Friction Stir Spot Welding of Aluminum Alloy to Carbon Fiber-Reinforced Thermosetting Resin Coated by Thermoplastic Resin Using Tools with Different Surface Shapes

Abstract

To achieve carbon neutrality, a reduction in car body weight is essential. Multi-material structures that use lightweight materials such as carbon fiber-reinforced polymers (CFRP) and aluminum (Al) alloy are used to replace parts of steel components. This multi-material method requires specific joining techniques for bonding dissimilar materials. Friction stir spot welding (FSSW) is one of the joining techniques used for joining dissimilar materials, enabling rapid and strong joints. FSSW for bonding A5052 Al alloy and carbon fiber-reinforced thermosetting resin (CFRTS) utilizing composite laminates with integrally molded thermoplastic resin in the outermost layer has been developed. However, joints using this method cause pyrolysis due to excessive frictional heating at the tool’s bottom, which may affect joint strength and promote corrosion in Al alloy. Therefore, this study developed new tools, a concave-shaped tool without a probe, a concave-shaped tool with a probe and a conventional FSSW tool, and investigated the influence of heat distribution and joint strength using the three new tools. The newly developed concave-shaped tool with a probe suppressed 7% of maximum heat input, decreased the pyrolysis area of epoxy resin by 47%, and increased joint strength by 4%. Finite element analysis also showed the suppression of heat input through the newly developed concave-shaped tool with a probe, achieved by reducing the contact area between the tool and Al alloy.

1. Introduction

Achieving carbon neutrality is essential to mitigate climate change [1]. It has been reported that a decrease in automobile weight leads to a decrease in emissions [2]. According to Japan’s Ministry of Transport, decreasing the weight of a single vehicle by 100 kg can reduce CO₂ emissions by an average of 20 g/km [3]. Therefore, carbon fiber-reinforced polymers (CFRP) [4] and lightweight metals, such as aluminum alloy, magnesium alloy and titanium alloy, are expected to contribute to lightweight structure design [5]. Nonetheless, these materials, especially CFRP, are costly compared to steel, limiting their application in mass-produced automobiles [4]. Consequently, the concept of multi-material design, which uses lightweight materials and steel in optimal locations, is expanding [6]. Optimal material selection can impact both cost and weight reduction. In order to implement this concept, effective adhesive and joining techniques for dissimilar materials are required.

Various joining techniques, such as mechanical joining and adhesive bonding, have been developed [7]. However, each technique presents certain drawbacks. For instance, mechanical joining induces stress concentration at the joint, which can negatively impact fatigue resistance, reduce material lifespan and increase weight due to additional materials such as bolts and rivets [7]. Furthermore, adhesive bonding is time-consuming [8]. Additionally, carbon fiber-reinforced thermosetting resin (CFRTS) with a thermosetting matrix resin cannot be bonded without modification since the matrix resin forms an irreversible crosslinked structure.

Friction stir spot welding (FSSW) is a solid-state welding process [9]. In the FSSW process for metal-to-metal joining, a rotating tool is inserted into the overlapping metal sheets. This action facilitates the flow of material both downward and outward from the tool shoulder, leading to metal–metal bonding [9]. On the other hand, in the metal-to-resin FSSW process, the frictional heat generated at the interface between the tool and the metal is utilized to raise the thermoplastic resin to its melting point [9]. FSSW is also applied for joining carbon fiber-reinforced thermoplastics (CFRTP) and metals, where the thermoplastics melt and facilitate the bond [10,11].

We developed FSSW for the joining of A5052 aluminum alloy with CFRTS composites that have an integrally molded thermoplastic resin in the outermost layer [12]. In this method, the thermoplastic resin acts as an adhesive, welding Al alloy and CFRTS [12]. This method has demonstrated a shear strength of 12.1 ± 0.6 kN, which exceeded the JIS standard value for steel spot welding [12]. However, joints formed using this method cause pyrolysis due to excessive frictional heating at the tool’s bottom [12]. This excessive heat leads to the decomposition of thermoplastic PA12 and epoxy resin pyrolysis under the welding tool [12]. Consequently, the joint strength is maintained at the outer area of the pyrolysis area [12] and galvanic corrosion of Al may occur due to direct contact between carbon fiber and Al [13]. Therefore, preventing excessive heating at the tool’s bottom is crucial, and it may improve shear strength. We introduced a multi-stage heating process, which included cooling periods as a method to manage heat input [14]. This research demonstrated that multi-stage heating with cooling intervals reduces excessive heating under the tool, leading to improved joint strength [14]. However, we reported that the pyrolysis area of resin remained visible on the fracture surface after the tensile test [14]. Additionally, Geng et al. improved the joint strength between 6061-T6 aluminum alloy and carbon fiber-reinforced polyamide-6 (CF/PA6) by modifying the tool shape [15]. They developed concave-shaped and ring-shaped tools, and compared them to a flat probe-less tool [15]. However, no research has been made in comparing it to an FSSW conventional tool which has a probe on top. Furthermore, FSSW joints between Al alloy and CFRTP composites potentially damage CFRTP composites since joints are achieved by melting their own matrix resin. This can be avoided by using CFRTS composites coated by thermoplastic resin as the other materials, where thermoplastic resin melts without CFRTS composite damage.

In this research, newly developed tools with different surface shapes will be compared to a FSSW conventional tool. A5052 aluminum (Al) alloy and CFRTS composites with integrally molded thermoplastic resin in the outermost layer will be joined using a multi-stage heating FSSW to minimize excessive heat input.

CFRP/Al joints are prepared using three tools: a newly developed concave-shaped tool without a probe, a newly developed concave-shaped tool with a probe and a conventional tool. With the conventional tool, the Al alloy is pressed against a flat circular surface of the tool, potentially generating excessive heat. In contrast, the newly designed tool contacts only the ring surface instead of the entire flat circular surface of the tool, possibly reducing excessive heat at the tool’s bottom. The temperature distribution during welding was observed using the IR thermography camera. Image analysis of the IR thermography data was used to calculate the PA12 molten area, epoxy resin pyrolysis area, and the maximum temperature input during FSSW. Additionally, the interfacial structure of joints was observed using a scanning electron microscope and X-ray observation, and bonding strength and fracture surface of the joints were obtained through tensile shear test. Based on the combined evaluation of thermal process and mechanical characteristics, the most suitable tool for FSSW of Al and CFRP will be determined. Finite element analysis (FEA) is also conducted to assess optimal tool shapes by analyzing temperature distribution during FSSW.

2. Materials and Methods

2.1. Materials

CFRP laminates with a thickness of 1.2 mm, a length of 100 mm, and width of 30 mm were molded from five sheets of CF/ epoxy prepreg (F6343B-05P, 0.24 mm thick, Vf = 55%, Toray Industries, Inc., Tokyo, Japan) with two sheets of polyamide film (PA12, 3014U, 0.1 mm thick, 179 °C M. P., UBE Corp., Tokyo, Japan) stacked on the upper and lower surfaces (Figure 1). CFRP laminates were formed by curing epoxy resin after melting PA12 film using a molding machine (STIP05-05, Sato Machinery Works Co., Ltd., Aichi, Japan) at a molding pressure of 5.7 MPa. The temperature history for curing was chosen based on previous research [12], as shown in Figure 2.

An A5052 Al alloy sheet with a thickness of 2 mm, a length of 100 mm, and a width of 30 mm was used as the other materials. Prior to the welding experiments, the surface of the A5052 Al alloy was subjected to electropolishing and phosphoric acid anodizing treatment (Figure 3a). Electropolishing was carried out for 1 min in a 1 mol/L hydrochloric acid solution at 50 °C and with a current density of 100 mA/ cm³ (Figure 3c). Phosphoric acid anodizing was performed at room temperature (25 °C) for 25 min, using a 100 g/L phosphoric acid solution and an applied voltage of 11.5 V (Figure 3d).

2.2. Welding Procedure

FSSW was conducted using a precision universal testing machine (Autograph AG-250 kN, Shimadzu Co., Ltd., Kyoto, Japan) equipped with a servo motor (1.0 kW, MHMF092L1U2, Panasonic Industry, Tokyo, Japan) (Figure 4a). A handmade jig was connected to the servo motor and a tool [14] (Figure 4b). The upper Al alloy sheet was overlapped on a 30 × 30 mm area of the CFRP laminates. The specimens were fixed to a glass epoxy jig during the welding process, as shown in Figure 4c. CFRP laminates and the Al alloy were welded using three different tools: Tool0, a conventional tool with a probe (Figure 5a); Tool1, a newly developed concave-shaped tool without a probe (Figure 5b); and Tool2, a newly developed concave-shaped tool with a probe (Figure 5c). These tools were manufactured from SKD61. Contact friction between the Al alloy and tool, as well as plastic deformation of the Al alloy, are the main heat sources in the FSSW process. Considering heat generated due to contact friction Q, it can be expressed by the following equation [16].

$$dQ=\left(P\times dA\times \mu \right)\times \left(\omega \times r\right)$$

(1)

where P [N] is the pressure, A [mm²] is the contact area, μ is the friction coefficient, ω [rad/s] is the angular velocity, and r [mm] is the radius [16]. Theoretically, the outermost region of the tool generates the maximum temperature according to Equation (1). However, previous research has indicated the maximum heat in a flat circular tool without a probe is generated at the 3/5 radius region of the rotating tool during FSSW [15]. This is because the outermost region of the tool can transfer heat to the Al alloy; in contrast, heat tends to remain at the 3/5 radius region of the tool. Therefore, our newly developed tools are expected to transfer heat to the Al alloy by inducing a concave shape, allowing heat to escape more effectively.

The tool rotation speed was set to 3000 rpm, the tool plunge speed was set to 0.6 mm/s and the plunge depth was set to 0.6 mm. Multi-stage heating was induced as a tool depth condition, as described in Figure 6. The multi-stage heating included a total dwell time of 4 s, divided into 0.5 s and 3.5 s and a cooling time of 2.0 s between the two dwell periods at a tool depth of 0 mm.

After achieving joints, a section of the joint was cut through the center of the tool, as shown in Figure 7, and polished using a cross-section polisher (SM-09010, JEOL, Tokyo, Japan). The joint section, obtained from three different tools, was observed using a scanning electron microscope (SEM, SU8020, Hitachi High-Tech Co., Ltd., Tokyo, Japan).

X-ray submicron CT system (nano 3DX, Rigaku Corporation, Tokyo, Japan) was used to observe the overlap region of CFRP/Al joints. The joint area under the center of the tool and right side of the tool with geometry and dimensions of 3.6 mm × 3.6 mm × 2.8 mm were observed, as illustrated in Figure 8. After X-ray observation, the void ratio of CFRP laminates within the dimensions of 1 mm × 1 mm × 0.2 mm, calculated from the top surface of CFRP laminates, was calculated using ImageJ (version: 1.54i) [17].

2.3. Observation of the Temperature Distribution During FSSW

An infrared thermography camera (InfReC R300SR, Nippon Avionics Co., Ltd., Kanagawa, Japan) was used to observe the temperature distribution at the interface between the Al alloy and CFRP laminates during the welding process. To facilitate the observation of temperature distribution, a quarter of the overlapped area (15 × 15 mm) of the CFRP laminates was cut away, and black color paint was applied to Al’s back surface (Figure 9). A 40 mm × 40 mm silver-plated copper plate was used as a reflector and placed under the observation area at a 45° angle to refract infrared rays. From the reflected images, the temperature distribution was captured, and the area percentages of melting and decomposition within the observed area and maximum temperature during FSSW were calculated using MATLAB (version: 9.14.0 (R2023a), The MathWorks, Inc., Natick, MA, USA). Infrared images were captured with a sampling time of 0.1 s. The maximum temperature was obtained from each infrared image as the highest temperature observed within the observation area at each 0.1 s of infrared image. Based on previous research, the melting temperature of PA12 resin and the pyrolysis temperature of epoxy resin was defined at 178 °C and 350 °C, respectively [14]. The molten region is defined as the area where the temperature exceeds 178 °C, whereas the pyrolysis region corresponds to the area where the temperature exceeds 350 °C. By counting the pixel number ratio within an observed area, we evaluated the molten area percentage of PA12 resin and the pyrolysis area percentage of epoxy resin within the observation area.

2.4. Evaluation of Mechanical Properties of Joints

To evaluate the joint strength, a static tensile shear test was performed. As shown in Figure 10, Al tabs with thicknesses of 1.2 mm and 2.0 mm were bonded to the ends of each specimen after FSSW to align the load axis in the direction of the plate thickness.

The tensile test was conducted using a precision universal testing machine (Autograph AG-100 kN Xplus, maximum load capacity 100 kN, Shimadzu Co., Ltd., Kyoto, Japan) at a displacement rate of 0.167 mm/min.

After the tensile test, the fracture surfaces were examined using a digital microscope (VHX-5000, Keyence Co., Ltd., Osaka, Japan), as illustrated in Figure 11.

2.5. Finite Element Analysis (FEA) of Temperature Distribution During FSSW

Finite element analysis (FEA) was conducted using the COMSOL Multiphysics (Stockholm, Sweden) simulation software. The temperature distribution between Tool0 and Tool2 during the initial dwell time of 0.5 s was compared in this simulation. In this research, Tool2 was shown to suppress heat efficiently among the three tools. Therefore, Tool2 will be compared to Tool0, the conventional FSSW tool. The heat transfer process was governed by the Fourier law of heat conduction, as described by the following equation.

$$\rho C_p{\displaystyle \frac{\mathsf{\partial }T}{\mathsf{\partial }t}}=\nabla ·\left(k\nabla T\right)+Q_{shoulder}+Q_{pin}$$

(2)

where $k$ represents the thermal conductivity, $\rho$ is the density, $C_p$ is the specific heat capacity, $Q_{shoulder}$ denotes the heat from tool shoulder (red line) and $Q_{pin}$ refers to the heat from tool pin (green line), as indicated in Figure 12.

During FSSW, the rotating tool contacts the Al alloy at the two surfaces: the tool’s shoulder ($Q_{shoulder})$ and the tool’s pin ($Q_{pin})$. Heat is generated at these both interfaces and transferred to the Al alloy. These heat sources are expressed by the following equations [16].

$$Q_{shoulder}=\left({\displaystyle \frac{\mu {\left(T\right)F}_n}{A_s}}\right)\omega r$$

(3)

$$Q_{pin}={\displaystyle \frac{\mu \left(T\right)}{\sqrt{3\left(1+{\mu }^2\right)}}}{r}_p\omega \overline{Y\left(T\right)}$$

(4)

where $\mu \left(T\right)$ represents the temperature dependence of friction coefficient, $F_n$ is the applied load, $A_s$ is the shoulder area, $\omega$ is the angular velocity, $r$ is the radius, $r_p$ is the pin radius, $\overline{Y\left(T\right)}$ denotes the average shear stress of the Al alloy. To achieve an accurate simulation, temperature dependence of the friction coefficient should be considered [18]. Therefore, in this study, the friction coefficient was expressed by the following equations for different temperature ranges [19].

$$\mu \left(T\right)=0.18+2.27\times {10}^{-3}T\left(0°\mathrm{C}<T\le 150°\mathrm{C}\right)$$

(5)

$$\mu \left(T\right)=0.9-2.54\times {10}^{-3}T\left(150°\mathrm{C}<T\le 280°\mathrm{C}\right)$$

(6)

$$\mu \left(T\right)=0.36-5.94\times {10}^{-3}T\left(280°\mathrm{C}<T\le 600°\mathrm{C}\right)$$

(7)

Based on the literature, the average shear stress is defined as the yield stress of the Al alloy with temperature dependence, as illustrated in Figure 13 [20].

The upper and lower surfaces of the Al alloy and CFRP laminates lose heat through natural convection and surface-to-ambient radiation, except in the overlapping region between the Al alloy and CFRP laminates. These heat fluxes are expressed by the following equations [21].

$$q_{Al}=h_{Al}\left(T_0-T\right)+\epsilon \sigma \left({T_{amb}}^4-T^4\right)$$

(8)

$$q_{CFRP}=h_{CFRP}\left(T_0-T\right)+\epsilon \sigma \left({T_{amb}}^4-T^4\right)$$

(9)

where $h_{Al}$ and $h_{CFRP}$ represent the heat transfer coefficients for natural convection, $T_0$ is an associated reference temperature, $\epsilon$ is the surface emissivity, $\sigma$ is the Stefan–Boltzmann constant, and $T_{amb}$ denotes the ambient air temperature. A surface-to-surface algorithm was applied to the heat interaction between the Al alloy and CFRP. The specific values used in this simulation are listed in

Table 1. Input data for the simulation [22].

Density		Heat Transfer Coefficient		Rotational Speed	Applied Force	Ambient Air Temperature	Surface Emissivity
Al Alloy	CFRP Laminates	Al Alloy	CFRP Laminates
2700 kg/m3	1225 kg/m3	70 W/(m2 K)	20 W/(m2 K)	3000 rpm	3000 N	293 K	0.3

and temperature dependence thermo-physical properties are listed in

Table 2. Thermo-physical properties of Al alloy and CFRP laminates [22].

Parameters	Materials	20 °C	100 °C	200 °C	300 °C	400 °C	500 °C	600 °C
Thermal conductivity (W/(m K))	Al alloy	195	200	206	208	206	199	195
	CFRP laminates	0.427	0.435	0.512	-	-	-	-
Specific heat (J/(kg K))	Al alloy	900	950	1000	1020	1050	1090	1200
	CFRP laminates	1310	1950	2860	-	-	-	-
Expansion (10−6/K)	Al alloy	22	24.6	26.6	27.8	3.01	-	-
	CFRP laminates	5.0	4.9	3.1	-	-	-	-

. These numbers were referenced from the previous research [22].

3. Results and Discussion

3.1. Results of the Temperature Distribution Observation

Examples of maximum temperature profiles during FSSW are shown in Figure 14. Tool2 exhibits lower heat input during welding compared to Tool0 and Tool1. Additionally, the maximum temperature was observed during the initial dwell time at 1 s–1.5 s after the start of welding. These findings also can be seen in Figure 15, presenting examples of temperature distribution images obtained using three tools at six time points during FSSW. In this study, we specifically focus on the time points of 0.8 s and 1.5 s, because the influence of tool modification on the temperature distribution for the three tools becomes apparent at these specific times. At 0.8 s, only the probe contacts the Al alloy for Tool0 and Tool2, and only the ring surface contacts the Al alloy for Tool1. At 1.5 s, the temperature distribution corresponds to the middle of the initial dwell time, where the maximum heat input occurs. Figure 16 provides schematic drawings of tool movement during FSSW up to 1.5 s after welding started. As noted, the maximum temperature occurs during this initial dwell time and Tool2 has the lowest temperature of 354 °C among the three tools, compared to 425 °C for Tool0 and 456 °C for Tool1. A high-temperature area is observed at the bottom of each tool. Tool0 shows a concentrated high temperature at the flat circular surface, Tool1 at the ring surface, and Tool2 at the ring surface with slight spreading towards the center of the tool. Additionally, the infrared image at 0.8 s shows that a probe inputs initial suppressing heat for Tool0 and Tool2 at 123 °C and 117 °C, respectively, while Tool1 has a large heat input of 213 °C at 0.8 s due to the fact that only the ring surface contacts the Al alloy at this stage. These temperature patterns correspond to the unique surface shapes of each tool. As illustrated in Figure 16, since Tool2 has the 0.5 mm depth probe, only the probe area contacts with the Al alloy up to 0.8 s after welding starts. After 0.8 s, the ring surface of Tool2 makes contact, penetrating only 0.1 mm into the Al alloy. In contrast, although Tool0 has the 0.5 mm depth probe, its flat circular surface contacts the Al alloy after 0.8 s instead of the ring surface. As reported by Geng et al., a flat circular surface shows a higher temperature input compared to a ring surface during FSSW [15]. By modifying the tool’s surface into a concave shape, the contact area between the tool and Al alloy is reduced, and heat can transfer to Al alloy more effectively, thereby decreasing the overall heat input. Tool1 exhibited excessive heat input since there is no probe in Tool1 and only the ring surface contacts the Al alloy, penetrating 0.6 mm into the Al alloy, as shown in Figure 16. Previous research has indicated that increasing the tool’s plunge depth results in an increased heat input [22]. For Tool1, because the ring surface penetrates 0.6 mm due to the absence of a probe, large heat was generated during FSSW. Therefore, Tool2’s probe can induce an initial heat while suppressing the heat in contrast with Tool1’s initial heat due to its concave shape. Subsequently, the ring surface of Tool2 can spread this heat while suppressing it compared to the flat circular surface of Tool0. These observations indicate that the modified surface shape of Tool2 effectively reduces heat input. Figure 17 shows the maximum temperature input during FSSW obtained from the three tools. The maximum temperature input for Tool2 was the lowest value of 380 ± 17 °C (mean ± S.D.), showing a significant difference (p < 0.05) compared to Tool0 and Tool1. Specifically, Tool2 demonstrated a 7% reduction in heat input compared to Tool0 and 17% compared to Tool1.

Figure 18 and Figure 19 show the molten area percentage of PA12 resin and pyrolysis area percentage of epoxy resin during welding, respectively, as calculated from the temperature distribution obtained using the three tools. Tool2 provides a sufficient molten area of PA12, with no significant difference observed among the three tools. For the pyrolysis area of epoxy resin, Tool2 exhibited the lowest value of 2 ± 1% (mean ± S.D.), showing a significant difference (p < 0.05) compared to Tool1, with a 68% reduction. Although there is no significant difference between Tool0 and Tool2, Tool2 achieved a 47% reduction in the pyrolysis area on average compared to Tool0. This reduction is attributed to the suppressed heat input using Tool2. Due to this reduction in heat input, Tool2 shows the smallest pyrolysis area of epoxy resin. These results, derived from the observation of temperature distribution, demonstrate that Tool2 effectively suppresses heat input, leading to the reduction in the pyrolysis area of epoxy resin during FSSW.

3.2. Tensile Shear Test

Figure 20a presents the maximum load from the tensile shear test for the three tools and the welding area calculated from the observation of the fracture surface after the tensile shear test. The welded area was defined as the region where PA12 resin adhered to the Al alloy. Figure 20b provides an example of a welded area, outlined in red, observed on the CFRP laminates after welding. The discolored resin area is visible in the center region, indicating resin pyrolysis; therefore, this region was not considered as a welded area. Tool2 achieved the largest welded area of 882 ± 26 mm² (mean ± S.D.), being significantly larger (p < 0.05) compared to Tool0 and Tool1, with an increase of 12% and 9%, respectively. These results correspond to the molten and pyrolysis areas in Figure 18 and Figure 19, which were obtained through the observation of temperature distribution with a thermography camera. Figure 21 shows the joint strength for the three tools, which is calculated from the maximum load divided by the welded area. Tool2 achieved the highest joint strength value of 16 ± 0.6 MPa (mean ± S.D.), which is a significant difference (p < 0.05) compared to Tool0 and Tool1. Using Tool2, joint strength can be increased by 4% compared to Tool0 and 7% compared to Tool1. These findings can be explained by the reduction in the pyrolysis area of epoxy resin using Tool2. By using Tool2, the largest welded area was obtained, resulting in the highest joint strength among the three tools.

Figure 22 shows the fracture surfaces after the tensile shear test. Fracture surfaces obtained from Tool0 and Tool1 exhibit discolored resin areas, indicated by red arrows. These discolored resin areas are considered to be the pyrolysis areas of resin caused by excessive heat input. In FSSW, where a thermoplastic resin acts as an adhesive and is subsequently melted to facilitate bonding, it has been reported that under the tool, the epoxy resin matrix of CFRP and PA12 adhesive resin undergoes thermal decomposition due to an excessive heat input, and the joint is maintained at the outer area of pyrolysis area of resin [23]. Although both Tool0 and Tool1 show these pyrolysis resin areas, Tool2 displays the minimal presence of such areas. This result also suggests that Tool2 effectively suppresses heat input, resulting in a reduction in the pyrolysis area of the resin and an increase in the welded area.

This trend can also be observed in the SEM images of joint cross-sections under the tool’s shoulder. As reported by Akbari et al., most of the heat during FSSW is generated by the tool shoulder rather than the probe [24]. Therefore, in this study, we focused on the area under the tool shoulder to assess the influence of tool modification by examining the region that is theoretically most affected by the heat input during FSSW. As shown in Figure 23, this cross-sectional area includes the Al alloy on the top part, while the rest is composed of CFRP laminates. Although joints formed using Tool0 and Tool1, show voids in CFRP laminates, outlined by red boxes, no such void areas were observed in the joints formed by Tool2. Previous research demonstrated that higher temperature input induces voids at greater depth regions in the thickness direction of CFRP laminates [14]. Therefore, the voids observed in Tool0 and Tool1 joints in this study are also thought to be due to the high-temperature input. These results further indicate that heat input can be suppressed using Tool2 during FSSW, minimizing voids in CFRP laminates.

The 3D images obtained from the X-ray observation are shown in Figure 24. Voids inside CFRP laminates and CFRP laminates are illustrated in red and yellow color, respectively. Joints achieved by Tool2 have fewer voids inside CFRP laminates both under the tool and at the right side of the tool. In contrast, many voids can be seen in CFRP laminates for Tool0 and Tool1. These voids are thought to be caused by excessive heat input during FSSW. These findings also demonstrate that Tool2 effectively suppresses heat input. Moreover, voids are observed in specific areas according to the unique shapes of Tool0 and Tool1. Tool0 has many voids both under the center of the tool and at the right side of the tool due to the flat circular surface and probe shape, whereas Tool1 has many voids under the right side of the tool due to its ring shape.

Table 3. Void ratio of CFRP laminates.

Tool Type	Tool0		Tool1		Tool2
Calculated Region	Center	Right	Center	Right	Center	Right
Void ratio	31.9%	6.7%	6.6%	7.7%	0.2%	0.1%

lists the void ratio of CFRP laminates within the observation region of 1 mm × 1 mm × 0.2 mm. These results also correspond to the Tool’s unique shape. Although joints achieved by Tool0 and Tool1 show a high void ratio in both the center and at the right side of CFRP laminates due to their shape and excessive heat input during FSSW, Tool2 has low void ratio compared to Tool0 and Tool1 and shows almost 0% void ratio within the observation region.

3.3. Finite Element Analysis (FEA) of Temperature Distribution

In order to validate the FEA model developed in this study, the predicted temperature history along the defined line on Al’s back side, as illustrated in Figure 25, was compared with the experimental data obtained from the observation of temperature distribution using an infrared camera. Figure 26 shows the comparison of temperature histories between FEA predictions and experimental results for Tool0 and Tool2. Although the FEA model predicts a higher temperature input compared to the experiment results, the trend of temperature history for the two tools is in good agreement between FEA and experiment results. This discrepancy can be explained by the fact that temperature observation was performed by removing away a quarter of the overlapping area of CFRP laminates to capture infrared images, which might cause slight heat loss from this exposed area. Thus, the actual temperature generated during FSSW is likely higher than the observed values. Based on these results, the FEA model developed in this study is considered to be able to predict the FSSW phenomenon and generate frictional heat.

Using this FEA model, the temperature distribution of Tool0 and Tool2 was simulated and the results are shown in Figure 27 and Figure 28. Figure 27 shows the temperature distribution on the surface of the Al alloy, while Figure 28 presents the distribution across the cross-section area between the Al alloy and CFRP laminates. As shown in Figure 27, Tool0 spreads heat from the probe and the flat circular surface (Figure 27a), whereas Tool2 distributes heat from the probe and the ring surface (Figure 27b). This trend corresponds to the temperature distribution observed using an infrared camera, as shown in Figure 15, where the higher temperatures occur under each tool’s unique surface shapes. Similarly, Figure 28 shows that while Tool0 transfers heat into Al alloy and CFRP laminates from the probe and flat circular interface (Figure 28a), Tool2 spreads from the probe and concave-shaped interface (Figure 28b). These FEA results demonstrate that Tool2 can input more effectively while suppressing excessive heat compared to Tool0 due to its concave-shaped surface, which reduces the contact area between the tool and Al alloy during FSSW.

Figure 29 shows temperature histories at the interface between the Al alloy and CFRP laminates. Evaluated interfaces are also shown in Figure 29a, illustrated by a red line. Tool0 increases in temperature more rapidly than Tool2 (Figure 29b), while Tool2 shows a gradual temperature increase due to its concave-shaped surface (Figure 29c). As reported by Ma et al., FEA demonstrated that a concave-shaped tool successfully suppresses heat input compared to a flat surface tool by preventing full contact between the friction surface and the Al alloy during FSSW [25]. In this study, Tool2 also suppresses heat input by reducing the contact area between the tool and Al alloy, achieved by modifying the flat circular surface of Tool0 into a concave-shaped surface.

This simulation model includes some simplification. For instance, tool movement such as rotation and vertical motion are not considered. As indicated in the methods section, this model assumes that the tool has already been inserted into the Al alloy and that heat is generated from the tool interface. Therefore, the actual movement of Tool0 and Tool2 during FSSW, where the probe contacts the Al alloy first, followed by the flat circular or ring surface at 0.8 s after the start of welding, is not considered. Moreover, this FEA model only simulates heat distribution during the first dwell time of 0.5 s for Tool0 and Tool2. The model also does not account for the stirring process in the Al alloy, which is highly complex due to phase changes and material flow around the rotating tool. Although the model developed in this study allows observation of the difference in heat distribution depending on the tool surface shapes, a new FEA model that considers the actual FSSW process could be developed in future research for a deeper understanding of the heat transfer phenomenon. Moreover, the optimization of tool geometry can be considered, especially the ring surface area of Tool2 can be reduced since most heat is generated from the tool shoulder. A reduction in the contact area between the tool, where the most heat is generated, and the Al alloy can expect a reduction in heat input during FSSW. In this research, FSSW was performed between the Al alloy and CFRTS composites coated by thermoplastic resin. However, CFRTP laminates are becoming increasingly popular due to their recyclability and short productive time. Thus, developing an FSSW method for Al alloy and CFRTP laminates without damaging the CFRTP material is critical. We anticipate that this could be achieved using our newly developed Tool2 shape.

4. Conclusions

To suppress excessive heat input under the tool during friction stir spot welding (FSSW), this study introduced newly developed tools used for FSSW. The effect of modifying tool shape on the static tensile shear strength of the joint between the Al alloy and CFRP laminates with integrally molded PA12 resin in the outermost layer was investigated. Additionally, temperature distribution during FSSW was observed using an infrared thermograph camera to determine the maximum temperature input, as well as the molten and pyrolysis resin areas. Finite element analysis (FEA) was also conducted to compare the difference in temperature distribution between the conventional tool and newly developed tools. Our study yields the following conclusions:

• Tool2, the newly developed concave-shaped tool with a probe effectively suppressed heat input compared to Tool0, the conventional tool with a probe, and Tool1, the newly developed concave-shaped tool without a probe. Tool2’s unique features of the probe induce a small initial heat input, and the ring surface spreads heat while suppressing it. Using Tool2, the maximum temperature input during FSSW and the pyrolysis area of resin can be reduced by 7% and 47%, respectively, compared to Tool0.
• The highest joint strength was achieved using Tool2, with a 4% improvement compared to Tool0. By reducing the pyrolysis area of resin, Tool2 led to a 12% increase in the welded area compared to Tool0.
• Finite element analysis (FEA) demonstrated the suppression of heat input using Tool2. While heat spreads from the probe and flat circular interface for Tool0, Tool2 distributes heat from the probe and concave-shaped interface. FEA also showed a reduced temperature rise at the interface between the Al alloy and CFRP laminates using Tool2 due to its concave shape.