Effect of Strain Rate on Aluminum–Polymer Friction Stir Joints Mechanical Performance

Abstract

Friction stir-based joining techniques offer a promising route for the integration of highly dissimilar materials into single structures, with potential applications in safety-critical sectors such as hydrogen storage and lightweight mobility systems. Ensuring structural integrity under dynamic loading is crucial for their industrial adoption, particularly given the strong inhomogeneity of metal–polymer interfaces. This study investigates the strain rate sensitivity of lap joints between an AA6082-T6 aluminum alloy, and a glass-fiber-reinforced polymer (Noryl™ GFN2) produced using a friction stir process. Quasi-static and intermediate strain rate (≈3 s⁻¹) tensile tests were performed on the joints, while both base materials were additionally characterized at quasi-static, and intermediate strain rate conditions using a custom accelerated electromechanical testing device. Digital image correlation was employed to monitor deformation. The results reveal that the joints exhibit clear strain rate sensitivity, with ultimate remote stress and bending angle stiffness increasing by approximately 30% and 23%, respectively, from quasi-static to intermediate strain rate loading. Fracture consistently initiated in the polymer, indicating that the joints mechanical performance is limited by the polymeric constituent, although the polymer strain rate hardening impacts the peel/shear mix in the loading scenario of intermediate strain rate loading. Overall, the findings highlight that while friction stir metal–polymer joints benefit from strain rate hardening, their performance envelope remains governed by the polymer base material.

1. Introduction

Dissimilar metal–polymer joints are central to lightweight, multifunctional structures in transport and energy systems, yet their structural reliability under dynamic events (crash, blast, drop) remains an open challenge. Friction stir-based technologies (FSW/FSJ) have emerged as powerful solid-state routes to join thermoplastics to aluminum without adhesives, enabling robust mechanical interlocking and controlled thermal exposure at the interface. In Type-III hydrogen vessels, an aluminum liner overwrapped with fiber-reinforced polymer provides hydrogen tightness and shares load with the composite shell; recent pressure-vessel studies and reviews document liner roles, burst-pressure scaling, and design trade-offs for Al/CFRP systems, underscoring the need for reliable metal–polymer interfaces under service transients [1,2,3]. In ground vehicles, hybrid Al fiber-reinforced polymer components and tubes show notable gains in crashworthiness and specific energy absorption compared with monolithic aluminum, motivating multi-material architectures in body structures and protective housings [4,5]. Recent reviews and applications underscore rapid progress but also emphasize gaps in rate-dependent performance and failure mechanisms for metal–polymer FSW joints [6,7]. Given the potential applications of such joints in crash-prone conditions, it is important to fill this research gap as to enable such applications.

In quasi-static conditions, processing parameters, tool design, and local thermal cycles govern interfacial morphology and joint strength. Recent studies on AA6082 thermoplastic systems (including Noryl™ GFN2) map how temperature windows and tool kinematics steer failure from polymer-dominated to interfacial modes [8,9]. However, under dynamic loading the constitutive response of thermoplastics and their interfaces change markedly. Polymers and structural adhesives often exhibit pronounced strain rate hardening and brittle-to-quasi-brittle transitions, while aluminum contributes comparatively modest rate effects, shifting the failure locus and load transfer across the joint [10,11]. Impact studies on metal composite hybrids further suggest complex energy-dissipation pathways (rebounding, secondary bending), but comprehensive rate-dependent tensile data for friction stir metal–polymer joints are still scarce [12].

Experimental characterization of polymer and composite materials is challenging as these materials exhibit high time-dependent mechanical behavior. These materials may change from rubbery to ductile plastic to brittle over a range of temperatures and strain rates [13]. Depending on the strain rate, various methods can be used, from specialized electromechanical testing systems to Split Hopkinson bar or Taylor impact testing. Capturing polymer and polymer–matrix composite behavior across 10⁻³–10³ s⁻¹ demands complementary rigs and careful signal conditioning. In Split Hopkinson bars, the core difficulty is low acoustic impedance and viscoelasticity of polymers, which attenuate the transmitted wave and distort it via dispersion. Recent solutions combine low-impedance polymer bars (PMMA/PC), viscoelastic dispersion/attenuation corrections, and tailored pulse shaping, yielding reliable tension data on very soft or thin sheet-like materials [14]. Digital-twin layouts and novel tension bars with viscoelastic bars amplify signals for low-impedance specimens, while quartz/semiconductor transducers boost SNR and high-speed DIC or even flash X-ray corroborate kinematics and equilibrium [15,16]. At intermediate rates (≈1–400 s⁻¹), high-speed servo-hydraulic or electromechanical machines face actuator inertia, system ringing, and the need to reach constant velocity before load application; current designs pre-accelerate the actuator and engage the specimen at speed, adopt signal-restoration algorithms to remove ringing, and use damping in load paths approaches validated on polymers and other rate-sensitive materials [17,18]. Even though assessment of high strain rate mechanical performance of polymer and polymer-based composite materials is challenging, these recent developments in methodologies and techniques demonstrate their feasibility.

Bonded and hybrid (metal–composite) joints, reveal a consistent rate-dependent strengthening, often accompanied by a shift toward more brittle, interfacial, or polymer-dominated failure, as documented in recent reviews [19], and targeted experiments on adhesively bonded Al–polymer and composite–metal systems under SHPB tension and impact modes [11,20]. For joints made by friction stir-based processes, most published work still emphasizes quasi-static thermo-mechanical maps and processing–microstructure–performance links, while high-rate tensile data remain scarce; impact studies on friction spot/-stir hybrids nevertheless indicate substantial energy absorption and residual strength sensitivity to impact side and interfacial morphology [8,21]. Overall, the state of the art indicates that (i) strength and stiffness metrics generally increase with rate while (ii) failure loci migrate toward polymer/interphase control, and (iii) interfacial texturing and thermal windows from friction stir processing strongly mediate the rate response [21,22,23].

Although friction stir-based routes now enable continuous metal–polymer lap joints, the literature remains dominated by quasi-static thermo-mechanical maps and failure-mechanism studies, with little to no strain-rate-resolved tensile data on continuous lap geometries. Recent work on AA6082–polymer friction stir composite joints rigorously charts processing windows, temperature effects, and failure loci under quasi-static and fatigue loading, but does not map joint-level strength/stiffness and failure migration across 10⁻³–10² s⁻¹ in tension, nor does it couple such tests with full-field kinematics to quantify secondary bending typical of single-lap coupons [9,21]. Beyond these systems, studies on continuous linear friction stir/extrusion-type joints likewise report process, microstructure, performance at low rates only, leaving dynamic tensile behavior largely unaddressed [24]. In contrast, the high-strain-rate joint literature is richer for adhesively bonded or hybrid metal–polymer specimens, where SHPB and related methods show rate-dependent strengthening and polymer/interphase-controlled failure, but these findings cannot be assumed to transfer to friction stir interfaces, whose meso-texturing and thermal history dominate load transfer [11,22]. Methodological advances now make joint-level dynamic tension feasible, yet have seldom been applied to continuous friction stir metal–polymer lap joints.

To fill the research gap, the present work investigates quasi-static to high strain rate tensile behavior of AA6082-T6/Noryl™ GFN2 friction stir lap joints alongside their base materials, using digital image correlation and accelerated tensile testing equipment to quantify ultimate response and secondary-bending metrics. Building on recent thermo-mechanical and failure-mechanism maps for these systems, we test the hypothesis that joint strength and bending-stiffness metrics increase with strain rate, but that failure remains polymer-controlled.

2. Materials and Methods

A 2 mm thick aluminum alloy AA6082-T6 was friction stir-joined in lap configuration with a 5 mm thick glass-fiber-reinforced polymer, Noryl^® GFN2. Noryl^® GFN2 is a blend of Poly(phenylene ether) (PPE) and high-impact Polystyrene (PS), with undisclosed proportion, reinforced with 20% weight of short glass fiber. The thermo-mechanical properties of these materials were further discussed by the authors in a previous study [8].

The joints were produced with a 40 mm overlap with the aluminum alloy on the top side of the joint and both plates are 200 × 125 mm. A custom-built friction stir welding machine (shown in Figure 1a) equipped with a tool composed of a flat scrolled shoulder and a threaded cylindrical probe, with 16 and 5 mm of diameter, respectively, was used to join the plates in position control. The main process parameters used are listed in

Table 1. Friction stir joining process parameters.

Dwell Time (s)	$\mathsf{\omega }$ (RPM)	v (mm/s)	$\mathsf{\alpha }$ (°)	Pin Length (mm)	Plunge Depth (mm)
15	1000	2.33	2	2	2.2

. The joint schematic including base plate positioning and processing direction is found in Figure 1b. The process parameters and tool design were chosen according to previous research of the work group where these resulted in the highest performing joints [21].

Upon joining, the plates were cut perpendicularly to the joining path into 25 mm wide specimens. Base material specimens were also extracted from the base material plates following ASTM E8M [25] and ASTM D638 standards [26]. Base material and joint specimens were then tensile loaded in quasi-static (QS) and intermediate strain rate (ISR) conditions. For QS tensile testing, a Shimadzu (Kyoto, Japan) AGX-V series equipped with a 50 kN load cell (Figure 2a) was used at 5 mm/min. ISR testing was conducted in a custom electromechanical accelerated setup with a 25 kN load cell (Figure 2b), enabling pre-accelerating the actuator and engaging the specimen at 800 mm/s. In the case of the joints in both loading conditions, the distance between clamps was 100 mm.

Digital image correlation (DIC) was used to assess the strain fields within the failure zone for all experiments with Istra4D^® and Vic-2D System^® for QS and ISR tests, respectively. In the case of ISR testing, given the testing speed, a Photron Fastcam Nova^® high-speed camera was used to acquire the images during testing. A comprehensive analysis of the macro- and microstructure of the joints achieved with the given process parameters and tool, including the aluminum–polymer interface, was conducted and presented in [21]. The joint width was measured at approximately 17.3 mm, as shown in Figure 3, which is 1.3 mm wider than the diameter of the tool shoulder. As previously discussed, the joints bonding is mainly governed by macro- and micro-mechanical interlocking features within joints’ interfaces, while the chemical bonding plays a secondary role [27].

3. Results and Discussion

In this section, the results of tensile testing at both strain rates of the base materials and joints are presented and discussed, followed by an assessment of the DIC data, namely on the effect of strain rate in the out-of-plane bending, which affects the ratio between peel and shear loading in the joint.

3.1. Joint and Base Material Characterization with Strain Rate

Tensile tests were performed on ten specimens of each base material, with half being tested in quasi-static and the rest at intermediate strain rate conditions. Figure 4 shows a resulting representative stress vs. strain plot for each base material in both quasi-static and intermediate strain rate condition. It is observable that both materials present positive strain rate sensitivity regarding ultimate tensile strength, with the aluminum alloy increasing 12.27% from 348.62 MPa at 1.33 × 10⁻³ s⁻¹ to 391.40 MPa at 48.61 s⁻¹ and the glass-fiber-reinforced polymer increasing 31.44% from 48.57 MPa at 7.5 × 10⁻⁴ s⁻¹ to 63.84 MPa at 14.67 s⁻¹ (average values).

Even though ultimate strain in the aluminum alloy increased with increasing strain rate, it is possible that such effect might instead be due to the challenge of simultaneously measuring load and displacement in accelerated loading conditions, as the dynamic phenomenon results in vibrations in the setup which were not possible to be fully mitigated, leading to a higher degree of uncertainty in the strain at failure measurement. This phenomenon is of particular importance in the aluminum specimens given the high stiffness of these specimens when compared to the glass-reinforced polymer specimens, resulting in higher amplitude vibration.

Noryl^® GFN2 behavior at quasi-static conditions presents some softening before catastrophic failure takes place, while at intermediate strain rate conditions, the specimen shows a stable and linear loading response until it reaches the ultimate tensile strength, after which the value rapidly decreases until failure takes place.

Both results reached for the metallic and polymeric specimens are in line with results from different studies where the enhancement of the brittle behavior of the polymers and the slight differences in the aluminum behavior across different strain rates were also evaluated [28,29,30,31,32].

Having assessed the strain rate sensitivity of the base materials, the mechanical response of the dissimilar joints when loaded in single-lap tensile testing was then assessed. Figure 4 shows a representative remote stress vs. strain of friction stir-joined specimens under the different test conditions, QS and ISR. To allow for easier benchmarking of the joint’s performance, the remote stress was calculated based on the cross-section area of the glass-reinforced polymer (5 × 25 mm) as in [8]. Representative remote stress vs. strain curves of the joints in quasi-static and intermediate strain rate are plotted in Figure 5. A significant increase in strength is accompanied by a decrease in ductility of the joint, as it becomes significantly stiffer with increasing strain rate. On average, there was an increase of 29.87% in strength, even with a more modest increase in strain rate, from 4.9 × 10⁻⁴ s⁻¹ to 3.27 s⁻¹.

Figure 6 shows the failure location in both quasi-static tensile testing and intermediate strain rate, where it is possible to observe a consistent failure in both strain rates within the polymer side at the edge of the friction stir-processed zone, similar to previous results in the literature for quasi-static single-lap joint tensile testing [8,9,21].

Figure 7 summarizes the results for both base materials and joints for both loading rates, regarding ultimate strength and maximum strain. Apart from the aluminum alloy, it is observed that with increasing strain rate, there is an increase in strength accompanied by a decrease in ductility.

3.2. Out-of-Plane Assessment with Strain Rate

The joints exhibit a higher strain rate sensitivity than the base materials, as is observed in Figure 8, where the linear trend for strength with strain rate exhibits a slope sensibly double of the exhibited by the glass-reinforced polymer, which is the more strain-rate-sensitive base material. Even though accelerated tensile loading was conducted at the same cross-rate speed of 800 mm/s, differences in specimen design led to differences in strain rate during testing.

The joints failure location was consistent between strain rates, as reported in [8], although when analyzing the DIC strain field and measuring joint rotation due to geometric misalignment of the single-lap configuration, differences were observed in out-of-plane bending stiffness from quasi-static to intermediated strain rate loading, as shown in Figure 9.

Not only did higher strain rate loading lead to higher stiffness, but also the stiffness remained linear up to fracture, while in quasi-static loading, there was an increase in stiffness as the joints approached failure. This behavior may be due to the more brittle behavior of the polymer at higher strain rates, which, although increases strength, limits deformation. This results in more shear and less peel in the load mix at higher strain rate conditions. Given that peel loading is more disadvantageous from a joint strength point of view, as it loads the joint edges as cracks in mode I, it may result in significantly increased strength with strain rate. Figure 10 presents this relationship between joint strength and out-of-plane bending by comparing the initial stiffness (initial linear section in the case of quasi-static loading) and joint strength in both load rates.

4. Conclusions

The strain rate sensitivity of glass-reinforced polymer and aluminum friction stir lap joints was assessed in the present research work. This experimental study resulted in the following findings:

• Dissimilar polymer—aluminum friction stir lap joints have a positive strength strain rate sensitivity. This increase in strength is accompanied by an increase in joint stiffness and decrease in ductility. The average remote stress increased by ≈30% and stiffness ≈23%;
• The glass-fiber-reinforced polymer has a higher strain rate sensitivity than the aluminum alloy, but the dissimilar joints have an even higher strain rate sensitivity;
• Failure remains polymer-controlled at both rates, occurring in the Noryl™ adhered near the processed zone edge; higher rates shift the load mix toward more shear/less peel, consistent with reduced out-of-plane rotation and a more brittle polymer response; the nonplanar bending of the joints differed from quasi-static to intermediate strain rate. A higher bending angle stiffness was found for higher strain rates, requiring higher values of joint load or joint displacement to achieve similar values of out-of-plane bending angle;
• The bending angle increased linearly with joint loading at an intermediate strain rate, while in quasi-static loading, a rapid increase was found when approaching joint failure;
• As design implication for friction stir metal–polymer lap joints, increased rate can raise strength/stiffness but narrows the deformation margin; joint performance envelopes are therefore governed by the polymer constituent and by secondary-bending kinematics.