Friction stir lap joining techniques effects on microstructure and tensile properties of high-strength automotive steel top hat sections

Dual Phase (DP) steel, a type of Advanced High Strength Steel (AHSS) with a thickness of 1.7 mm, is used to fabricate single-hat components that are then joined to the base plate using two friction stir welding processes: friction stir lap welding (FSLW) and friction stir spot lap welding (FSSLW). It is difficult to join this assembly using fusion welding techniques. The welding variables for the FSLW process, tool rotation speed (TRS), tool traverse speed (TTS), and plunge depth (PD), were optimized using the design of an experiments-based response surface method by experimentally measured tensile shear failure load (TFL) of top hat assembly. For the FSSLW process, the welding variable TTS was replaced by dwell time (DT). Peak temperature, microstructure at different zones, microhardness mapping, and energy absorption capacity of both processes were evaluated under optimal welding conditions. For both processes, the stir zone and the heat-affected zone had the highest and lowest microhardness, which can be correlated with the level of martensite tempering, martensite lath spacing, polygonal ferrite volume, and precipitated carbides. Under optimum welding conditions, the TSL and energy absorption of FSLW joints were 14 kN and 170 J, respectively, which is 20% and 47 higher than the TSL and energy absorption of FSSLW joints.


Introduction
The adoption of advanced high-strength steels (AHSS) for the construction of automobiles has steadily increased in recent years due to their higher strength (i.e., to have improved passenger safety) and lighter weight ( i.e., for improved fuel economy) [1,2]. These materials are the frontal and rear side energy-absorbing members, which absorb most energy during crash collisions [3]. Besides the many advantages of AHSS for car body construction, the microstructural changes due to welding lead to brittleness and reduced formability. Resistance spot welding (RSW) has been used for several years in car manufacturing due to its utility and economy. Still, this process suffers from disadvantages such as high-stress concentration, spatters, blowholes, brittle microstructure, electrode wear, hydrogen-induced cracking, softer heat-affected zone (HAZ), and higher residual stresses [4,5]. Heat treatment is required to solve these issues, but it is time-consuming and costly due to decreased productivity.
Laser beam welding is also a popular process in car manufacturing; however, they exhibited poor joint integrity due to the brittle martensitic microstructure at HAZ caused by the higher cooling rates [6,7]. Hence an equivalent solution to this problem is using a solid-state welding process, where the heat input and cooling rates can be controlled to obtain the desired properties and microstructure [8]. Such a solid-state procedure, in which the workpiece is not melted and recast, is friction stir welding (FSW) and friction stir spot welding (FSSW) [9,10]. Moreover, solidification defects like porosity, cracks, and voids are eliminated in both processes [11,12].
A literature survey indicated that, for medium to high-volume automotive production, there is a 20% saving in FSW compared to arc welding processes and 25% savings in FSSW compared with RSW [13].
Though much research works are available on friction stir welding variants [14][15][16][17][18] of AHSS, they are limited to microstructural characteristics and basic mechanical properties in basic lap configuration. However, to gain more understanding and confirm the feasibility of using AHSS for energy absorbing and other structural applications, a study on friction stir welding on lap configuration, i.e., friction stir lap welding (FSLW) and friction stir spot lap welding (FSSLW), must be studied in detail. The microstructure and joint performance of AHSS FSLW joints have received relatively little research so far. Wang et al [19] applied FSLW to join DP1180 grade steel by utilizing constant tool rotation, welding speed, and tools with varying pin lengths and reported that the desired microstructural features (i.e., martensite with good toughness) and maximum fracture load of 12.4 kN were noticed in the stir zone (SZ) when the tool pin length is 1.57 mm and lower sheet penetration depth of 0.4 mm was used.
Ghosh et al [20] attempted to join martensitic low-carbon steel by FSLW and found that the microstructure and joint performance are greatly affected by the cooling rate of the process and peak temperature at different travel speeds. Irrespective of travel speed values, the SZ and three different HAZs underwent softening. Das et al [21] attempted to control the temperature during the FSLW of DP590 within the A 1 -A 3 range, which involved adjusting the welding speed between 100 and 300 mm min −1 with a fixed pin rotation of 100 rpm. Although the authors claimed 100% joint performance with the fracture reported in the base metal (BM), the unbonded regions (hook defects) shown from the presented macrostructure may affect the joint performance.
In different zones of FSSLW joints made of 590-grade DP steel, OHASHI et al [22] observed uneven microstructural characteristics caused by the variation in the peak temperatures obtained by the zones during the welding process. Sarkar et al [23] examined the effect of FSSLW variables on the quality of DP590 joints. They revealed that variations in the heat input with respect to the parameters used significantly impact the failure mode, bonding area, weld metal phases, and their sizes. Zhao et al [24] investigated the effects of FSSLW variables on the weld zone size, microstructure features, phase fractions, and features of the bonding interface of DP590 grade steel sheets. They concluded that the heat input needed to be regulated to make high-quality joints. When using FSSLW to join Zn-coated DP590 steel, Rong et al [25] showed that the shoulder depth required to be controlled to achieve the desired distribution of Zn at the joint interface and achieve the highest failure load.
Although the data on the FSLW and FSSLW joints are available, a direct comparison of mechanical properties, microstructure, and the weld quality between the FSLW and FSSLWed joint of high-strength steel is not scanty. Moreover, data on the steel top hat sections welded by FSLW and FSSLW processes are unavailable. As the top hat sections have different weld geometry from conventional lap joining of steel sheets, the FSLW and FSSLW parameters must be optimized. Hence, this study is carried out to understand the effect of process parameters to join the top hat high-strength DP590 grade steel section by FSLW and FSSLW processes and correlate the properties with the weld zone microstructures under optimum welding conditions.

Experimental setup 2.1. Base sample preparation
This study employed the base sample of 1.7 mm thick cold-rolled Dual Phase (DP) 590 steel sheets. The elemental composition and the mechanical property of the raw material are mentioned in table 1. The sheet is bent into a top hat shape using GMT electro-hydraulic press with dimensions as mentioned in figure 1(a).

FSLW and FSSLW processes
Two variants of friction stir welding, namely friction stir continuous lap welding (FSLW) and friction stir spot lap welding (FSSLW), were used to join the top hat structure with the base member of 1.7 mm thick x 80 mm wide and 200 mm length dimensions. The base member and top hat section sample faying surfaces were polished and chemically cleaned before welding, and the samples were securely attached using a specially-made fixture. The design of experiments-based response surface methodology (RSM) approach was used to fabricate the top hat section by arriving at optimum welding variables. Based on previous studies, among primary and secondary welding variables, tool rotational speed (TRS), tool transverse speed (TTS), and shoulder plunge depth (PD) were selected for FSLW [26,27]. Instead of TTS, the parameter dwell time (DT) was considered for FSSLW [28]. Based on the visual inspection, working limits were adjusted by carefully accounting for the material loss on the top sheet due to adhesive wear, unbonded interface, cold lap, hooking defect, subshoulder development in the tool due to wear, tunnel and surface imperfections and weaker bonding ligament. The detailed data are presented in the results and discussion section. The welding parameters selected for FSLW are TRS with a range of .600-1200 rpm, PD between 0.25-0.55 mm, and TTS of 40-80 mm sec −1 . For FSSLW, the parameters are TRS and PD with the same range as the FSLW, and the third parameter is DT, ranging from 5-25 sec. To make it simpler to collect and evaluate experimental data, the maximum and lower levels (i.e., values) of the welding variables have been encoded as + 1.682 and − 1.682, respectively. The encoded values of any intermediate values (−1, 0, and 1) can be calculated using the relationship shown in equation (1) below: ) Y min denotes the variable's lowest level, Y max denotes its highest level, and Yi represents the variable's necessary coded value [29]. The actual values of selected FSLW and FSSLW welding parameters compared to the coded values, together with the ranges for each parameter, are finalized and tabulated in tables 2 and 3. Because of its greater hardness value and is robust enough to weld steel plates, the tungsten lanthanum oxide (WLa 2 O 3 ) tool was chosen. The pin height of the taper tool is 2 mm, the pin root diameter is 5 mm, the pin base diameter is 6 mm, and the shoulder diameter is 12.5 mm. To evaluate the temperature dissipation and its effect, a FLIR BX420 thermal camera was employed to record the joining processes.

RSM and characterization of FSLW and FSSLW top hat section assembly
The alternation of variables still affects the mechanical action and heat input, which in turn affects material flow, microstructure, and interface joint features in various regions of dual-phase steel FSLW and FSSLW joints. This is true even though the range of welding variables is fixed to obtain macro-level defect-free top hat weld assembly. The welding parameters TRS, TTS, PD, and DT must be employed at their ideal values to obtain a quality top hat weld assembly. Applying experimental design methods that can assist in locating the ideal values of welding variables is helpful because trial-and-error procedures are time and resource intensive [30].
Through the use of a uniaxial tensile test, RSM was used in the study to identify the ideal welding parameter values. Response Surface Methodology (RSM) is a set of mathematical and statistical approaches to analyze problems in which several independent variables influence a response variable to maximize the responses [31]. Tensile Shear Fracture Load (TSFL) is taken at the output response for both types of welding, which is supposed to be a function of the input parameters of the corresponding process. The samples for evaluating tensile shear failure load were extracted via wire electrical discharge machining (WEDM) from each welded top hat assembly fabricated as per the design matrix presented in table 4. With the help of a specifically designed fixture (figure 1(b)) that can hold a welded top hat assembly, the samples are tested using a crosshead speed of 1 mm min −1 on the 100 kN universal testing machine (UTM) to determine the tensile shear failure load (TFL). Three samples under each experimental run were tested; the results are presented in table 4. The top hat tensile sample before and after tensile shear failure loading are presented in figures 1(c)-(f).
The microstructural characteristics of the welded top hat assembly fabricated under optimum welding variables are investigated. For the microstructural examination, the samples are taken from the weld in both the cross-sectional and transverse directions. Silicon carbide abrasive sheets with grits ranging from #80 to #2500 and a final diamond paste are used to polish the mounted samples. After etching with 2% Nital, an optical microscope (OM) and scanning electron microscope are used for the microstructural study (SEM). Using a Vickers microhardness tester and a 0.2 kg force for a dwell length of 15 s, the microhardness mapping is done along several zones at the transverse direction of the weldment. Three rows are chosen, with the hardness values measured at 0.8, 1.6, and 2.4 mm distances from the top. The weld is kept in the middle as 50 places in each row are measured for hardness.

Outcome of initial trail experiments
The key welding variables, TRS, TTS, PD, and DT, are the ones that determine a bonding interface free of flaws in FSLW and FSSLW joints. A suitable working range must be established using trail FSLW and FSSLW studies to conduct experiments and produce high-quality joints free of macro-level flaws. Increasing TRS, PD, and DT and decreasing TTS give more heat to the weld zone. Additionally, the plunge depth and TRS impact the vertical and horizontal material flow around the pin, respectively. The material flow from the tool's front-preheating zone to its back-cooling zone is similarly influenced by the TTS in FSLW. In figures 2(a)-(c), the results of the first trials with FSLW joints are displayed. Faster welding and slower rotational speeds cause the cold lap flaws in figure 2(a). This resulted in incorrect bonding because there was inadequate heat and material flow, which impeded diffusion at the bonding interface. Due to turbulence in the material flow generated by a greater rotational speed and a slower welding speed, figure 2(b) illustrates the tunnel flaw at the fabricating surface. Cao et al [32] noticed a similar tunnel defect at the sheet interface during FSLW of mg alloys when using a tool with low TRS values, which can result in insufficient deformation. The surface flaws in figure 2(c) are caused by a faster welding speed and a shallower plunge, which lowered the heat provided to the welding region, accelerated material flow from the tool's front to its back, and  3.2. Developing relationships between TSFL and welding variables of FSLW and FSSLW processes RSM investigates the connections between a number of variables and one or more response variables in statistics [34]. Response Surface Methodology (RSM) is a set of statistical techniques that can be used to analyze problems in which a number of independent variables affect a response variable, intending to optimize the response [35].
RSM was used in this study to develop empirical relationships in the form of multiple regression equations for the quality attributes of friction stir lap and friction stir spot welded 590 grade DP steel hat section assembly. When using response surface methodology, the independent variable was viewed as a surface to which an empirical model was fitted [36].  [37]. It is a power series polynomial that is expressed as follows: Where a 0 is the mean of the responses, a 1 , a 2 , a 3 , andK..a 33 are the coefficients that depend on the primary and interaction factors. The statistical software package 'Design-Expert' and a central composite rotatable design were used to calculate the coefficients. After determining the significant coefficients (at a 95% confidence level), the empirical correlations were developed. The following empirical correlations have been proposed to estimate the tensile shear failure load of FSLW (equation (3)) and FSSLW (equation (4)) joints.

ANOVA results
The analysis of variance (ANOVA) procedure and the results of 2nd order response surface model fitting were used to test the appropriateness of the above-established correlations as given in the table 5. The difference between the Predicted R 2 and the Adjusted R 2 is less than 0.2, which is considered to be a fair agreement for both the welded conditions. In the same way, the model's F-value and P-value are both significant, which is an essential requirement. The P value also affirms that the welding variables considered for FSLW and FSSLW are significant. For FLSW top hat weld assembly, TRS is found to be the foremost significant factor on the TSFL, followed by TTS and PD, whereas for FSSLW, TRS, followed by PD and DT, is found to have a significant effect on the TFLS. The interaction effect between TTS and PD in the case of FSLW and the interaction effect between TRS and PD for FSSLW are found to significantly influence the quality attribute of top hat welded assembly. Figure 3 shows the perturbation plot for FSLW and FSSW with respect to the input parameters. From figure 3(a), it is very evident that the TRS is the most influencing factor for FSLW. After a certain level of increase, TTS produces lower heat as the tool movement increases. This allows the formation of a hook in the interface, which affects the tensile property of the weld. Z.W. Wang et al [19] also discussed the tool movement's influences on the joint's strength and fracture mode. As the TTS increases, because of the lower heat, the effective lap width reduces, reducing the joint's strength. Lower TRS results in weaker bonding in both cases since less material flow and less heat are produced during the welding operation. G.M. Xie et al [38] stated that the higher TRS generates a lot of heat, which causes the material to flow upwards and causes the connecting surface's interface to pull up and produce a hook which induces fracture. PD also has a considerable effect on the TFL. Lower PD means the bond between the elements is weak, whereas the severity of the upward bend of the hook increased with increasing PD, which affected the joints' mechanical strength. Ya-peng Zhao et al [24] also reported in their work that on the increase in PD, the area of SZ increases as well as the length of the bonded region increases, which is because of the increase in heat due to friction and the flow of material. M.Miles et al [39] inferred that higher PD increases the axial load on the area of contact between the workpiece and tool, which results in increased material flow in the interface, leaving more material movement in the joint behind the hook structure, which reduces the strength of the joint thus achieved. In the case of FSSLW as in figure 3(b), DT greatly impacts the temperature decapitated during the process. Lower the DT, the bonding thus formed weaker because of insufficient mixing of the materials. A. Al-Shahrani et al [40], concluded that higher DT leads to the generation  of more heat as the tool is in contact with the material for a longer period. This heat is sufficient to mobilize the material making the bond weaker. The optimization functionality of the design-expert software was applied to determine the best input process variables values for maximum tensile shear failure for FSLW and FSSLW top hat welded assemblies. According to the numerical optimization results for the developed empirical correlations, the best possible FSLW welding parameters for joining the top hat-base member assembly are a TRS of 900 rpm, TTS of 60 mm min −1 , and PD of 0.40 mm. Similarly, the best possible FSSLW parameters are a TRS of 900 rpm, PD of 0.40 mm, and DT of 15 secs.

Macrolevel zones in FSLW and FSSLW joints
Microstructural development throughout the process primarily influences friction stir welded joint's mechanical characteristics and fracture behavior. Optimizing the process parameters helps to attain favorable microstructure and phases in various weld locations [41]. The macrostructural zones in the FSLW joints under best welding conditions, as shown in figure 4, are categorized into three distinct zones, namely base metal (BM), heat-affected zone (HAZ), and stir zone (SZ). It was noticed that there was no visible thermomechanical affected zone (TMAZ) in the welds. The allotropic phase transitions that occurred during carbon steel friction stir welding remove the evidence of (TMAZ) [19]. The HAZ in the FSLW joint is divided into three zones, namely the crucial heat-affected zone (CHAZ), nuclear heat-affected zone (NHAZ), and peripheral heat-affected zone (PHAZ), as quoted by Mahdi Mahmoudiniya et al [4]. Similarly, the zones found in the FSSLW joint under best welding conditions are BM, HAZ, TMAZ, and SZ. These differences in zones and sizes are primarily due to variations in process physics between FSLW and FSSLW. TMAZ is discernible in FSSLW joints due to slightly higher strain rate, temperature, and lower heat transfer towards the transverse direction when compared to FSLW joints.

Thermal history of FSLW and FSSLW joints
Since the joining processes used here are solid-state processes, the material is not heated to the melting point of the material (1529°C), as seen in figure 5. A FLIR BX420 camera was used to record the temperature while the lap joints were done to understand the effect of the peak temperature during the welding processes. The area at which the tool interacts with the BM produces heat which eventually increases as the parameters vary. It was revealed in work by Kim et al [42] that the heat produced at the shoulder interface and the BM is more than the pin-affected region. Ac 1 and Ac 3 phase transition temperatures of DP590 steel were estimated to be 745°C and 860°C, respectively [25]. The peak temperature of the SZ of both the joints produced is higher than the Ac 3 temperature of the material. M. Ghosh et al [20] also noticed the influence of cooling rate on the formation of the complete martensite phase and stated that the cooling rate should be more than 21°C to attain the same. It was observed that the maximum temperature during FSLW is comparatively lower than the peak temperature attained during FSSLW. Also, when comparing the cooling rate of both processes, FSSLW has a higher cooling rate than the FSLW, as in figure 5. TTS, one of the main parameters during FSLW, is considered to be the reason for the lower peak temperature than FSSLW. The tool moves along the weld direction, interacting with new regions to join, because of which the heat dissipates to a larger area, but in FSSLW, the area under load is specific. During FSSLW, PD and DT are the main parameters for this increase in peak temperature.

Microstructure of base metal
Cold-rolled BM was made up of a ferrite matrix and martensite islands that are found inside and along the ferrite grain boundaries. Figure 6 shows the SEM image of the BM and volume fraction of the martensite. It is found that the martensite constitution in the BM is around 21.3%.

Stir zone microstructure of FSLW and FSSLW joints
During FSLW, the stir zone (SZ), as in figure 7(a), experiences dynamic recrystallization while exposed to high temperatures and extreme strain. SZ seems to have been heated to Ac 3 in the single-phase austenitic region. After severe deformation by the FSW tool, the material is rapidly cooled to achieve the martensite transition. The stir zone's microstructure morphology differs significantly from that of the other zones due to the presence of phase change caused by the high strain rate imposed by the tool [43]. The final microstructure of SZ of FSLW and FSSLW joints depends on the highest temperature and time to which it is exposed. Relatively higher peak temperature and more resident time in FSSLW caused an increase in the austenite size before cooling. However, a slower cooling rate in FSSLW joints caused fine lath spacing in the martensite as compared to the one in the FSLW joints. The lower peak temperature and less resident time restricted the austenitic growth before cooling, which caused reduced martensitic block grain size as compared to the SZ of FSSLW joints. However, a slower cooling rate during FSLW resulted in the precipitate of carbides in the SZ. The SZ of the FSLWed joint possessed around 30.23% of martensite, whereas the FSSLWed joint had 32.78%, as in figure 7(b). In both cases, the SZ has the highest martensite composition making this region harder than the other zones.
3.9. Thermomechanically affected zone TMAZ is a region that gets affected by thermal energy and deformation during the FSSLW process. When the temperature is raised between Ac 1 and Ac 3 , refined ferrite grains are produced, and the banded character of martensite is induced by full austenitization, as in figure 8. Due to significant strain stresses and the existence of an upward-flowing pattern surrounding the stir zone, the grains in the TMAZ are excessively elongated [24]. The microstructure of this region consists of ferrite and martensite phases. When the temperature was raised above Ac 1 during FSSLW, transformed austenite with fine microstructure and the untransformed ferrite phase might  emerge as a result of dynamic recrystallization. During subsequent cooling, the austenite changed into martensite, maintaining the mixed fine microstructure leaving around 25.81% of martensite in this zone. As mentioned earlier, in FSLW, the material flow and phase transformation during cooling are insufficient, thus destroying the elongated grain structure of the TMAZ, so no distinct TMAZ region is detected. The austenite to martensite transformation in DP steels is quite well recognized to be accompanied by a volume increase of 2%-4%, resulting in geometrically necessary dislocations (GNDs) in the ferrite base at the interface of ferrite and martensite. Because these dislocations accommodate plastic strain between grains, ductility increases as grain size decreases [44,45].

Heat-affected zone
In the case of the FSLWed joint, the outermost part of the HAZ adjacent to the BM is called the peripheral heataffected zone (PHAZ), which consists of ferrite and tempered martensite, as shown in figure 9(a). The maximum temperature in this region is found to be less than Ac 1 . Higher cooling rates in the HAZ are correlated with quicker heat dissipation of heat input through the HAZ. The amount of time that martensite is exposed to heat at a high-temperature decrease as transverse speed increases, which reduces the degree of tempering [46]. The ferrite and martensite phases in the BM get transformed into ferrite and tempered martensite due to incomplete recrystallization. The tempered martensite thus formed makes the PHAZ softer, wherein the martensite percentage was also found to be less than BM, around 18.7%. One of the most efficient techniques for reducing the PHAZ softening is to minimize the heat flow into the area, which can be achieved by increasing the TTS [47].
The nuclear heat-affected zone (NHAZ), as in figure 9(b), is the central part of the HAZ where the martensite and ferrite get heated to a maximum temperature between Ac1 and Ac 3 . Because of the thermal impact, the martensite's ability to self-temper was limited to form ferrite and austenite phases [16]. Higher heat input because of the increasing TRS results in a higher austenite formation rate. Around 23.57% of martensite was calculated in this area which is higher than the BM. The crucial heat-affected zone (CHAZ) is the closest region to the SZ, which consists of polygonal ferrite, bainite, and martensite phases, as in figure 9(c). This part of HAZ experiences a peak temperature of above Ac 3 during welding. Higher peak temperature at this zone because of lower TTS leaves lower heating and cooling rate, allowing the austenite grains to grow and become coarser grains upon cooling [48]. In this region, the martensite percentage was found to be higher in the HAZ, which is around 25.63% In FSSLW, HAZ adjacent to BM consists of tempered martensite and ferrite, where the peak temperature seems to be around Ac 1 . The martensite phase becomes coarser, which is responsible for this region's softening, contributing to a level of 19.67%, as in figure 10. Due to its greater distance from the shoulder face, the HAZ gets less heat input through conduction heat transfer.

Comparative assessment of the microhardness of FSLW and FSSLW joints
The hardness plot for the welded samples reveals that the hardness range of the SZ is much higher than the BM (280 HV), as in figure 11. The area surrounding the tip of the stirring tool underwent significant plastic deformation and displayed a high hardness value of around 540 HV. It is revealed that HAZ in FSSLW and PHAZ in FSLW had softened and had a lower microhardness rating of around 240 HV. The martensite volume fraction and the microhardness value are connected; thus, those with a greater volume percentage of martensite have a higher microhardness value. The carbide precipitation and martensite tempering are connected to the reduction in microhardness in HAZ [8]. The heat induced in SZ during the welding process increased as a result   of the higher TRS and higher PD. A greater quantity of coarse martensite may have formed because of the increased temperature, increasing the material's hardness. Localized ferrite deformation occurs during FSLW due to the ferrite phase's plastic deformation being limited by the surrounding martensite. The ferrite near the martensite islands is expected to have a localized high dislocation density as a result of the local shear strain during the FSLW process. Together, these findings strongly imply that the volume percentage of martensite and ferrite grain size affects the joint's strength [21].

Failure load evaluation of the welded samples
The FSSLWed joint under tensile shear load exhibits two types of failures: nugget pull-out and interfacial failure [31]. With lower PD, it is observed that the joints fractured in interfacial failure mode. When the PD is raised, the bonded area of the interface, as well as the mechanical characteristics of joints, increases. According to the  fractured specimen, Y.F. Sun et al [49] observed significant plastic deformation in both the lower and the upper plates along the welded region. After beginning at the welded area's edge, the fracture eventually spread along the depression's perimeter. In this case, the mode of failure is nugget failure. The length of the bonded interface may be expanded even more by the change in rotational speed, which may also have an impact on the thermal aspects during the welding process [50,51]. The TSFL for FSSLW was observed around 11.565 kN when the TRS was 900 rpm with 0.40 mm PD and 15 sec DT, whereas, for FSLW, the value was 13.68 kN, as shown in figure 12. This highest TSFL was obtained at the optimal condition of 900 rpm at PD of 0.40mm and 60 mm min −1 as TTS. It was also noticed that the joint from FSLW absorbed more energy and elongation before deformation than the FSSLW joint. The energy absorbed is the area under the curve up to the maximum failure load of the weld. The relationship between the tensile shear failure load and the energy absorbed before failure was also reported by Pouranvari et al [52]. The energy absorbed can also be found using the following expression (equation (5) where W is the energy absorbed before failure, P max is the maximum fracture load, L max is the maximum displacement at failure load, and F is the load. From the above equation, the total energy absorbed by the FSLW sample before fracture was calculated to be 170.13 J, and for the FSSLWed sample, it was 115.32 J. From the above results, it was found that the sample from FSLW absorbs more energy than the FSSLWed sample. This higher energy absorption value indicates higher reliability towards accidents and other loads.

Conclusion
The following key findings were derived from an attempt to join a single top hat section with a base member using two variants of friction stir welding, namely friction stir lap and friction stir spot welding.
• FSLW and FSSLW joints were produced using TRS between 600 and 1200 rpm and PD between 0.25 and 0.55 mm. TTS in the range of 40 to 80 mm min −1 was used to fabricate FSLW joints, whereas DT in the range of 5 to 25 secs was used to fabricate FSSLW joints. Joints fabricated outside of this range of welding variables resulted in a material loss on the top sheet due to adhesive wear, an unbonded interface, a cold lap, a hooking defect, sub-shoulder development in the tool due to wear, tunnel and surface imperfections, and a weaker bonding ligament.
• Empirical correlations were developed for estimating the maximum TFL of top hat weld assembly in terms of welding variables of FSLW and FSSLW processes. The best joining condition was obtained using a numerical optimizer based on the response surface model. The highest TFL of 11.6 kN was achieved for FSSLW joint assembly at the optimized condition of 900 rpm with PD of 0.40 mm and DT of 15 sec, whereas the highest TSFL was 13.768 kN at 900 rpm of tool rotation with PD of 0.40 mm and TTS 60 mm min −1 .
• Regardless of the welding variables used, the temperature in the stir zone exceeds Ac3, causing material deformation in a single-phase austenitic field that cools to a martensitic structure. However, there is a significant difference in the volume percentage, lath spacing, and carbide precipitates between the FSLW and FSSLW processes due to differences in the peak temperature, resident time, and cooling rate.
• The heat-affected zone in both FSLW and FSSLW underwent softening due to the tempering of martensitic structure and the presence of polygonal ferrite. The reduction in microhardness was around 14 to 18% as compared to the base metal. The Stir zone exhibited the highest hardness of 540 HV0.5 which is 92% higher than the base metal. This is mainly due to the presence of untampered martensite and fresh carbides.
• The total energy absorbed by the FSLW top hat welded assembly before fracture was calculated to be 170 J. At the same time, the total energy absorbed by the FSSLW top hat welded assembly was 115 J. This shows that the top hat welded assembly by the FSLW process has higher reliability towards crashworthiness than the joint assembly by the FSSLW process.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).