ultrasonically consolidate, which was attributed to difficulties with oxide layer removal, thought to be due to the presence of MgO in surface oxide layers of these alloys. Satisfactory bonding between Al 6061 foils was achieved by acid etching the
foil surfaces just prior to welding for removing the surface oxide layers. Due to surface roughness, when two metal surfaces are put together, 100% surface contact is not possible and the no-contact areas are inevitable. Generally, metal surfaces with a low-level of surface roughness, which produce fewer and smaller no-contact areas, are desirable for bond formation during ultrasonic consolidation. When the consolidation parameters are right, these no-contact areas are eliminated due to adequate plastic deformation, leading to nearly 100% linear weld densities. The effect of surface roughness on bond formation was investigated by Janaki Ram et al. (2007b). Examining the surfaces of ultrasonically consolidated foils, the authors noticed significant roughening of foil top surfaces as a result of sonotrode motion. This roughness was shown to be the major source of defects for unbonded regions in ultrasonically consolidated parts. Further, a simple technique, called surface machining, was proposed by the authors, which involved removal of a thin layer (∼30 m) of metal from consolidated foil surfaces prior to subsequent layer deposition. With this technique, nearly 100% linear weld densities were achieved even under processing conditions that would have otherwise resulted in a large number of unbonded regions along foil interfaces. It was claimed that, besides smoothening the substrate surface, surface machining mechanically removes the substrate surface oxide layers, which facilitates bond formation. 4.2. Fiber embedment during UC Microstructural examination produced no evidence of significant diffusion or chemical reactions between fibers and matrix materials during fiber embedment using UC. Therefore, it appears that fibers are mechanically trapped between the top and bottom foils during UC, with fiber embedment made possible by plastic deformation of matrix materials. When a fiber is placed between two metal foils, a large gap is created between the metal foils in the vicinity of the fiber. During consolidation, the fiber acts as a stress riser leading to larger stresses and more extensive matrix material plastic deformation in the vicinity of the fiber than at regions away from the fiber. Consequently, metal flow occurs into the gap around the fiber resulting in sound fiber embedment. Similar phenomena were noticed by Janaki Ram et al. (2007a) during stainless steel mesh embedment between Al 3003 foils using UC. If the fiber is sandwiched between metal foils of similar hardness, both metal foils deform to the same extent, resulting in a fiber being embedded with its center aligned with the metal/metal interface (as noticed in the present study in the case of Al 3003/SiC/Al 3003, Fig. 8). When dissimilar metal foils with a large difference in hardness are used, the softer metal tends to deforms more than the harder metal, resulting in the fiber being embedded more into the softer metal side (as noticed in the current study in the case of Al 3003/SiC/Cu, Fig. 10). To further confirm the suggestion of mechanical entrapment of fibers, tensile fractured Al 3003 specimens containing multiple SiC fibers were examined under SEM. Studies revealed a few instances of fiber pull-out (Fig. 14a). Examination of a pulled-out fiber at higher magnifications showed the fiber surface morphology (Fig. 14b and c). The surfaces of the holes created in the matrix by fiber pull-out showed a precisely matching surface morphology (Fig. 14d), indicating that the matrix tightly gripped the fiber and, as a consequence, took the negative image of the fiber surface. Clearly, such features would not show up if there had been any chemical bonding between the fiber and matrix. These observations thus provide strong evidence that fibers are mechanically entrapped in the matrix material during UC.
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
4923
Fig. 14. (a) Tensile fractured Al 3003 specimen containing SiC fibers. (b and c) Surface morphology of the pulled-out fiber. (d) Surface morphology of the “matrix hole” created by fiber pull-out.
5. Conclusions
Acknowledgements
The interface microstructures of various ultrasonically consolidated similar (Al 3003/Al 3003 and Ni 201/Ni 201) and dissimilar (Al 3003/Ni 201, Al 3003/Cu) metal samples were reported. None of the samples showed evidence of mechanical interlocking, localized metal melting, significant diffusion, or recrystallization at the weld interface. Bond formation between metal foils during ultrasonic consolidation under standard processing conditions appears to be essentially solid-state, which is caused by atomiclevel forces across the nascent metal contact points. Two crucial conditions for bond formation during ultrasonic consolidation are: (i) removal of surface oxide layers, and (ii) intimate contact between clean metal surfaces. Frictional effects at the weld interface ensure removal of surface oxide layers. Intimate nascent metal contact is facilitated by interfacial plastic deformation. The process of ultrasonic consolidation seems to involve repeated occurrence of contact and bond stages, with each leading to the other. During the contact stage, new contact points are generated by plastic deformation of existing bonds. During the bond stage, oxide layers are removed and metallurgical bonds are formed at the contact points generated in the preceding contact stage. Al 3003/SiC/Al 3003 and Al 3003/SiC/Cu fiber embedded samples showed no evidence of diffusion or chemical reactions between fiber and matrix materials. Fibers are mechanically entrapped in the matrix. Matrix material plastic deformation facilitates sound fiber embedment during ultrasonic consolidation.
The authors thank the National Science Foundation (DMI 0522908) and the Office of Naval Research (under Grant no. N000140710633) for supporting this work. The National Science Foundation provided travel support (International Research & Education in Engineering) facilitating the experiments at Loughborough University in the UK. The authors would like to thank Dr. Rupert Soar and Dr. Dezhi Li at Loughborough University for their constructive thoughts and suggestions. The authors would also like to thank Dr. Brent Adams and other personnel at Brigham Young University, Provo, UT for their assistance in OIM studies. References Adams, B.L., Nylander, C., Aydelotte, B., Ahmadi, S., Landon, C., Stucker, B.E., Janaki Ram, G.D., 2008. Accessing the elastic–plastic properties closure by rotation and lamination. Acta Materialia 56, 128–139. Allameh, S.M., Mercer, C., Popoola, D., Soboyejo, W.O., 2005. Microstructural characterization of ultrasonically welded aluminum. Journal of Engineering Materials and Technology 127, 65–74. Daniels, H.P.C., 1965. Ultrasonic welding. Ultrasonics 3, 190–196. De Vries, E., 2004. Mechanics and Mechanisms of Ultrasonic Metal Welding. PhD Thesis. Ohio State University. Enjio, T., 1986. Effects of ultrasonic vibration on diffusion welding of aluminum. Transactions of JWRI 15, 289–296. Eaves, A.E., Smith, A.W., Waterhouse, W.J., Sansome, D.H., 1975. Review of the application of ultrasonic vibrations to deforming metals. Ultrasonics 13, 162–170. George, J.L., 2006. Utilization of Ultrasonic Consolidation in Fabricating Satellite Decking. Masters Thesis. Utah State University, Logan, UT.
4924
Y. Yang et al. / Journal of Materials Processing Technology 209 (2009) 4915–4924
Gunduz, I.E., Ando, T., Shattuck, E., Wong, P., Doumanidis, C., 2005. Enhanced diffusion and phase transformations during ultrasonic welding of zinc and aluminum. Scripta Materialia 52, 939–943. Janaki Ram, G.D., Robinson, C., Yang, Y., Stucker, B., 2007a. Use of ultrasonic consolidation for fabrication of multi-material structures. Rapid Prototyping Journal 13, 226–235. Janaki Ram, G.D., Yang, Y., Stucker, B., 2007b. Effect of process parameters on bond formation during ultrasonic consolidation of aluminum alloy 3003. Journal of Manufacturing Systems 25, 221–238. Jones, J.B., Powers, J.J., 1956. Ultrasonic welding. Welding Journal 35, 761–766. Joshi, K.C., 1971. The formation of ultrasonic bonds between metals. Welding Journal 50, 840–848. Kenik, E., Jahn, R., 2003. Microstructure of ultrasonic welded aluminum by orientation imaging microscopy. Microscopy and Microanalysis 9, 720–721. Kong, C.Y., Soar, R.C., Dickens, P.M., 2003. Characterization of aluminium alloy 6061 for the ultrasonic consolidation process. Materials Science and Engineering A A363, 99–106. Kong, C.Y., Soar, R.C., Dickens, P.M., 2004. Optimum process parameters for ultrasonic consolidation of 3003 aluminium. Journal of Materials Processing Technology 146, 181–187.
Kong, C.Y., 2005. Investigation of ultrasonic Consolidation for Embedding Active/Passive Fibers in Aluminum Matrices. Doctoral Thesis. Loughborough University, Loughborough, United Kingdom. Kreye, H., 1977. Melting phenomena in solid state welding processes. Welding Journal 56, 154s–158s. Langenecker, B., 1966. Effects of ultrasound on deformation characteristics of metals. IEEE Transactions on Sonic and Ultrasonic 13, 1–8. Li, L., Nagai, K., Yin, F., 2008. Progress in cold roll bonding of metals. Science and Technology of Advanced Materials 9, 1–11. Siggard, E.J., 2007. Investigative Research into the Structure Embedding of Electrical and Mechanical Systems Using Ultrasonic Consolidation (UC). Masters Thesis. Utah State University, Logan, UT. Weare, N.E., Antonevich, J.N., Monroe, R.E., 1960. Fundamental studies of ultrasonic welding. Welding Journal 39, 331s–341s. Yang, Y., Janaki Ram, G.D., Stucker, B., 2007. An experimental determination of optimum processing parameters for Al/SiC metal matrix composites made using ultrasonic consolidation. Journal of Engineering Materials and Technology 129, 538–549.