Self-Piercing Riveted Joint of Vibration-Damping Steel and Aluminum Alloy

: In this study, the self-piercing rivet (SPR) joining of vibration-damping steel and aluminum alloy (Al5052-H32) is successfully carried out, for the ﬁrst time to our knowledge, and the e ﬀ ects of die type and joint conﬁguration on the mechanical performance, failure mode, and geometrical characteristics of the new joint are investigated. The vibration-damping steel and Al5052-H32 SPR joint exhibits the largest tensile–shear load when a ﬂat die is used. An increase in the die taper angle and diameter decreases the mechanical performance of the joint due to the increase in volume of the die, leading to a smaller interlock width of the joint. The joint conﬁguration with Al5052-H32 as a top sheet has superior mechanical performance compared with the reverse conﬁguration, owing to the increase of the interlock width. All SPR joints of vibration-damping steel and Al5052-H32 show consistent rivet pull-out failure, regardless of the joint conﬁguration, because of relatively small interlock width. It is also found that these SPR joints show better mechanical performance than those of SPFC590DP (a skin material of the vibration-damping steel) and Al5052-H32 under the Al5052-H32–top conﬁguration.


Introduction
A vibration-damping steel plate is a sandwich-structured steel panel with a viscoelastic adhesive layer between two steel sheets. This plate is mainly considered for car cowl and dash panels, since it not only reduces noise and vibration, but also provides proper mechanical strength. In response to the environmental regulations and market forces for improved fuel economy, the automotive industry has made an effort to enhance fuel efficiency and reduce carbon emissions. One of the key solutions to meet current regulations is weight reduction by application of lightweight materials, including advanced high-strength steel, aluminum alloy, and composites [1,2]. Consequently, dissimilar joining of the vibration-damping steel to lightweight materials is inevitable to realize a variety of lightweight designs for cowl and dash panels. These days, aluminum alloy, in particular, is being increasingly adopted to replace conventional steels as the lightweight material in automobiles. Accordingly, it is necessary to develop dissimilar joining methods of aluminum alloys to the vibration-damping steel sheets.
However, it is challenging to join the vibration-damping steels to aluminum alloys, since conventional fusion joining methods for metals, such as resistance spot welding [3][4][5] and laser welding [6][7][8], are not applicable to aluminum alloys and steel dissimilar joints due to their large The materials used in this study were vibration-damping steel, aluminum alloy (Al5052-H32), and steel alloy (SPFC590DP). As shown in Figure 1, the vibration-damping steel consisted of two 0.7mm-thick SFC590DP sheets and one viscoelastic adhesive polymer layer with 0.1 mm thickness between them. As a result, the vibration-damping steel had a 1.5 mm thickness. The tensile strengths and thicknesses of the materials used for the SPR joining process are listed in Table 1. The tensile strengths of all materials were evaluated using the specimens fabricated according to the ASTM E8M specification.

SPR Equipment, Rivet, and Die
The SPR joints were made from a hydraulic-type riveting machine (Rivset Gen2, BÖLLHOFF) that had a maximum setting force of 78 kN. The rivet (boron steel with Almac ® coating (combination of aluminum and zinc), 480 ± 30 HV in hardness, supplied by BÖLLHOFF) with a flat countersunk was used in the SPR process. The engineering drawing of the rivet is illustrated in Figure 2. The rivet shank diameter and rivet length were 5.3 mm and 4.5 mm, respectively. The rivet length of 4.5 mm was selected to be approximately 2 mm thicker than the total thickness of base materials (2.6-2.7 mm). In this study, three die types (supplied by BÖLLHOFF) were tested to investigate the effect of the die geometry on the joining performance. Geometrical information of these dies is presented in Figure 3. Type A1 (diameter: 8.8 mm, depth: 1.8 mm) die was a benchmark die that had a basic flat-bottom shape, and both type B (diameter: 9.5 mm, depth: 1.8 mm) and type C (diameter: 9.0 mm, depth: 2.0 mm) had a hump at the center of the die cavity. The hump of type B was conical and the height of the hump was smaller than the depth of the die cavity. However, the hump height of type C was larger than the depth of the die cavity, resulting in more extrusion. In addition, for the benchmark die, the influence of cavity diameter and cavity taper angle on the joint characteristic was examined. In this examination, two modified benchmark dies (type A2 and A3) were used. Type A2 die had a cavity taper angle of 5° (which was 5° more than that of type A1), and type A3 die had a diameter of 9.2 mm (which was 0.2 mm more than that of type A1).

SPR Equipment, Rivet, and Die
The SPR joints were made from a hydraulic-type riveting machine (Rivset Gen2, BÖLLHOFF) that had a maximum setting force of 78 kN. The rivet (boron steel with Almac ® coating (combination of aluminum and zinc), 480 ± 30 HV in hardness, supplied by BÖLLHOFF) with a flat countersunk was used in the SPR process. The engineering drawing of the rivet is illustrated in Figure 2. The rivet shank diameter and rivet length were 5.3 mm and 4.5 mm, respectively. The rivet length of 4.5 mm was selected to be approximately 2 mm thicker than the total thickness of base materials (2.6-2.7 mm). In this study, three die types (supplied by BÖLLHOFF) were tested to investigate the effect of the die geometry on the joining performance. Geometrical information of these dies is presented in Figure 3. Type A1 (diameter: 8.8 mm, depth: 1.8 mm) die was a benchmark die that had a basic flat-bottom shape, and both type B (diameter: 9.5 mm, depth: 1.8 mm) and type C (diameter: 9.0 mm, depth: 2.0 mm) had a hump at the center of the die cavity. The hump of type B was conical and the height of the hump was smaller than the depth of the die cavity. However, the hump height of type C was larger than the depth of the die cavity, resulting in more extrusion. In addition, for the benchmark die, the influence of cavity diameter and cavity taper angle on the joint characteristic was examined. In this examination, two modified benchmark dies (type A2 and A3) were used. Type A2 die had a cavity taper angle of 5 • (which was 5 • more than that of type A1), and type A3 die had a diameter of 9.2 mm (which was 0.2 mm more than that of type A1).

Cross-Sectional Analysis of SPR Joint
Cross-sectional analysis of the SPR joint was carried out to determine the geometrical features of the SPR joint. For this, three geometrical indices of the SPR joint were measured, as presented in Figure 4-head height (the distance between the surface of the rivet head and upper surface of the top sheet), interlock width (the distance between the tip of the deformed rivet shank and the pierced point of the top sheet), and bottom thickness (remaining thickness of the bottom sheet after the riveting). In this study, after the preliminary SPR joining experiments, the setting force (or riveting load) of the riveting machine was carefully selected to be 27-42 kN, so that all joints made from various die types and joint configurations could have a head height of 0 ± 0.05 mm. The head height of the joint was measured with a dial gauge, and the interlock width and bottom thickness were measured by an optical microscope system with an image analysis tool.

Cross-Sectional Analysis of SPR Joint
Cross-sectional analysis of the SPR joint was carried out to determine the geometrical features of the SPR joint. For this, three geometrical indices of the SPR joint were measured, as presented in Figure 4-head height (the distance between the surface of the rivet head and upper surface of the top sheet), interlock width (the distance between the tip of the deformed rivet shank and the pierced point of the top sheet), and bottom thickness (remaining thickness of the bottom sheet after the riveting). In this study, after the preliminary SPR joining experiments, the setting force (or riveting load) of the riveting machine was carefully selected to be 27-42 kN, so that all joints made from various die types and joint configurations could have a head height of 0 ± 0.05 mm. The head height of the joint was measured with a dial gauge, and the interlock width and bottom thickness were measured by an optical microscope system with an image analysis tool.

Cross-Sectional Analysis of SPR Joint
Cross-sectional analysis of the SPR joint was carried out to determine the geometrical features of the SPR joint. For this, three geometrical indices of the SPR joint were measured, as presented in Figure 4-head height (the distance between the surface of the rivet head and upper surface of the top sheet), interlock width (the distance between the tip of the deformed rivet shank and the pierced point of the top sheet), and bottom thickness (remaining thickness of the bottom sheet after the riveting). In this study, after the preliminary SPR joining experiments, the setting force (or riveting load) of the riveting machine was carefully selected to be 27-42 kN, so that all joints made from various die types and joint configurations could have a head height of 0 ± 0.05 mm. The head height of the joint was measured with a dial gauge, and the interlock width and bottom thickness were measured by an optical microscope system with an image analysis tool. point of the top sheet), and bottom thickness (remaining thickness of the bottom sheet after the riveting). In this study, after the preliminary SPR joining experiments, the setting force (or riveting load) of the riveting machine was carefully selected to be 27-42 kN, so that all joints made from various die types and joint configurations could have a head height of 0 ± 0.05 mm. The head height of the joint was measured with a dial gauge, and the interlock width and bottom thickness were measured by an optical microscope system with an image analysis tool.

Tensile-Shear Test for SPR Joint
The mechanical performance of the SPR joint was examined by a tensile-shear test (test machine: Universal Testing Machine of Shimadzu AG-300kNX with a maximum load capacity of 30 ton). Specimen dimensions and test procedure for the tensile-shear test conformed to KS B ISO 14,273 standard. Figure 5 shows the schematic diagram of the tensile-shear test specimen. The test was carried out at a crosshead speed of 5 mm/min.

Tensile-Shear Test for SPR Joint
The mechanical performance of the SPR joint was examined by a tensile-shear test (test machine: Universal Testing Machine of Shimadzu AG-300kNX with a maximum load capacity of 30 ton). Specimen dimensions and test procedure for the tensile-shear test conformed to KS B ISO 14,273 standard. Figure 5 shows the schematic diagram of the tensile-shear test specimen. The test was carried out at a crosshead speed of 5 mm/min.  Figure 6 shows the cross-sectional views of the SPR joints of vibration-damping steel (top) to aluminum alloy (bottom) formed with different dies, as presented in Figure 3. While the cavity inside the rivet shank (or rivet cavity) was formed for the dies with a flat-bottomed cavity (types A1, A2, and A3), the rivet cavity was significantly reduced when non-flat dies (type B and C) were used due to the protruded part of the dies. The punched slug from the top vibration-damping steel embedded into the shank cavity separated in common. On the other hand, the top sheets of the joints barely separated, showing a little slipping between the outer steel sheets of top vibration-damping steel. Figure 6. Cross-sectional images for SPR joint of vibration-damping steel (top) and Al5052-H32  Figure 6 shows the cross-sectional views of the SPR joints of vibration-damping steel (top) to aluminum alloy (bottom) formed with different dies, as presented in Figure 3. While the cavity inside the rivet shank (or rivet cavity) was formed for the dies with a flat-bottomed cavity (types A1, A2, and A3), the rivet cavity was significantly reduced when non-flat dies (type B and C) were used due to the protruded part of the dies. The punched slug from the top vibration-damping steel embedded into the shank cavity separated in common. On the other hand, the top sheets of the joints barely separated, showing a little slipping between the outer steel sheets of top vibration-damping steel.

SPR Joint between Vibration-Damping Steel (Top) and Aluminum Alloy (Bottom)
aluminum alloy (bottom) formed with different dies, as presented in Figure 3. While the cavity inside the rivet shank (or rivet cavity) was formed for the dies with a flat-bottomed cavity (types A1, A2, and A3), the rivet cavity was significantly reduced when non-flat dies (type B and C) were used due to the protruded part of the dies. The punched slug from the top vibration-damping steel embedded into the shank cavity separated in common. On the other hand, the top sheets of the joints barely separated, showing a little slipping between the outer steel sheets of top vibration-damping steel. The rivet head gap between the rivet head edge and the upper surface of the bended top sheet in the joint represents the amount of top surface sealing. The gap increased with the increase of the die taper angle and diameter, because the die-to-rivet volume ratio (R) increased as the die taper and diameter increased, as presented in Figure 3. As described by Ma et al. [19], for R > 1, the die with the larger cavity imposed fewer constraints on the material stacks, and thus the bending curvature of the top sheet increased with increasing R and resulted in the gap increase. In the case of the not-flat dies (type B and C), the projection (or tip) of the dies did not reduce the gap. Especially for the type C die, Figure 6. Cross-sectional images for SPR joint of vibration-damping steel (top) and Al5052-H32 (bottom) with respect to SPR die type. (a) Type A1 die, (b) type A2 die, (c) type A3 die, (d) type B die, and (e) type C die.
The rivet head gap between the rivet head edge and the upper surface of the bended top sheet in the joint represents the amount of top surface sealing. The gap increased with the increase of the die taper angle and diameter, because the die-to-rivet volume ratio (R) increased as the die taper and diameter increased, as presented in Figure 3. As described by Ma et al. [19], for R > 1, the die with the larger cavity imposed fewer constraints on the material stacks, and thus the bending curvature of the top sheet increased with increasing R and resulted in the gap increase. In the case of the not-flat dies (type B and C), the projection (or tip) of the dies did not reduce the gap. Especially for the type C die, its R was less than that of type A1 die but the gap was more than that of the type A1 die. Hence, the tip of the not-flat die was not effective for top surface sealing, but only for reducing the rivet cavity. Figure 7 shows the measurement results of geometrical indices for the SPR joint of the vibration-damping steel (top) and Al5052-H32 (bottom) with respect to the SPR die type. As seen in the figure, the joint made from the type A1 die has the largest interlock width, and it decreases in the order of type A2, type A3, type B, and type C. The flat dies (types A1, A2, and A3) induced a larger interlock width compared with the non-flat dies (types B and C). For the flat dies, the interlock width was not only reduced by the increase of the die taper angle, but also the increase of the die diameter. The increase in both the die taper angle and die diameter contributed to the increase of the die volume in the radial direction. Consequently, the bottom sheet (Al5052-H32) was more deformed in the radial direction, leading to less interlocking between the rivet shank and lower sheet. The bottom thickness of the joints was much larger than the interlock width of the joints, due to the small penetration made by the rivet shank in the bottom sheet. For the flat die, the bottom thickness of the joints had an opposite trend with the interlock width; however, no distinct trend was observed when the bottom thickness of the non-flat die was compared with that of the flat die.
Appl. Sci. 2019, 9, x FOR PEER REVIEW 6 of 13 its R was less than that of type A1 die but the gap was more than that of the type A1 die. Hence, the tip of the not-flat die was not effective for top surface sealing, but only for reducing the rivet cavity. Figure 7 shows the measurement results of geometrical indices for the SPR joint of the vibrationdamping steel (top) and Al5052-H32 (bottom) with respect to the SPR die type. As seen in the figure, the joint made from the type A1 die has the largest interlock width, and it decreases in the order of type A2, type A3, type B, and type C. The flat dies (types A1, A2, and A3) induced a larger interlock width compared with the non-flat dies (types B and C). For the flat dies, the interlock width was not only reduced by the increase of the die taper angle, but also the increase of the die diameter. The increase in both the die taper angle and die diameter contributed to the increase of the die volume in the radial direction. Consequently, the bottom sheet (Al5052-H32) was more deformed in the radial direction, leading to less interlocking between the rivet shank and lower sheet. The bottom thickness of the joints was much larger than the interlock width of the joints, due to the small penetration made by the rivet shank in the bottom sheet. For the flat die, the bottom thickness of the joints had an opposite trend with the interlock width; however, no distinct trend was observed when the bottom thickness of the non-flat die was compared with that of the flat die. The tensile-shear load of the specimens is shown in Figure 8. The joint made from the type A1 die had the highest tensile-shear load, owing to the large interlock width, and it decreased with the decrease of the interlock width (refer to Figures 7 and 8). However, the tensile-shear load of the joint The tensile-shear load of the specimens is shown in Figure 8. The joint made from the type A1 die had the highest tensile-shear load, owing to the large interlock width, and it decreased with the decrease of the interlock width (refer to Figures 7 and 8). However, the tensile-shear load of the joint formed by a type C die was relatively high, even though it had the smallest interlock width, as presented in Figure 7. This is due to the fact that the type C die had a relatively small die-to-rivet ratio (R = 1.254), which imposed more constraints on the bottom material. This may have further induced the strain hardening of the bottom sheet (Al5052-H32), contributing to the overall improvement in the joint's mechanical performance. The failure images in the lap shear testing of the SPR joints are presented in Figure 9. Regardless of the die type, rivet pull-outs from the bottom sheet were observed for the stack-up. Neither top sheet fracture nor rivet fracture was found for all cases. In addition, the interlocking between the rivet and bottom sheet completely failed, as presented in the figure. Thus, the rivet pull-out failures were mainly due to the small interlock widths of the joints (<0.3 mm for all cases of this study), where less than approximately one-fourth of the rivet shank was only entrapped by the bottom sheet. Consequently, the rivet was apt to be pulled out from the bottom sheet because of the weak interlocking between the rivet shank and bottom sheet. The rivet pull-out failures of the SPR joints due to the small interlock width have been reported in other studies as well [19,23,24].  Figure 10 shows the cross-sectional views of SPR joints of aluminum alloy (top) to vibrationdamping steel (bottom) formed with different die types, as presented in Figure 3. Different from the reversed sheet material's stack configuration, the rivet shank in these SPR joints were fully filled with the stack materials, so no rivet cavity was formed, which is commonly reported when a ductile metal such as aluminum alloy is used as the top material [19,24,25]. The rivet penetrated through the top sheet of the bottom vibration-damping steel, but not into the bottom sheet of the vibration-damping The failure images in the lap shear testing of the SPR joints are presented in Figure 9. Regardless of the die type, rivet pull-outs from the bottom sheet were observed for the stack-up. Neither top sheet fracture nor rivet fracture was found for all cases. In addition, the interlocking between the rivet and bottom sheet completely failed, as presented in the figure. Thus, the rivet pull-out failures were mainly due to the small interlock widths of the joints (<0.3 mm for all cases of this study), where less than approximately one-fourth of the rivet shank was only entrapped by the bottom sheet. Consequently, the rivet was apt to be pulled out from the bottom sheet because of the weak interlocking between the rivet shank and bottom sheet. The rivet pull-out failures of the SPR joints due to the small interlock width have been reported in other studies as well [19,23,24]. The failure images in the lap shear testing of the SPR joints are presented in Figure 9. Regardless of the die type, rivet pull-outs from the bottom sheet were observed for the stack-up. Neither top sheet fracture nor rivet fracture was found for all cases. In addition, the interlocking between the rivet and bottom sheet completely failed, as presented in the figure. Thus, the rivet pull-out failures were mainly due to the small interlock widths of the joints (<0.3 mm for all cases of this study), where less than approximately one-fourth of the rivet shank was only entrapped by the bottom sheet. Consequently, the rivet was apt to be pulled out from the bottom sheet because of the weak interlocking between the rivet shank and bottom sheet. The rivet pull-out failures of the SPR joints due to the small interlock width have been reported in other studies as well [19,23,24].  Figure 10 shows the cross-sectional views of SPR joints of aluminum alloy (top) to vibrationdamping steel (bottom) formed with different die types, as presented in Figure 3. Different from the reversed sheet material's stack configuration, the rivet shank in these SPR joints were fully filled with the stack materials, so no rivet cavity was formed, which is commonly reported when a ductile metal such as aluminum alloy is used as the top material [19,24,25]. The rivet penetrated through the top sheet of the bottom vibration-damping steel, but not into the bottom sheet of the vibration-damping steel. Hence, the top steel sheet of the vibration-damping steel actually sustained the mechanical  Figure 10 shows the cross-sectional views of SPR joints of aluminum alloy (top) to vibration-damping steel (bottom) formed with different die types, as presented in Figure 3. Different from the reversed sheet material's stack configuration, the rivet shank in these SPR joints were fully filled with the stack materials, so no rivet cavity was formed, which is commonly reported when a ductile metal such as aluminum alloy is used as the top material [19,24,25]. The rivet penetrated through the top sheet of the bottom vibration-damping steel, but not into the bottom sheet of the vibration-damping steel. Hence, the top steel sheet of the vibration-damping steel actually sustained the mechanical interlocking for the joint as the bottom sheet. The bottom steel sheet of the vibration-damping steel increased the mechanical performance of the top steel sheet and made a bottom sealing. The two outer sheets of the vibration-damping steel were stuck to each other by the adhesive layer. The rivet head gap was hardly observed compared with the reversed stack configuration, regardless of die type.  Figure 11 shows the measured geometrical indices of the SPR joints for this material stack with respect to the SPR die type. The results show that the stack of aluminum alloy (top) to vibrationdamping steel (bottom) had more interlock width than the reverse stack for all types of dies. The rivet shank cavity was fully filled with the stack material and the rivet head gap was tightly sealed because the top sheet aluminum alloy had a better formability than the top vibration-damping steel sheet for the reversed stack. As a result, more rivet shank penetrated into the bottom sheet, and thus less shank was entrapped by the top sheet. The type A1 die gave the largest interlock width, and the smallest interlock width was obtained with the type B die. The interlock widths of other joints were similar to each other, as shown in the figure. It was also observed, as is the case for the reverse material stack, that for dies with a flat-bottom shape (types A1, A2, and A3), the increase of the die taper angle, as well as die diameter, reduced the interlock, since the number of constraints on the bottom sheet reduced with increasing die cavity. The bottom thickness had a contrary trend with the interlock width for dies with a flat-bottom shape (types A1, A2, and A3); however, a clear trend of the bottom thickness was not found for the non-flat die (types B and C), whose trend was also observed in the case of the reverse material stack configuration. The tensile-shear test results for the SPR joints of aluminum alloy (top) and vibration-damping  Figure 11 shows the measured geometrical indices of the SPR joints for this material stack with respect to the SPR die type. The results show that the stack of aluminum alloy (top) to vibration-damping steel (bottom) had more interlock width than the reverse stack for all types of dies. The rivet shank cavity was fully filled with the stack material and the rivet head gap was tightly sealed because the top sheet aluminum alloy had a better formability than the top vibration-damping steel sheet for the reversed stack. As a result, more rivet shank penetrated into the bottom sheet, and thus less shank was entrapped by the top sheet. The type A1 die gave the largest interlock width, and the smallest interlock width was obtained with the type B die. The interlock widths of other joints were similar to each other, as shown in the figure. It was also observed, as is the case for the reverse material stack, that for dies with a flat-bottom shape (types A1, A2, and A3), the increase of the die taper angle, as well as die diameter, reduced the interlock, since the number of constraints on the bottom sheet reduced with increasing die cavity. The bottom thickness had a contrary trend with the interlock width for dies with a flat-bottom shape (types A1, A2, and A3); however, a clear trend of the bottom thickness was not found for the non-flat die (types B and C), whose trend was also observed in the case of the reverse material stack configuration. that for dies with a flat-bottom shape (types A1, A2, and A3), the increase of the die taper angle, as well as die diameter, reduced the interlock, since the number of constraints on the bottom sheet reduced with increasing die cavity. The bottom thickness had a contrary trend with the interlock width for dies with a flat-bottom shape (types A1, A2, and A3); however, a clear trend of the bottom thickness was not found for the non-flat die (types B and C), whose trend was also observed in the case of the reverse material stack configuration. The tensile-shear test results for the SPR joints of aluminum alloy (top) and vibration-damping steel (bottom) are presented in Figure 12. The results show that this material stack configuration had a greater tensile shear load than the reversed stack for all types of die. This is attributable to the overall increase of the interlock width when the vibration-damping steel was used as a bottom sheet. As for the die type dependence, the difference in the tensile-shear load of specimens was insignificant The tensile-shear test results for the SPR joints of aluminum alloy (top) and vibration-damping steel (bottom) are presented in Figure 12. The results show that this material stack configuration had a greater tensile shear load than the reversed stack for all types of die. This is attributable to the overall increase of the interlock width when the vibration-damping steel was used as a bottom sheet. As for the die type dependence, the difference in the tensile-shear load of specimens was insignificant for this stack (the difference in measured tensile-shear load was 0.34 kN) compared to that of the reversed stack (the difference in measured tensile-shear load was 2.52 kN), as presented in Figure 8. This indicates that the mechanical performance of the joint using vibration-damping steel as a bottom sheet was less affected by die type than the joint using it as a top sheet. The joint made from type A1 had the highest tensile-shear load, which was the same as that of the reversed stack. For the joint configuration using vibration-damping steel as a bottom sheet, the second highest tensile-shear load was obtained with the type C die. This could be ascribed to the second highest value of the interlock width of the joint.

SPR Joint of Aluminum Alloy (Top) and Vibration-Damping Steel (Bottom)
Appl. Sci. 2019, 9, x FOR PEER REVIEW 9 of 13 for this stack (the difference in measured tensile-shear load was 0.34 kN) compared to that of the reversed stack (the difference in measured tensile-shear load was 2.52 kN), as presented in Figure 8. This indicates that the mechanical performance of the joint using vibration-damping steel as a bottom sheet was less affected by die type than the joint using it as a top sheet. The joint made from type A1 had the highest tensile-shear load, which was the same as that of the reversed stack. For the joint configuration using vibration-damping steel as a bottom sheet, the second highest tensile-shear load was obtained with the type C die. This could be ascribed to the second highest value of the interlock width of the joint. The failure images of the joints are shown in Figure 13. The rivet pull-out failure was observed for all specimens. As discussed in Section 3.1, these pull-out failures were due to the fact that the interlock width, which was much greater than that of the reversed stack, was still not enough to induce other failure modes such as top sheet tearing and rivet fracture. The failure images of the joints are shown in Figure 13. The rivet pull-out failure was observed for all specimens. As discussed in Section 3.1, these pull-out failures were due to the fact that the interlock width, which was much greater than that of the reversed stack, was still not enough to induce other failure modes such as top sheet tearing and rivet fracture.
with respect to SPR die type.
The failure images of the joints are shown in Figure 13. The rivet pull-out failure was observed for all specimens. As discussed in Section 3.1, these pull-out failures were due to the fact that the interlock width, which was much greater than that of the reversed stack, was still not enough to induce other failure modes such as top sheet tearing and rivet fracture.

Vibration-Damping Steel vs. SPFC590DP
In this study, the geometrical and mechanical characteristics of the SPR joint of vibrationdamping steel were compared with those of SPFC590DP (a skin material of the vibration-damping steel). The thickness of SPFC590DP was chosen to be 1.4 mm, which was the same as the sum of thicknesses (0.7 mm) of the outer sheets in the vibration-damping steel. For the comparison, SPR joints were formed with the type A1 die. In the cross-sectional images of the SPR joints in Figure 14, the stack of joining the vibration-damping steel to the aluminum alloy (Al5052-H32) as the bottom material had a smaller cavity and rivet head gap than the stack of joining SPFC590DP to the aluminum alloy. This was because the vibration-damping steel was more flexible than the solid Figure 13. Failure modes for SPR joint of Al5052-H32 (top) and vibration-damping steel (bottom) with respect to SPR die type. (a) Type A1 die, (b) type A2 die, (c) type A3 die, (d) type B die, and (e) type C die.

Vibration-Damping Steel vs. SPFC590DP
In this study, the geometrical and mechanical characteristics of the SPR joint of vibration-damping steel were compared with those of SPFC590DP (a skin material of the vibration-damping steel). The thickness of SPFC590DP was chosen to be 1.4 mm, which was the same as the sum of thicknesses (0.7 mm) of the outer sheets in the vibration-damping steel. For the comparison, SPR joints were formed with the type A1 die. In the cross-sectional images of the SPR joints in Figure 14, the stack of joining the vibration-damping steel to the aluminum alloy (Al5052-H32) as the bottom material had a smaller cavity and rivet head gap than the stack of joining SPFC590DP to the aluminum alloy. This was because the vibration-damping steel was more flexible than the solid SPFC590DP, because of the viscoelastic adhesive layer in the vibration-damping steel (refer to Figure 14a,b).
Appl. Sci. 2019, 9, x FOR PEER REVIEW 10 of 13 SPFC590DP, because of the viscoelastic adhesive layer in the vibration-damping steel (refer to Figure  14a,b). For the Al5052-H32 top-stack configuration, the rivet cavity is completely filled with ductile Al5052-H32 for both the SPR joints (refer to Figure 14c,d). However, slightly higher deformation of the vibration-damping steel is observed at the bottom-center of the joint due to the separation of the outer sheets of the vibration-damping steel.  Figure 15 shows the result of the measured geometrical indices for the SPR joints. It was found that the interlock width and bottom thickness of the SPR joint of the vibration-damping steel were almost the same as those of the joint of SPFC590DP when Al5052-H32 was used as a bottom sheet. However, when Al5052-H32 was used as a top sheet, both the interlock width and bottom thicknesses of the joint made from the vibration-damping steel were larger than those of the joint made from SPFC590DP. For the Al5052-H32 top-stack configuration, the rivet cavity is completely filled with ductile Al5052-H32 for both the SPR joints (refer to Figure 14c,d). However, slightly higher deformation of the vibration-damping steel is observed at the bottom-center of the joint due to the separation of the outer sheets of the vibration-damping steel. Figure 15 shows the result of the measured geometrical indices for the SPR joints. It was found that the interlock width and bottom thickness of the SPR joint of the vibration-damping steel were almost the same as those of the joint of SPFC590DP when Al5052-H32 was used as a bottom sheet. However, when Al5052-H32 was used as a top sheet, both the interlock width and bottom thicknesses of the joint made from the vibration-damping steel were larger than those of the joint made from SPFC590DP.
(top) and Al5052-H32 (bottom), (c) Al5052-H32 (top) and SPFC590DP (bottom), and (d) Al5052-H32 (top) and vibration-damping steel (bottom). Figure 15 shows the result of the measured geometrical indices for the SPR joints. It was found that the interlock width and bottom thickness of the SPR joint of the vibration-damping steel were almost the same as those of the joint of SPFC590DP when Al5052-H32 was used as a bottom sheet. However, when Al5052-H32 was used as a top sheet, both the interlock width and bottom thicknesses of the joint made from the vibration-damping steel were larger than those of the joint made from SPFC590DP. The tensile-shear loads of the SPR joints of the stack configurations are presented in Figure 16. For the Al5052-H32 bottom configuration, the tensile-shear load of the joint made from SPFC590DP was 22.6% larger than that made from the vibration-damping steel, despite the interlock width of the two joints being similar. This difference (~1 kN) in the tensile-shear load of the two joints was attributable to the difference in the intrinsic mechanical properties of the base materials. The flexural rigidity of the vibration-damping steel was less than SPFC590DP, so more bending of the vibrationdamping steel occurred than the SPFC590DP as the top sheet in the lap shear test, which led to inter- The tensile-shear loads of the SPR joints of the stack configurations are presented in Figure 16. For the Al5052-H32 bottom configuration, the tensile-shear load of the joint made from SPFC590DP was 22.6% larger than that made from the vibration-damping steel, despite the interlock width of the two joints being similar. This difference (~1 kN) in the tensile-shear load of the two joints was attributable to the difference in the intrinsic mechanical properties of the base materials. The flexural rigidity of the vibration-damping steel was less than SPFC590DP, so more bending of the vibration-damping steel occurred than the SPFC590DP as the top sheet in the lap shear test, which led to inter-locking failure in the earlier stage [19,24]. However, the tensile-shear load of the joint made from the vibration-damping steel was slightly larger than that of the joint made from SPFC590DP when Al5052-H32 was used as a top sheet. This was mainly due to the larger interlock width of the joint made from the vibration-damping steel.
Appl. Sci. 2019, 9, x FOR PEER REVIEW 11 of 13 locking failure in the earlier stage [19,24]. However, the tensile-shear load of the joint made from the vibration-damping steel was slightly larger than that of the joint made from SPFC590DP when Al5052-H32 was used as a top sheet. This was mainly due to the larger interlock width of the joint made from the vibration-damping steel. In this study, the mechanical performance of the SPR joint between the vibration-damping steel and Al5052-H32 was always improved when the thinner Al5052-H32 was used as a top sheet, due to the increase of the interlock width. However, for the SPR joint in the SPFC590DP and Al5052-H32, using the thicker SPFC590DP as a top sheet gave more tensile-shear load by 3.9%, despite the smaller interlock width.
The failure images of the joints are illustrated in Figure 17. According to the figure, the joints in SPFC590DP and Al5052-H32 had the preferential fracture of the Al5052-H32 sheet, instead of showing the rivet pull-out failure, regardless of the joint configuration. The high mechanical strength of the SPFC590DP gave more solid interlocking between SPFC590DP and rivet (high resistance to bending deformation caused by the tensile-shear load), which in turn caused more stress concentration at the interlocking region between Al5052-H32 and the rivet during the tensile-shear test. This stress concentration in the Al5052-H32, with a much lower mechanical strength (nearly In this study, the mechanical performance of the SPR joint between the vibration-damping steel and Al5052-H32 was always improved when the thinner Al5052-H32 was used as a top sheet, due to the increase of the interlock width. However, for the SPR joint in the SPFC590DP and Al5052-H32, using the thicker SPFC590DP as a top sheet gave more tensile-shear load by 3.9%, despite the smaller interlock width. The failure images of the joints are illustrated in Figure 17. According to the figure, the joints in SPFC590DP and Al5052-H32 had the preferential fracture of the Al5052-H32 sheet, instead of showing the rivet pull-out failure, regardless of the joint configuration. The high mechanical strength of the SPFC590DP gave more solid interlocking between SPFC590DP and rivet (high resistance to bending deformation caused by the tensile-shear load), which in turn caused more stress concentration at the interlocking region between Al5052-H32 and the rivet during the tensile-shear test. This stress concentration in the Al5052-H32, with a much lower mechanical strength (nearly three times smaller) than SPFC590DP, would eventually drive it to be preferentially fractured.
In this study, the mechanical performance of the SPR joint between the vibration-damping steel and Al5052-H32 was always improved when the thinner Al5052-H32 was used as a top sheet, due to the increase of the interlock width. However, for the SPR joint in the SPFC590DP and Al5052-H32, using the thicker SPFC590DP as a top sheet gave more tensile-shear load by 3.9%, despite the smaller interlock width.
The failure images of the joints are illustrated in Figure 17. According to the figure, the joints in SPFC590DP and Al5052-H32 had the preferential fracture of the Al5052-H32 sheet, instead of showing the rivet pull-out failure, regardless of the joint configuration. The high mechanical strength of the SPFC590DP gave more solid interlocking between SPFC590DP and rivet (high resistance to bending deformation caused by the tensile-shear load), which in turn caused more stress concentration at the interlocking region between Al5052-H32 and the rivet during the tensile-shear test. This stress concentration in the Al5052-H32, with a much lower mechanical strength (nearly three times smaller) than SPFC590DP, would eventually drive it to be preferentially fractured.

Conclusions
In this study, the SPR joining between dissimilar materials of vibration-damping steel, SPFC590DP, and Al5052-H32 was carried out, and the mechanical performance, failure mode, and geometrical features of the joints with different die types and material stack configurations were investigated. The important conclusions of this research are as follows: (1) The SPR joint between the vibration-damping steel and Al5052-H32 showed mechanical performance similar to that between SPFC590DP and Al5052-H32. For the Al5052-H32 bottom configuration, the SPR joint of the vibration-damping steel had 22.6% lower tensile-shear load than SPFC590DP. However, it had 3.9% higher tensile-shear load than SPFC590DP when Al5052-H32 was used as a top sheet. These results suggest that the SPR technique could be a promising solution for dissimilar joining between vibration-damping steel and aluminum. (2) The SPR joints of the vibration-damping steel and Al5052-H32 with the Al5052-H32-top configuration had a consistent superior mechanical performance, mainly due to the increase of the interlock width. In addition, due to the small interlock width, all SPR joints of the vibration-damping steel and Al5052-H32 showed consistent rivet pull-out failure, regardless of the joint configuration. (3) The SPR joints of the vibration-damping steel and Al5052-H32 showed the largest tensile shear load with a flat die (a benchmark die). For the flat die, the increase of die taper angle and diameter reduced the mechanical performance of the joint, owing to the increase of the die's volume, resulting in the smaller interlock width of the joint.