Analyses of Friction Stir Riveting Processes: A Review

Today's manufacturing industry is inclined toward evolving new technologies to meet rapidly changing needs [1]. Among those, weight savings and increasing fuel efficiency emerged as particular demands [2] in industries like aviation and automotive. Lightweight materials such as Mg and Al alloys, carbon steels, and polymer composites are widely used in combination for achieving the desired results [3,4]. These manufacturing advancements call for new efficient joining technologies. Dissimilar material joining poses more challenges than similar materials because of difference in their mechanical, chemical, and thermal properties. Nonetheless, dissimilar materials are difficult to be joined in multimaterial structures using conventional fusion-based welding methods. Mechanical fasteners, e.g., bolting and riveting, can be used independently as well as in combination with adhesives [5,6]. Current mechanical fastening substitutes for joining materials are solid riveting, blind riveting (BR), and self-piercing riveting (SPR), to name a few [7].

Solid riveting is a conventional fastening technique [8] where a solid rivet is placed inside a predrilled hole. The rivet shank deforms under an axial compressive load to fill the hole/cavity and forms the rivet clinch (Fig. 1(a) [9]). BR [10] is one type of solid riveting; however, in this process, more intricate rivets are used. A blind rivet has two components, mandrel and shank. The mandrel is a long rod with an increased diameter at one end; and the shank comprises a hollow tube with a flat cap on the end. The internal diameter is capable of housing the mandrel. The three steps of BR are predrilling, placing, and pulling [11]. The first step, predrilling, is to drill a hole on the work materials with a diameter larger than the rivet's body diameter. The second step, placing, as the name suggests, involves a blind rivet seating into the predrilled hole. The third step, pulling, is to drag the mandrel of the blind rivet until the mandrel breaks off. Figure 1(b) [12] gives a schematic representation of the process. In a self-piercing riveting, a hollow rivet is driven into the specimen materials that are incorporated in a shaped die. Initially, the rivet is driven into the upper surface of the work materials because of the pressure applied (through the press punch). At this stage, the rivet penetrates through the upper sheets of materials. At the same time, it also partly pierces the lower sheets before undergoing the effect of the lower upsetting die, which radially dislocates the hollow rivet end outward [13]. Simultaneously, the work material is displaced to fill any created voids. The entire process generates a mechanical interference and interlocking. However, two-sided access is mandatory for this process. Details of the process are illustrated in Fig. 1(c) [14]. In recent years, mechanical riveting methods such as SPR and BR are widely used in automotive industry for joining of dissimilar materials [15,16].

The welding institute is the pioneer in inventing friction stir welding (FSW) in 1991 [17]. Friction stir spot welding (FSSW) is a derivative of this process [18]. In FSSW, the material pieces are fastened by a welding tool through both sides. The welding tool is brought into contact with the parts at a desirable rotational speed (RS). Consequently, frictional heat is generated. The tool is further driven into the work material under pressure until it partially penetrates through both the work materials. The frictional heat generated during the process results in softening and subsequently joining of work materials [19–23]. The process is explained in Fig. 1(d) [24]. In automotive industry, FSW of aluminum and steel is utilized for manufacturing of vehicle components for instance, the trunk hinge on the Mazda MX-5 [25] and the hybrid steel/aluminum sub frame presented by Honda on the 2013 Accord [26].

From the previously mentioned discussion, it can be inferred that there are several mechanical riveting or friction stir mechanisms for dissimilar materials. In addition, technologies including adhesive bonding, laser welding, or other solid-state methods (e.g., ultrasonic welding) are being applied in joining dissimilar materials. However, there is a continuous rise in the expansion of substitute joining technologies to minimize the limitations of contemporary processes, such as joint performance, production time, and cost. This results in evolution of new techniques, which can meet the stringent manufacturing requirements, particularly in robustness and ease of implementation. Therefore, this paper aims at analyzing and comprehending an emerging joining technique, FSR processes, which integrates the advantages of both friction stirring and mechanical riveting. FSR are novel joining processes, which eradicate the necessity of predrilling for rivet insertion, and thus surmount the difficulties in hole alignment. The processes are fast, only taking a few seconds to form a joint, and ready with implementation of robot systems. There are different variants in use for FSR processes, which are friction riveting, friction self-piercing riveting (F-SPR), friction bit joining (FBJ), two-sided friction stir riveting by extrusion, and friction stir blind riveting (FSBR). This paper covers a comprehensive review of the different FSR processes starting from the description of different processes and fundamentals of the process design, followed by the analytical modeling studies, the underlying process parameters, different bond formations, and characterization of affected zones; presents research on mechanical behavior of FSR joints, behavior under corrosive environment; and in the end, concludes with synopsis and future outlook.

This section will cover important aspects related to process physics involved in the accomplishment of FSR processes. Each process is discussed individually, and processes sharing commonality are grouped together.

Many researchers joined various dissimilar materials by exploiting the advantages of friction stirring and mechanical fastening. Researchers at Helmholtz-Zentrum, Geesthacht, Germany [27] invented friction riveting process (called FricRiveting) to join hybrid thermoplastic-metal structures through spot connections. One or more thermoplastic components are joined with metal by inserting a round profiled (or plain) metallic rivet in them. In this process, the surface of the plastic component is pressed by a rotating metallic rivet (Fig. 2(a)) at high rotational speeds. The applied axial force and high rotational speed generate frictional heat, thereby creating a plasticized/molten film around the rivet tip (Fig. 2(b)). The frictional heat results in softening of materials, thus allowing the rivet to penetrate further. The tip of the rivet softened at the culmination of the heating stage, owing to a local temperature rise. This marks the beginning of forging phase, where the axial force is increased and the rotation is decelerated. The rivet tip thrusts the remaining soften material to the flash; however, the rivet faces resistance from the colder material and is deformed during the process to adopt a wider diameter (Fig. 2(c)). After consolidation under pressure, anchoring and adhesive forces hold the joint together (Fig. 2(d)) [28].

Researchers have used F-SPR process to join Al–Mg alloys [29,30]. The process is similar to self-piercing rivet other than the step where rivet rotates at high speeds and penetrates to soften the materials. Further, the shank of rivet flares and locks into the bottom workpiece similar to conventional SPR process. Consequently, a joint forms, which consists of the mechanical joining mechanism of SPR along with the solid-state joining technique of FSSW. The entire technique is distributed into four stages, namely (1) rivet feed stage where worksheets are fastened on the fixture, and the downward moving rotating rivet interacts with the material; (2) hot riveting stage in which the rotating rivet penetrates into the upper worksheet, and meanwhile softens and inserts into the bottom material for interlock formation; (3) friction stage where the downward movement of the rivet is stopped; however, it keeps spinning for a preset time to produce additional heat required for solid-state joining; and (4) off-stage where the rotation movement is halted to generate a static contact among the rivet and encompassing materials. Figure 3 gives the schematic of the process.

The hybrid friction-stir riveting method, developed at the University of Toledo, Toledo, OH [1,31], is very similar to F-SPR. In this process, a joint is formed by spinning and pressing a self-piercing rivet into sheet metals. Once the cycle is completed, the rivet is left in the sheets to form a joint.

Friction bit joining was developed by Mega Stir in which a consumable bit was used as the friction stir tool [32]. FBJ process mainly involves two steps, cutting and friction [33–35]. Initially, the consumable bit cuts through the top workpiece (Fig. 4(b)), which is followed by a joining procedure where the friction heat results in heating the bit and work materials (Fig. 4(c)). As a result, the bit acts as filler material to combine the sheets. In the first step (cutting), the consumable bit is rotated at relatively slow speeds. However, after cutting, the bit rotation speed is increased for more heat generation and subsequent strong bond formation. The purpose of spindle speed settings is to achieve two-pronged effects. Lower speeds help the bit to maintain its cutting edge necessary for proper cutting, whereas higher speeds result in softening the bit material (due to frictional heat generation) and bonding with work material. The joining bit is separated from the weld at the end of the joining process (Fig. 4(d)).

In 2015, a new method was developed for joining dissimilar materials and termed as two-sided friction stir riveting by extrusion. The method is an extension of one-sided friction stir extrusion process [36]. Two-sided FSR by extrusion combines elements of friction-stir extrusion (FSE) [37] and rotating anvil friction stir spot welding (RAFSSW) [38]. Two-sided FSR by extrusion uses FSE idea of extruding material by friction stirring, but does so at a single spot with the two-sided RAFSSW process. In this process, first in plunging process (Fig. 5(a)), the two FSSW tools are brought into contact at top and bottom sheets. Because of the stirring action, the softened materials flow into a predrilled hole at the middle sheet in the second step (Fig. 5(b)). In the third step, the tools are retracted from the work pieces leaving behind long rivetlike structures at both top and bottom sheets (Fig. 5(c)). Unlike other FSR procedures, this process uses two pinless FSSW tools, which feature a convex taper with scroll-like features.

FSBR was invented by General Motors, Detroit, MI, in 2006 [39–41], which is a blend of friction stir riveting and BR processes. In FSBR, a rotating blind rivet is moved downward and brings in contact with work materials. The generated frictional heat from the interaction between the rivet and the work materials softens the work materials, thereby permitting the rivet to be driven into under reduced axial force. The blind rivet is finally pulled once it is fully penetrated into the work materials [41,42]. FSBR eradicates the necessity of predrilling as in the traditional BR process while retaining the advantage of one-sided accessibility. The process is explained through a schematic diagram in Fig. 6 [43], and the process has been used for joining similar as well as dissimilar materials.

Researchers successfully joined a metal (Al or Mg) with a glass fiber-reinforced plastic (GFRP) sheet in lap joint configuration through a process termed as spin-blind-riveting (SBR) [44]. The process integrates flow drilling and conventional blind riveting. In this process (as shown in Fig. 7), a rotating rivet is simultaneously pressed into the sheets by a penetration force and generates frictional heat. The metal because of the rivet movement forms a sleeve analogous to flow drilling [45]. The formed sleeve is then pressed against the GFRP sheet and thermoplastic matrix begins to melt due to heat transfer by conduction. Because of the molten polyamide matrix, the fibers can be expatriated by the rivet rather being damaged. Once the rivet is entirely penetrated in work materials, the rivet mandrel is pulled back, and formation of mandrel head in rivet body occurs. SBR process is similar to FSBR process by considering the rivet movement and the final joint formation.

From the previously mentioned discussion, it is evident that there are different processes available under the umbrella of term “friction stir riveting.” Every process is distinct from the other in process setup and implementation. However, the underlying principle for each process is similar, which is the utilization of frictional heat generated by tool–workpiece interaction to allow an easy penetration of mechanical rivet or form a special material flow. However, the last stage of every process is different, which defines the role of the rivet in the joint. A brief summary of all the FSR processes is presented in Table 1.

From the previous discussion, it is clear that except friction-stir extrusion process, in all other processes, the tool, e.g., the rivet or the bit, is clamped inside a fixture in a machine, which is also used to drive it. This condition necessitates the elimination of any comparative motion or slippage among the tool and the spindle fixture [46]. Failure of abiding this requirement results in two situations. (1) Only relative rotational motion (torque) occurs. In this case, the compromised energy transfer from the fixture to the tool ultimately reduces the tool rotation speed to zero. This scenario is also devoid of any friction stirring, and the entire process only results in penetration, which eventually causes tool failure. (2) Translational relative motion along the feed direction is involved. This case results in pressing the fixture against the tool head before it fully penetrates the work materials. In addition to these two generic failure modes during the FSR processes, for FSBR and SBR processes, there is an additional failure mode. The mandrel body of a blind rivet has an indentation for easy tensile break-off. This notch can degrade the torsional strength resulting in shear rupture of the mandrel before it can fully penetrate into the samples. This failure mode occurs when the necessary penetration torque exceeds the design torque limit of the mandrel.

Torque and transverse force are two important components that act in synergy for an efficient FSR process. Deviation from their optimum conditions will result in either damaging the equipment or a weak joint. In addition, torque and transverse force vary throughout the FSR processes according to different contact conditions between rivet head and the workpiece.

Different analytical models were proposed by researchers to estimate the generation of frictional heat and resultant forces in FSR processes. The models were implemented to optimize the process parameters by correlating the forces with the performance indicators. Altmeyer et al. [47], when investigating FricRiveting process, deduced the mechanical energy relationship as a consequence of forces from rotational motion only. The authors proved that translational forces had the negligible impact on mechanical energy input. Blaga et al. [48] derived an analytical model to define the anchoring efficiency of the rivet in FricRiveting process. The authors called the model as “the volumetric ratio,” and used it to link the anchoring efficiency with the tensile strength of the joints. Amancio-Filho et al. [49] studied the thermal degradation of a polymer under increasing rotational speeds using Fourier transform infrared spectroscopy, gel permeation chromatography, and X-ray computer microtomography in the FricRiveted zone. Based on the results, the researchers correlated the molecular weight measurements with thermal degradation phenomenon. Min et al. [50] utilized the force and torque considerations from solid mechanics theory [51] to develop analytical models for FSBR process. The developed models were used to compute the material penetration force, material removal rate, and torque during the FSBR process. The models have the potential to extend further to other FSR or friction stir drilling processes, where material removal is involved. The underlying assumptions in the Min's model, derived from Ref. [52], suggested that there could be three possible contact conditions between the rivet and the work materials, which are (1) pure sliding, (2) absolute sticking, and (3) mixed sliding and sticking in any friction stir process. The researchers utilized those assumptions to assess the contact condition by equating the ratios of torque to penetration force in FSBR [41]. The authors proposed torque to penetration force ratio for evaluating the communication condition in the FSBR process. They concluded that initial rivet penetration was pure sliding which later changed to the mixed contact beyond a critical penetration depth. In the pure sliding condition, a linear relationship between material removal rate and the penetration force or torque was observed. This relationship was found independent of the rotation speed but was greatly influenced by the feed rate.

Amancio-Filho [53] suggested an analytical model to calculate the heat input during FricRiveting of unreinforced aluminum/thermoplastics. The model allowed the calculation of the total heat generated; however, the model was restricted only to thermoplastic materials as the constitutive model was based on the shear rates and variations in viscosity of the molten polymer.

Available analytical models cover the estimation of rivet anchoring efficiency, frictional force, and torque required during the FSR processes. However, the models do not account for the varying frictional heat generation due to the interaction of rivet and work material during the process. The heat development regime during FSR processes is an extremely intricate phenomenon, which involves solid friction and viscous dissipation (internal shearing of macromolecules) between the work and rivet materials. A comprehensive thermomechanical model catering for thermal softening and phase changes in work material during the FSR processes is not available. The present models can be augmented by incorporating the thermal effects to obtain a superior estimation of the associated forces.

The successful joint formation through the friction stir processes depends on performing the operations on their optimized parameters [54]. Available literatures give an insight on the combination of different parameters to assess the feasibility of the process. Studies conducted by Min et al. [55] explored the process parameters for FSBR joints with two different shank diameter rivets and various Al alloy sheets. The authors used three different settings of spindle speed and feed rate (for each parameter) to optimize the process parameters. The study suggested that maximum penetration force and peak torque increased with feed rate. However, increasing the spindle speed had a reverse influence on maximum penetration force and peak torque. Gao et al. [56] confirmed the importance of rivet design, parameters like off-axis angle, and rivet cap diameter, which significantly contributed toward the joint strength. The findings showed the inconsequential effect of rivet diameter on joint strength. Lathabai et al. [57] studied the effect of rivet design on force and torque parameters for FSBR joints. They suggested that penetration force required in case of blind rivets with hollow mandrel heads was much lower than the one with solid mandrel heads. Wang [58] probed the process parameter effects on the FSBR process. The materials used for this purpose were Al, Mg, and carbon fiber-reinforced polymer (CFRP) composite materials. FSBR joints were made using different stacking sequences of the materials. Through the experimental results of tensile tests and torque and force measurements, the authors found the configuration of the material sheets as the most important factor for FSBR joint. Feed rate also appeared as a primary contributing factor in quality issues like the gap between the work materials and rupture at the bottom sheet. However, spindle speed was not as important as configuration and feed rate.

In Ref. [44], Podlesak et al. discovered that high-quality joints were formed by SBR process when the spindle speed was higher than a critical value, i.e., 3500 rpm, with a stable process time. The researchers recommended these settings for all material combinations. They discovered that a lower spindle speed resulted in less frictional heat and consequently worse plastic deformation especially in the case of Mg.

Blaga et al. [48] used a full-factorial design with a center point to assess the influence of the process parameters in FricRiveted joints on the joint formability and mechanical behavior under tensile loading. The process parameters were rotational speed (RS), friction pressure (FP), forging pressure (FoP), and friction time (FT). The evaluated parameters were the mushrooming efficiency, which was defined as a measure of anchoring efficiency of the rivet in the base plate, the rivet penetration depth (T), the mechanical energy input, and the pull-out force (F_pull-out). The authors suggested that lower RS, FT, and FP values were suitable for less mechanical energy inputs to avoid likely thermal degradation of polymer matrix. Amancio-Filho et al. [59] described the rivet geometry in anchoring zone (AZ) as the most important parameter and linked it with the mechanical performance of joints. They argued that anchoring region experienced the bulk of the load faced by the joint. They further formulated that anchoring zone geometry was substantially affected by the frictional heat in addition to the deformation force borne by the rivet during the forging phase. By evaluating the thermal degradation of Al 2024/polyetherimide (PEI) FricRiveted joints, it was established in Ref. [60] that the rotation speed appeared as a prime contributing factor in governing the rate of heating, heating time, and temperature change. An increasing trend in temperature and heating rates was established with high rotational speeds, and the heating time increased only slightly.

Miles et al. [33] found that in friction-bit joints, the joint strengths were profoundly influenced by material and features of the joining bit in addition to the spindle speeds and feed rates. Squire et al. [61] in their research work of joining high strength steel (HSS) and Al alloys through FBJ found that increasing spindle speed and feed rates had a positive influence on joint strength. However, it was found that after a certain rpm (3250), there was a sharp decline in the tensile strength. The researchers attributed this phenomenon to the excessive heat generation.

For friction stir riveting by extrusion, Evans et al. [36] determined a noteworthy variance in the joint strength once the two aluminum sheets were only extruded through the hole but not joined as compared to when they were extruded and joined. The joints made through joined and extruded aluminum sheets displayed superior strengths compared to the joints where the aluminum sheets were only extruded.

Extensive investigations have been devoted to evaluate the process windows and parameters influencing the performance of FSR processes. Spindle speeds and feed rates along with the tool design emerged as important considerations for all FSR methods. However, these parameters (spindle speeds and feed rates) need to be optimized within a process window for an FSR process. Improper selection of these parameters may result in excessive or insufficient heat generation along with the undesired plunging/cutting force, which affects the bond formation. The gap between work materials is a common defect, which is caused by the material flow under improper process parameters. Also, the configuration of materials in a joint and rivet geometry serves as a major contributing factor for the processes like FSBR, SBR, and FricRiveting. In general, the softer work material is preferred to be placed above the other one to facilitate the material flow, and in some cases, form the mechanical interlocking.

Understanding the microstructural changes in stirred area is important as it develops the foundation for successful studies into thermomechanical modeling and mechanical performance of the FSR processed joints. Min et al. [62] examined the microstructural development of an FSBR aluminum alloy sheet (AA6111) through electron-backscattered diffraction (EBSD) technique. The researchers found distinct zones near the rivet, which were one stir zone (SZ), three thermomechanical-affected zones (TMAZs), and one heat-affected zone (HAZ). HAZ and TMAZ terminologies represented the same meaning as in an FSW process [63]. The amount of shear deformation varied in all TMAZs, whereas with the increasing distance form rivet hole characterized the HAZ formation. Through EBSD, it was found that only the SZ underwent recrystallization. Microhardness of HAZ was found to be on the lower side than the parent base material. Temperature and the tangential shear stress decreased with the increasing distance from the hole edge. These phenomena resulted in lowering of the amount of deformation in the three TMAZs. Figure 8 [62] depicts the grain structure of various zones.

Croom et al. [64] investigated the distributions of fiber, metal inclusion, and pore volume fractions in the stirred region of a CFRP in an FSBR joint by using micro X-ray computed tomography in combination with volumetric digital image correlation (V-DIC). The authors found significant microstructural differences between stirred region and the bulk material. The study results showed an apparent increase in volume fractions of fiber, metal inclusion, and pore in the stir zone. An interesting observation was the existence of sporadically spread large size metal inclusions. Large size metal compositions were aluminum, which believed to be coming from aluminum workpiece placed as the top sheet. The authors pointed out that the presence of aluminum inclusions was owing to the high operating temperatures occurred during the process.

Amancio-Filho [65] characterized the microstructural zones of FricRiveted joints between amorphous polymers and aluminum. Five distinct microstructural zones (Fig. 9) namely, the anchoring zone (AZ) of the deformed rivet tip, the polymer heat affected zone (PHAZ), the polymer thermomechanically affected zone (PTMAZ), the metal heat affected zone (MHAZ), and the metal thermomechanically affected zone (MTMAZ) were identified. The characterization was conducted based on the temperature and high deformation rates. The researchers found metal–polymer interface in MTMAZ. PTMAZ was between the MTMAZ and PHAZ, characterized by a weld line with consolidated polymer. It was inferred from the analysis that interfaces were mainly held by adhesive forces. Rodrigues et al. [66] revealed microstructure changes in a friction-riveted joint of Al 2024-T351/polycarbonate (PC), such as different microstructural zones in the anchoring zone and grain realignment in the MTMAZ. However, no microstructural changes were visible in MHAZ, compared with the base material. The authors observed that there was no noticeable change in polycarbonate as it was amorphous and transparent.

The changes in metals are related to the grain structure modification caused by phase transformation. The changes occur due to softening and shear deformations. Based on the grain refinement and adjustments, the researchers categorized the affected regions in various zones—mainly stir zone, thermomechanical, and heat affected zones. In a frictionally stirred thermoplastic polymer, melting is the main observation, which may have the adhesive effect on the metal joined next to it.

To understand the mechanical behavior of the FSR joints, a detailed insight of different bonds formability during the joint formation is essential. Gao et al. [56] tried to interpret the interface morphology of the FSBR joint. The authors cited that frictional heat that was generated throughout the process might have caused the formation of the rivet to work material bonding and interwork piece welding. However, their results exhibited minor indication of a direct association between frictional heat generation and joint strength. Wang et al. [67] stated that during the FSBR process, the bottom material tried to flow up. The authors articulated that in the case of bottom material being stronger than the upper work material, it penetrated into the upper sheet and resulted in the formation of a mechanical interlocking at the interface of two layers. On the other hand, a gap formed between two sheets if the bottom material was softer, as the lower material could not be pierced into an upper layer. Authors performed experiments with different configurations of CFRP, Al, and Mg sheets. They found that CFRP/Al and Mg/Al configurations had different mechanical interlocking. CFRP layer on the top created a mechanical interlocking around the rivet body since the residue was corn shaped, which caused the bottom material to flow up with an angle. However, aluminum or magnesium on the top could not generate mechanical interlocking around the rivet body because their residue shapes were concaves, and the material flows were different. Additionally, the mechanical interlocking of Mg/Al configuration was about 1 mm away from the rivet body. Intermetallic compound (IMC) layer was not observed in FSBR joint. Figure 10(a) illustrates the mechanical interlocking bond formation for CFRP/Al configuration [67].

Miles et al. [35] used a thin intermediary film of interstitial free steel between the cast iron and the aluminum to yield adequate frictional heating for bond formation in FBJ process. The researchers observed diffusion bonding between the joining bit and cast iron because of the steel plate. In another research work [34], Miles et al. while joining HSS and Al alloy through FBJ found that the bonding area of the materials was larger than the characteristic weld nugget diameter of a resistance spot weld steel used in automotive applications. They argued that the increased bond diameter helped in improving the joint strength in comparison to spot welding. Good metallurgical bonding between the steel and Al alloy was seen (Fig. 10(b)) and validated by microchemistry analysis. However, no intermetallic phases were found between Fe and Al. In F-SPR process [29] (Fig. 10(c)), IMC layer was observed at different stirred regions along with the thick part of mixed materials. The regions became thicker with increasing dwell time.

In summary, in addition to mechanical connection, the bonding phenomena include mechanical interlocking and chemical bonding, e.g., IMC and non-IMC layers. The bond formation is highly dependent on the material flow through the process, which in turn depends on the process dynamics and the tool design.

For the optimal utilization of any structure, it is imperative to comprehend its mechanical behavior when subjected to various loading conditions. This section covers various aspects of mechanical characterization for FSR joints.

Extensive literatures were reported in identifying the failure modes of FSR joints under different loading conditions. In Ref. [44], the researchers found lateral, shear-out, and bearing failure as the failure modes for spin riveting.

Wang et al. [67] classified the failure modes of different configurations of FSBR lap-shear joints under static tensile loads. The authors used an in situ acoustic emission analysis technique to evaluate the commencement and growth of damage. For as-fabricated CFRP/Al FSBR joints, cleavage and tension failures occurred in the CFRP workpiece. Three different failure modes were observed for as-fabricated FSBR Mg/Al joints, i.e., tension, shearing, and bearing followed by cleavage patterns. The dominant mode of failure was tension mode. Likewise, for the as-fabricated FSBR CFRP/Mg joints, failure occurred either at the CFRP (tension failure) or at the Mg side (mixed failure of tension and shearing). Moreover, a rivet pullout was also observed. Min et al. [68] classified the failure modes of FSBR AA6022/AA6111 joints. They found that the thicker AA6022 workpiece underwent minimal bending deformation during tensile testing and the majority of the displacement occurred due to thinner AA6111 workpiece tearing. The authors concluded that in AA6111/AA6022 joints, mechanical properties were governed by the properties of the AA6111 workpiece, which was the fragile constituent material in the joint.

Amancio-Filho et al. [69] classified five failure modes for a FricRiveted thermoplastic material joint subjected to static tensile loading. They are given as follows: (1) Through the rivet (type I), in which ductile fracture occurred in the metallic rivet outside the joint. (2) Rivet pulling out with a back plug (type II) took place because of crack nucleation at the rivet deformed tip in the anchoring zone. The term “back plug” referred to the leaving part once the rivet was pulled out. (3) Full rivet pulling out (type III) was hallmarked with the complete removal of the rivet leaving behind a hole of similar diameter as of deformed end. (4) Rivet pulling out (type IV) was characterized by large deformations at the rivet tip with small insertion depths, which caused the anchoring zone close to the material surface. (5) Rivet pulling out with secondary cracking (type V) involved an intricate failure mechanism [63]. In this failure mode, nucleation was observed at different positions nearby the anchoring zone. The initial phases of crack promulgation resembled type IV failure, but the final fracture occurred as typified as type III failure. Figure 11 describes different failure modes [69].

Two-sided friction stir riveting by extrusion processed joints depicted shearing of the aluminum extrusion at both edges of the steel along with a breaking of the aluminum/steel bond [36]. In most of the cases, aluminum shearing occurred at both the top and the bottom of the steel hole leaving a deformed, cylindrical section of aluminum in the hole.

Different FSR processes show various failure modes because of the different material interactions across the interfaces. Also, the fastening mechanism in different FSR processes (e.g., tail forming in FSBR and anchoring zone in FricRiveting) changes the distribution of forces in the joints, thereby resulting in different failure modes.

Mechanical properties are important characteristics in estimating a joint performance under different service conditions. The preceding paragraphs provide an insight of mechanical performance of FSR joints.

All published literatures on FSR joints conclude that the resultant joint through FSR is of higher structural strength than many other joining methods, such as adhesive bonding. Min et al. [70] investigated the tensile strength of Mg/Al FSBR joints (manufactured at different spindle speeds and feed rates) and compared them with the ones having predrilled holes. They used three materials cast Mg AM60, rolled 1.5 mm thick Al AA6022, and extruded 3.15 mm thick Al AA6082 for the study. FSBR joints in all the cases were found to be of superior tensile strength compared to the joints with predrilled holes. Based on the analysis, tensile strength was linked to parameters such as tail forming process, frictional penetration, material matching, and sheet position. Min et al. [71] also found superior tensile strength for CFRP/CFRP and CFRP/Al FSBR joints. Zhang et al. [72] studied HSS/Mg alloy FSBR joints. They found that tensile strength varied with changing feed rate and material configuration. It was noted that with increasing feed rates, load-carrying capability and percentage elongation increase for FSBR samples.

In Ref. [66], the tensile strength of the AA2024-T351/polycarbonate (PC) FricRiveted joints exhibited excellent values of ultimate tensile forces in comparison to riveted joints, i.e., 68.4% greater strength. The similar range of superior tensile strength values was also reported by Amancio-Filho [65] for AA2024-T351/polyetherimide (PEI) FricRiveted joints. Blaga et al. [48] used tensile testing to investigate the mechanical anchoring behavior of FricRiveting joints. The researchers measured the contact volumes between the polymer and the rivet to define the mechanical anchoring. FBJ and SPR joints [73] made of Al 7075-T6/dual phase (DP) 980 steel alloy were compared in terms of shear strength. Results revealed that FBJ average lap shear strength was 6.4 KN in comparison of 5.0 KN shear strength of SPR joints.

Reported literatures reveal that FSR processes usually depict higher strength compared to conventional fastening techniques. Although all the researchers have cited metallurgical/mechanical bond formation as the primary contributor to this improved performance, different reasons were proposed for the bond formation in each individual FSR process.

In the literature, to evaluate the softening and thermal effects of the process in stirred zone, microhardness tests were conducted for different FSR processes. In Ref. [74], researchers discovered that the microhardness of a friction riveted polymer metal joint was associated with microstructural changes due to thermomechanical processing. For metallic rivet portions within the worksheets, a reduction in microhardness with respect to the parent metal was observed. However, the decrease in the values was smaller for the MHAZ than TMAZ. Moreover, regions in the vicinity of the rivet experienced an increase in microhardness compared to the base material. The authors associated these changes in hardness in the PHAZ with physical aging and hardening. The phenomenon occurred as a result of loss of structural water, coming as a by-product when different monomers joined together to form polymer.

Min et al. [68] carried out microhardness tests for the top sheet (AA6111) and presented the results in the form of relative hardness values, where the interrupted and complete joints were measured at four different layers through the thickness. In an intermittent joint, layer-I (layer closest to the AA6111 top surface) had the maximum hardness. The authors related this phenomenon with the massive deformation of upper surface of the AA6111 workpiece that was near the shank. Min et al. [70] also investigated the hardness profile of AM60 alloys. The authors explained that the increase of hardness phenomenon in FSBR stir zone was due to frictional penetration and tail forming (expansion of rivet's tail once the mandrel is pulled out) process. Croom et al. [64] explored the local mechanical properties in the stirred region of CFRP. V-DIC was used for 22 and 44 MPa compression loads in the near rivet hole specimen to calculate the axial strain. The results revealed that enhanced metal inclusion and fiber contents improved the material stiffness in the stir zone. Also, localized bands under high axial compression were detected on the rivet hole surface highlighting a nonuniform deformation in different stirred zone areas, which was related to kinking introduced by localized buckling (Fig. 12) [64].

Fatigue life performance consideration is necessary as the joints undergo dynamic loads throughout their service life. Moreover, it helps in devising the fail-safe methods. In that regard, Gao et al. [56] performed the fatigue life comparison of Al 5052-H32 FSBR joints with the similar material joints made using resistance spot welding and blind riveting processes. Experimental results concluded that the fatigue life of FSBR was one to two orders of magnitude higher than spot welding and blind riveting process. Fatigue testing of Al7075/DP980 steel FBJ [75] joints revealed that the failure occurred in the aluminum material while the FBJ bit and the weld zone remained intact and undamaged. It was concluded from the experiments that FBJ bonds had improved fatigue properties compared to the base aluminum material.

Available studies regarding mechanical properties show improved tensile and fatigue strengths compared to conventional riveting mechanisms, but most of the studies are focusing on linking the mechanical properties to process window optimization. Limited literature is available in establishing the relationship among microstructure evolution, mechanical properties, and the process mechanics, which inhibits the development of comprehensive understanding of the FSR processes.

A thorough understanding of mechanical joints under corrosive environment is vital, particularly when dissimilar materials involved, which increase the likelihood of galvanic corrosion at the joint interface [76–80]. FSR joints are also susceptible to this phenomenon. Li et al. [81] examined the behavior of FSBR Mg/Al joints under severe marine environment. The authors found through Fourier transform infrared spectroscopy and X-ray powder diffraction (XRD) analyses that FSBR Mg/Al joints experienced both crevice and outside of the crevice corrosion. Mg passivated inside the crevice, and substantial corrosion was found on Al. Within the crevice, hydrogen-evolution cathodic reaction yielding OH⁻ resulted in high pH environment. The additional OH⁻ then prevented Cl⁻ in the crevice, thereby further increasing the passivation of Mg. Outside crevice, minimal corrosion was observed on Al, contrary to a significant amount of corrosion on Mg. The authors suggested that preventions should be taken to protect both Al and Mg metals when a crevice is formed. In addition, the authors found that at the interface, the rivet was free of corrosion indicating little or no seepage of moisture from the environment due to the tight contact between rivet and work materials.

Researchers studied the corrosion behavior of spot joined Al 7075-T6/DP980 steel alloy by FBJ [82]. The study was conducted considering two scenarios: (1) with adhesives in the joint (weld-bonding specimens); and (2) without adhesives (FBJ only specimens). Using tensile testing, it was instituted that adhesive-FBJ bonded specimens showed higher strengths than the FBJ samples. The researchers found a large crevice between the steel and the aluminum sheets only in FBJ specimens, providing an entry path for corrosive material. The path was closed because of the existence of the adhesive for the weld-bonding joints, therefore resulting in better corrosive resistance.

In addition to galvanic corrosion, which is believed to be the major corrosion mechanism in the dissimilar material joint, crevice corrosion should be considered in FSR joint since the crevices are usually found between work materials. The size of crevice between the rivet and work materials, however, is much smaller than the conventional riveting where predrilling hole is required; and thus, minimize the corrosion of rivet itself. If the rivet does not corrode significantly, the joint maintains its majority of the strength. Nevertheless, corrosion behavior of FSR joints needs further intensive investigations.

The study brings forth various advantages of FSR processes, some of which are summarized below:

• Most of the FSR processes eliminate the need for a predrilled hole, thereby reducing the difficulties in laminating the multiple holes during joining.

• The methods are capable of joining both similar and dissimilar materials with a wide variety of material selections.

• They are highly suitable for batch production and automated feeding system.

• FSR processes require a shorter span of process times including post processing and sample preparations as compared to other joining methods, e.g., adhesive bonding combined with welding.

• FSR joints have relatively high strengths and large displacements before fracture.

• FSR processes allow the development of newer products and sophisticated design that were previously not possible with conventional joining processes.

Like any other process, FSR processes also have some limitations. The primary limitations are described below:

• The nature of processes only permits production of spotlike joints.

• FSR processes are limited to opting for optimized worksheet configurations for high quality joints.

• Friction-riveted joints like other bonded and riveted joints cannot be re-opened.

The paper presents a detailed review of FSR processes, which includes process physics, mechanical behavior, process parameters, microstructural studies, corrosion behavior, process modeling, process advantages, and limitations.

The techniques under “FSR” umbrella are promising joining techniques that synthesize the advantages of friction stirring and mechanical fastening. All FSR processes rely on frictional heat generation resulted from the interaction of the rivet and joining material. The frictional heat leads to softening the material, thereby allowing sufficient stirring, which subsequently causes mechanical interlocking where material flow acts as an additional bonding mechanism to riveting itself. In some cases, metallurgical bonding between the rivet and work materials can also occur. Thereby, improved mechanical properties of FSR joints are usually seen when compared to conventional riveting method. The process parameters, including spindle speed, feed rate, and work material stacking sequence, emerge as crucial factors for defect-free joints. Rivet geometry and material are also important considerations for the process's success. In the current FSR processes, most rivets are commercially available. Special rivet geometry design is still in the early investigation stage. Three major microstructural zones are usually observed in FSR processes. Stir zone is characterized by refined grain structures; thermomechanically affected zone experiences medium deformations and temperature rises and is characterized by deformed grains; and heat-affected zone involves only temperature rises without apparent grain structure change but sometimes precipitate coarsening happens.

For the future research on FSR, several key issues need to be addressed to broaden its applications. First, accurate process modeling is essential to FSR process design, automation, and optimization; however, it is limited by its complex thermomechanical and metallurgical nature. Surmounting this limitation requires multiscale modeling with efficient submodels to describe heat transfer, material flow, changes of metallurgical and mechanical properties, and contact conditions during the process. Additionally, mesh-free method is suggested to use in the simulation to avoid mesh distortion caused by large deformations in FSR process. Second, corrosion as an application issue for dissimilar material FSR joints is still not well understood. It is believed that galvanic is the common corrosion mechanism in dissimilar material joints. However, due to the insertion of conductive rivet, the crevice in the lap joints, and the different natures of metals and nonmetals joined through FSR processes, the corrosion mechanisms are complicated. To reduce corrosion, rivet coating or integrating adhesive with FSR is a possible solution. The research on rivet coating for dissimilar material joints is still in its infancy, and the coating should be able to sustain during the friction stir process. Third, fundamental understanding and predictive modeling for structural performance and integrity of FSR joints under dynamic loading and extreme environments are missing. To overcome the research gap, it requires the physics-based lifetime and reliability models that quantitatively predict the materials performance and durability of dissimilar material FSR joints operating at different environments; modeling and prediction of the crack generation and distribution introduced by application; and cause diagnosis that accounts for processing, material structure and environment.

This work was funded by the U.S. National Science Foundation Civil, Mechanical, and Manufacturing Innovation Grant Nos. 1363468 and 1651024.