Friction Stir Welding: Process Physics, Microstructure Zones, and Applications
Friction stir welding (FSW), invented at The Welding Institute (TWI) in 1991, is the most significant advance in metal joining technology of the past three decades. By keeping the workpiece below its melting point throughout the entire process, FSW sidesteps the solidification-related defects that limit fusion welding of high-strength aluminium alloys, and delivers joints with fatigue performance, residual stress profiles, and dimensional stability that fusion welding cannot match. This article covers the full physics of the process — tool geometry, heat generation, material flow, and the four-zone microstructural model — followed by quantitative process parameter guidance, defect classification, materials applicability, and the industrial sectors where FSW has displaced or supplements conventional fusion welding.
Key Takeaways
- FSW is a solid-state process: peak weld temperatures are 70–90% of the solidus temperature in absolute Kelvin, avoiding melting, solidification cracking, porosity, and loss of alloying elements by evaporation.
- The rotating FSW tool generates heat by two mechanisms: frictional sliding between the tool shoulder and workpiece surface, and viscoplastic energy dissipation in the plasticised material around the pin.
- Four distinct microstructural zones are created: the weld nugget (fine dynamically recrystallised grains), TMAZ (deformed but unrecrystallised), HAZ (thermally affected only), and unaffected base metal.
- In precipitation-hardened alloys (2xxx, 7xxx series), the HAZ is the weakest zone due to over-ageing of strengthening precipitates, producing a characteristic W-shaped hardness profile; joint efficiency is typically 70–90% of base metal UTS.
- The ratio of tool rotation speed to traverse speed (ω/v, RPM per mm/min) is the primary composite parameter governing heat input, material flow, and defect formation.
- FSW is commercially deployed in aerospace (fuselage panels, launch vehicle tanks), shipbuilding (deck panels, extrusion joining), railway (floor panels), and automotive sectors, with all major aluminium-intensive manufacturers holding or licensing TWI FSW patents.
Invention, Principle, and the Solid-State Advantage
FSW was invented by Wayne Thomas at TWI (The Welding Institute, Cambridge, UK) and patented in December 1991 (International Patent PCT/GB92/02203). The original application was joining aluminium alloys; the fundamental insight was that the combination of frictional heating and mechanical stirring could achieve intimate metallurgical bonding without exceeding the material’s solidus temperature.
The solid-state constraint is the process’s defining advantage and its primary engineering motivation. Aluminium alloys of the 2xxx series (Al-Cu) and 7xxx series (Al-Zn-Mg-Cu) — which account for the bulk of aircraft primary structural material — are classified as non-weldable or difficult-to-weld by fusion processes. When melted and resolidified, these alloys form low-melting eutectic phases at grain boundaries (Al-Cu eutectic at 548 °C, Al-Zn-Mg phases below 500 °C), leading to solidification hot cracking under the tensile stresses of weld contraction. Gas porosity from dissolved hydrogen rejected during solidification is a further problem, requiring expensive NDT and repair operations in fusion-welded aerospace structures. FSW eliminates both failure modes because the material is never liquid.
The comparison with fusion welding is most relevant to the HAZ microstructure discussion: while fusion welding creates a fully melted weld pool and a classical coarse-grain/fine-grain/intercritical HAZ structure, FSW creates a thermomechanically processed zone (the nugget and TMAZ) surrounded by a purely thermal HAZ. The absence of a fusion zone and its associated segregation, solidification structure, and columnar grain growth is FSW’s most consequential metallurgical characteristic.
Heat Generation: Friction, Viscoplastic Dissipation, and Temperature
FSW heat generation has two distinct physical origins that operate simultaneously and whose relative contributions depend on process stage and local conditions.
Frictional Heat (Shoulder-Dominated)
The primary heat source during steady-state FSW is the sliding friction between the rotating tool shoulder and the workpiece surface. The heat generated per unit time at the shoulder contact annulus is:
Shoulder frictional heat generation rate:
Q_shoulder = (2/3) × π × μ × P × ω × (R_s³ − R_p³)
where:
μ = friction coefficient (0.3–0.5 for Al alloys, tool-material pair)
P = axial pressure (MPa) = F_axial / A_shoulder
ω = angular velocity (rad/s) = 2πN/60 (N = RPM)
R_s = shoulder radius (m)
R_p = pin root radius (m)
Typical values for 6061-T6 aluminium FSW:
N = 1000 RPM → ω = 104.7 rad/s
R_s = 9 mm, R_p = 3 mm
P = 50 MPa
μ = 0.4
Q_shoulder ≈ 1,100–1,800 W
Viscoplastic Dissipation (Pin-Dominated)
In the plasticised material immediately surrounding the pin (where shear stress is high and material is flowing), viscoplastic energy dissipation contributes additional heat:
Viscoplastic dissipation rate (simplified):
Q_viscoplastic = τ × γ̇ × V_deformation_zone
where:
τ = flow stress of plasticised material at peak temperature
(typically 30–80 MPa for Al alloys at 400–500°C)
γ̇ = strain rate in deformation zone (10¹–10³ s⁻¹)
V_deformation_zone = volume of plasticised material
Total heat generation per unit length of weld (Pseudo-Heat-Input):
PHI = (M_torque × ω) / v_traverse
where:
M_torque = measured spindle torque (N·m)
ω = angular velocity (rad/s)
v_traverse = traverse speed (m/s)
PHI units = J/m (analogous to kJ/mm in fusion welding)
Peak Temperature and the Solid-State Constraint
Peak temperatures in the weld nugget of aluminium alloys during FSW fall in the range 0.7–0.9 Tsolidus (expressed in absolute temperature, Kelvin). For 6061-T6 (solidus 582 °C = 855 K), this corresponds to approximately 330–490 °C. For 2024-T3 (solidus 502 °C = 775 K), peak nugget temperatures are typically 350–450 °C. The temperature never exceeds the solidus because above this point the material loses shear resistance and the thermomechanical coupling that drives the process breaks down — the tool would effectively lose grip on the workpiece material.
Direct thermocouple measurement in the weld nugget is impractical during production FSW but is used in research via pin thermocouples in specially prepared workpieces. In production, tool temperature is inferred from spindle torque and power draw (higher temperature = lower flow stress = lower torque), and peak temperature estimates are made from microstructural analysis of the nugget zone using precipitate dissolution maps calibrated by isothermal annealing experiments on the base alloy.
Tool Geometry and Design
The FSW tool is the critical component of the process: its shoulder diameter, pin profile, and material determine heat generation, material flow pattern, nugget shape, and tool life. Tool design is the subject of extensive proprietary development and academic research.
Shoulder Geometry
The shoulder serves two functions: generating frictional heat and containing the plasticised material beneath it (acting as a forging surface). Shoulder geometries include:
- Flat shoulder: Simplest geometry; effective for flat plate butt welds; material tends to expel at edges (flash) if downforce is excessive.
- Concave (scrolled) shoulder: The slight concavity (typically 3–6°) drives material inward toward the pin, improving material containment and reducing flash. Most common in production tooling for aluminium.
- Scrolled shoulder: Spiral grooves on the shoulder face actively pump material inward; allows lower tilt angle (even 0°) while maintaining material containment; preferred for robotic and complex-geometry welding where variable tilt is problematic.
Pin (Probe) Geometry
The pin geometry determines how material flows around it and fills the void left as the pin advances. Pin designs have evolved from simple cylinders to complex profiles offering enhanced material pumping and reduced defect susceptibility:
| Pin Profile | Description | Advantages | Typical Application |
|---|---|---|---|
| Cylindrical, smooth | Plain cylinder, no features | Simple; long tool life | Research; thin sheet |
| Threaded cylinder | Right- or left-hand thread on pin body | Axial material transport; good nugget consolidation | General aluminium FSW (<10 mm thick) |
| Tapered (conical) | Pin diameter reduces from root to tip | Easier plunge; lower axial force | Thick plate, difficult materials |
| Trivex / MX Triflute (TWI) | Asymmetric 3-flat or 3-fluted profile | Greatly increased material flow; reduced traverse force; better weld quality at higher speeds | High-productivity aluminium; aerospace |
| Whorl / Re-stir (TWI) | Tapered with helical flutes | Improved downward material flow for thick section; used in Bobbin tool variant | Thick plate (>15 mm); shipbuilding |
| Bobbin tool (self-reacting) | Upper and lower shoulders; no backing anvil required | Eliminates backing bar; enables one-sided access; symmetric weld profile | Hollow structures; tank sections; pipes |
Tool Materials
For aluminium alloys, the tool can be made from tool steel (H13, D2), W-based alloys, or PCBN (polycrystalline cubic boron nitride). Tool steel is adequate for aluminium FSW (softening temperature of tool steel well above Al plasticising temperature of 350–500 °C). For harder materials:
| Workpiece material | Peak weld T (°C) | Tool material options | Key challenge |
|---|---|---|---|
| Aluminium alloys (all series) | 300–530 | H13 tool steel, MP159, carbide | Minimal; tool life >1,000 m |
| Magnesium alloys | 300–440 | H13, M2 tool steel | Oxidation; volatile Mg fumes |
| Copper alloys | 700–900 | W-25Re, PCBN, Si&sub3;N&sub4; | High flow stress; rapid tool wear |
| Titanium alloys | 700–950 | W-Re, PCBN, Si&sub3;N&sub4; | Chemical reactivity with tool; extreme forces |
| Austenitic stainless steel | 800–1100 | PCBN, WC-Co, Si&sub3;N&sub4; | Extreme wear; limited tool life (<5 m per tool) |
| Low-carbon steel / structural steel | 800–1000 | PCBN, W-Re, Mo-based | Tool cost; thermal management |
Microstructural Zones: The Threadgill Classification
The four-zone microstructural model for FSW welds was formalised by Threadgill (1997) and remains the standard framework for describing FSW cross-sections. Each zone is distinguished by its thermal history, strain history, and resulting microstructure, and each contributes differently to joint mechanical properties.
Zone 1: Weld Nugget (Stir Zone)
The nugget is the region directly processed by the rotating pin. It experiences the highest temperatures (0.8–0.9 Tm) and the highest plastic strains (effective strain >10–40). The microstructural outcome is dynamic recrystallisation (DRX), producing a fine, equiaxed grain structure with high-angle grain boundaries. Nugget grain sizes in aluminium FSW are typically 2–10 µm, compared with 30–100 µm in the wrought base material.
In precipitation-hardened alloys, the high nugget temperature dissolves most of the strengthening precipitates (GP zones, η′, T-phase). On cooling, precipitation of coarser precipitates occurs in the nugget; the strengthening contribution from fine coherent precipitates is lost unless a post-weld ageing treatment is applied. The nugget hardness in 7075-T6 FSW is typically 160–180 HV, compared with 175–190 HV base metal — a moderate reduction offset by the Hall-Petch contribution of the fine grain size.
The nugget also exhibits an onion-ring pattern in macro-etched cross-sections of aluminium welds — concentric elliptical bands of slightly different grain size and precipitate density corresponding to successive tool rotations depositing material as the tool traverses. This characteristic banding is the metallographic signature of FSW and distinguishes it from all other joining processes.
Zone 2: Thermomechanically Affected Zone (TMAZ)
The TMAZ immediately surrounds the nugget and is characterised by severe plastic deformation without complete dynamic recrystallisation. The grain structure is elongated and rotated in the direction of tool rotation, with grains swept into characteristic curved morphologies. Sub-grain boundaries and a high dislocation density are present. In 2xxx and 7xxx aluminium alloys, precipitate coarsening in the TMAZ is more pronounced than in the nugget because: (a) temperatures are high enough for diffusion but not enough for full dissolution and re-precipitation, and (b) the deformation accelerates diffusion. The TMAZ/nugget boundary on the advancing side is very sharp — sometimes a single grain boundary — while the TMAZ/nugget boundary on the retreating side is more gradual.
Zone 3: Heat-Affected Zone (HAZ)
The HAZ in FSW is analogous to the HAZ in fusion welding in that it experiences only thermal exposure with no significant plastic deformation. In solid-solution-strengthened alloys (5xxx series), the HAZ shows grain growth and slight softening. In precipitation-hardened alloys (2xxx, 7xxx, 6xxx series), the HAZ is the critical zone because:
- Temperatures in the range 150–400 °C accelerate precipitate coarsening (over-ageing) of the fine coherent strengthening precipitates, reducing their strengthening contribution.
- In 7075-T6, the MgZn2 (η-prime) coherent precipitates coarsen to incoherent η phase in the HAZ, reducing local hardness to 140–160 HV against 175–190 HV base metal.
- The minimum hardness location in the FSW cross-section is in the HAZ, typically 5–15 mm from the weld centreline, and this is the failure location under tensile loading of the joint.
This HAZ hardness minimum is the origin of the W-shaped hardness profile characteristic of FSW in precipitation-hardened aluminium alloys: high hardness in base metal, lower in HAZ (both sides), slightly higher in TMAZ and nugget. The HAZ concept in FSW parallels the heat-affected zone in fusion welding, though the microstructural changes are precipitate-related rather than transformation-related in aluminium alloys.
Zone 4: Unaffected Base Metal
Beyond the HAZ, the material is thermally unaffected by the welding process. Properties, grain structure, crystallographic texture, and precipitate state are unchanged from the original temper. In rolled plate, this region retains the characteristic pancake grain structure and rolling texture.
Process Parameters and Their Effects
Rotation Speed and Traverse Speed
The dominant parameter governing FSW thermal input is the ratio of rotation speed (N, RPM) to traverse speed (v, mm/min), often written as N/v or ω/v. This ratio determines how much energy is deposited per unit length of weld:
Heat index (simplified):
HI = N / v (RPM / mm·min⁻¹)
Higher HI → more heat input per unit weld length
→ higher peak temperature
→ greater precipitate dissolution (nugget) and over-ageing (HAZ)
→ larger softened zone width
→ potential for surface defects (excess flash)
Lower HI → less heat, lower peak temperature
→ higher residual dislocation density in nugget
→ risk of tunnel defect / wormhole if insufficient plasticisation
Typical ranges for 6xxx aluminium (6 mm thick plate):
N: 600–1500 RPM
v: 150–500 mm/min
HI: 2–8 RPM·min/mm (optimal ≈ 3–5 for 6061)
For higher-strength 7xxx alloys:
Lower HI preferred (2–4) to minimise HAZ over-ageing
Lower v (100–300 mm/min) to allow adequate plasticisation
Axial Force and Plunge Depth
The axial (downward) force maintains contact between the tool shoulder and workpiece, governing the frictional heat generation rate and the forging pressure that consolidates the weld nugget. FSW machines operate in two control modes:
- Force control: The machine servo maintains constant axial force (typically 5–30 kN for 6 mm aluminium plate). Preferred for flat plate welding where workpiece flatness is controlled.
- Position (plunge depth) control: The shoulder is held at a fixed depth below the original workpiece surface. More robust to workpiece thickness variation and distortion; preferred for production and robotic FSW.
Tilt Angle
The FSW tool is typically tilted 1–3° from vertical so that the trailing edge of the shoulder contacts the workpiece more aggressively than the leading edge. This trailing tilt drives plasticised material rearward into the wake of the pin, improving consolidation and surface quality. At 0° tilt (required for some robotic and curved-joint applications), the shoulder must be actively scrolled to maintain material containment.
| Parameter | Typical range (Al alloys) | Effect if too high | Effect if too low |
|---|---|---|---|
| Rotation speed (RPM) | 400–2000 | Excess heat; flash; surface galling; HAZ over-ageing | Insufficient plasticisation; tunnel defect; excess traverse force |
| Traverse speed (mm/min) | 50–600 | Deficiency of heat per unit length; wormhole; cold lap | Excessive heat input; over-ageing; coarse nugget grain; surface flash |
| Axial force / plunge depth (kN / mm) | 5–30 kN / 0.1–0.3 mm | Excessive flash; tool damage; workpiece thinning | Insufficient shoulder contact; kissing bond; low heat |
| Tilt angle (°) | 1–3 | Increased trailing edge wear; surface gouging | Flash; poor consolidation; leading-edge lifting |
| Pin length (mm) | Plate thickness − 0.1 to 0.3 | Pin contact with backing bar; tool damage | Root flaw (incomplete penetration); unbonded root region |
FSW Defects: Classification, Causes, and Detection
FSW is generally a low-defect process relative to fusion welding, but a distinct set of defect types can occur when process parameters deviate from the qualified window. Each defect has a characteristic morphology, cause, and detection method.
Wormhole (Tunnel Defect)
The wormhole is a continuous or intermittent void running along the advancing side of the weld, at or near the pin tip level. It forms when insufficient plasticised material is transported from the front to the rear of the pin to fill the void left by pin advance. Cause: excessive traverse speed, insufficient rotation speed, or insufficient axial force. Detection: radiography (RT) or phased-array ultrasonic testing (PAUT). Prevention: reduce N/v ratio by increasing v or decreasing N.
Kissing Bond (Cold Lap / Oxide Entrapment)
The kissing bond is the most insidious FSW defect because it can appear sound on visual inspection and conventional RT while severely reducing fatigue strength. It forms when the surface oxide film (Al2O3) on the original faying surfaces is not fully dispersed by the stirring action, leaving a planar, oxide-contaminated, partially bonded interface in the nugget, typically on the advancing side. Cause: insufficient deformation at the joint line; joint fit-up gaps; excessive traverse speed. Detection: PAUT with focused beam at advancing side; metallographic cross-section. Prevention: pre-weld brushing of faying surfaces; optimising N/v ratio.
A kissing bond that passes RT inspection can reduce fatigue life by 50–80% relative to a defect-free FSW joint, because the oxide interface acts as a pre-existing crack under cyclic loading. PAUT and time-of-flight diffraction (TOFD) are required to reliably detect kissing bonds in safety-critical FSW joints (aerospace, railway, pressure vessels).
Root Flaw
An incomplete penetration defect at the weld root caused by a pin that is too short to reach the full thickness of the joint. The root region remains unprocessed and unbonded. Unlike fusion welding root flaws (lack of fusion), the FSW root flaw is mechanically very sharp and highly detrimental to fatigue performance. Pin length must be set to within ±0.1 mm of the nominal value; tool length is verified before each production run using a calibrated height gauge.
Flash and Surface Irregularity
Flash is material expelled from under the shoulder due to excessive heat input or axial force. A small amount of flash is normal; excessive flash indicates over-heating, results in weld thinning, and can entrain oxide into the weld surface. It is removed by post-weld machining in applications where flush surface is required (aerospace skin panels).
Joint Line Remnant (Lazy S)
Surface oxide from the original joint line that is entrained and redistributed into the nugget, visible in macro-etched sections as a sinuous line of dispersed oxide particles following the material flow path. Distinct from the kissing bond in that the material is bonded across the oxide dispersion. Reduced by: pre-weld wire brushing, planing, or chemical cleaning of faying surfaces; optimising material flow parameters.
Materials: Applicability and Joint Efficiency
Aluminium Alloys
FSW is most mature and commercially deployed for aluminium alloys. All wrought alloy series (1xxx through 8xxx) have been successfully FSW’d in research; the most industrially significant are:
| Alloy series | Strengthening mechanism | FSW weldability | Joint efficiency (%UTS) | Notes |
|---|---|---|---|---|
| 1xxx (pure Al) | Work hardening | Excellent | 95–100% | Soft; high ductility; no precipitate effects |
| 2xxx (Al-Cu) | Precipitation (GP, θ′) | Very good | 75–90% | Fusion-unweldable; FSW enables joining 2024, 2219 |
| 5xxx (Al-Mg) | Solid solution + work hardening | Excellent | 90–100% | Shipbuilding; marine; no precipitate HAZ softening |
| 6xxx (Al-Mg-Si) | Precipitation (β′, β) | Excellent | 80–95% | Automotive, railway; post-weld age restores HAZ |
| 7xxx (Al-Zn-Mg-Cu) | Precipitation (η′, T-phase) | Good | 70–88% | Aerospace; HAZ softening limits efficiency; PWAA improves |
| Dissimilar Al-Al | Combined | Good–moderate | 60–85% | Composition gradient in nugget; TMAZ mixing |
The 5xxx series aluminium alloys, which are solid-solution strengthened rather than precipitation hardened, show essentially no HAZ softening because there are no coherent precipitates to coarsen. This makes them ideal FSW candidates with joint efficiencies approaching 100% — a major advantage in marine and naval structures where 5083, 5383, and 5059 alloys are widely used. Compare with the grain boundary strengthening mechanisms in these alloys for context.
Post-Weld Artificial Ageing (PWAA)
For precipitation-hardened alloys, PWAA can recover some of the HAZ strength lost to over-ageing. After FSW, the weld is artificially aged at the standard ageing temperature for the alloy (e.g., 120 °C/24 h for 7075-T6). The nugget, where precipitates were dissolved, responds strongly to ageing and recovers to near-base-metal strength. The HAZ also recovers partially. Joint efficiency in 7050-T7451 can increase from approximately 75% as-welded to 85–90% after PWAA. For 6xxx alloys, PWAA at 160–180 °C/8 h typically restores joint efficiency to 90–95%.
Dissimilar Metal FSW
FSW is uniquely capable of joining combinations that are impossible or impractical by fusion welding, because no liquid phase is formed and the mixing of dissimilar metals occurs in the solid state. Key dissimilar systems under active development and deployment:
- Aluminium–Copper: Used in electrical conductor joining (battery tab connections in electric vehicles, busbar joints). The solid-state mixing avoids the brittle Al-Cu intermetallic phases (Al2Cu, Al4Cu9) that form in fusion welds and cause joint embrittlement.
- Aluminium–Steel: Critical for automotive mixed-material structures (steel body-in-white with aluminium panels or structural members). FSW produces a thin (<1 µm) Al-Fe intermetallic layer at the interface that is thin enough to maintain joint strength; fusion welding produces thick, brittle FeAl and Fe3Al layers that are catastrophic for joint ductility.
- Aluminium–Titanium: For aerospace structures where weight reduction requires joining 2xxx/7xxx airframe alloys to Ti-6Al-4V fastener regions or fittings. FSW avoids Ti oxidation (which requires shielding or vacuum in fusion processes) and produces fine-grained mixed zones.
Residual Stress Distribution
Residual stress in FSW joints is substantially lower in magnitude and more benign in distribution than in fusion welded joints of similar geometry. The absence of a liquid-solid phase transformation eliminates the dominant source of high tensile residual stress in fusion welds (volume change on solidification combined with plastic constraint during cooling). The FSW residual stress pattern arises from: differential thermal expansion and contraction of the thermomechanically processed zone versus the base metal; and plastic deformation by the pin and shoulder.
The characteristic FSW residual stress profile shows longitudinal tensile residual stress (in the welding direction) within the nugget and TMAZ — typically 30–80% of the material yield strength — transitioning to compressive stress in the base metal on either side. The maximum residual stress in 6mm 6061-T6 FSW is approximately 60–100 MPa longitudinal tension in the nugget, compared with 150–200 MPa (near yield) for MIG welds in the same material. This lower residual stress significantly improves fatigue life, because fatigue crack growth rate is a strong function of mean stress: the lower tensile residual stress in FSW joints reduces the effective R-ratio and extends life by a factor of 2–5 relative to fusion welds in the same alloy.
Industrial Applications
Aerospace and Space Launch
FSW was adopted first by the aerospace industry, driven by the ability to join 2xxx and 7xxx aluminium alloys that were previously riveted (mechanical fastening adds weight from fasteners and requires hole drilling that reduces fatigue life). Key deployments include:
- Eclipse 500 business jet: First production aircraft with FSW primary structure (fuselage panels, 2003). Over 1,500 m of FSW per aircraft replaced approximately 7,300 rivets.
- Boeing 747 Advanced: Fuselage stringer-to-skin panels joined by FSW in specific sections.
- NASA Space Shuttle External Tank: Self-reacting FSW (bobbin tool) used on 2219 aluminium alloy barrel sections. Replaced variable-polarity plasma arc welding, improving weld quality and reducing rework.
- SpaceX Falcon 9 and Falcon Heavy: LOX and kerosene tank sections in 2219 aluminium joined by FSW.
- United Launch Alliance Delta IV and Atlas V: Cryogenic tank manufacture using FSW on 2014 and 2195 lithium-aluminium alloys.
Shipbuilding and Marine
The shipbuilding industry was among the earliest commercial adopters of FSW for marine-grade aluminium (5xxx and 6xxx series) panel fabrication. FSW offers significant productivity advantages over MIG welding for long straight butt joints in deck panels, bulkheads, and superstructure panels. Advantages include: reduced distortion (lower heat input), improved corrosion resistance (no sensitisation risk with 5xxx alloys), and elimination of fume extraction requirements. Norwegian shipbuilder Sapa (now Hydro Extrusion) and Japanese shipbuilders have industrialised FSW panel production for high-speed ferries and naval vessels.
Automotive and Rail
In automotive manufacturing, FSW is applied to battery trays and enclosures for electric vehicles (joining 5xxx and 6xxx aluminium extrusions and castings into sealed, structurally efficient enclosures). The joining of aluminium-to-steel in tailored blanks (for body-in-white weight reduction) is an active application area. In rail, FSW produces lightweight aluminium floor panels for high-speed trains; Hitachi, Bombardier (now Alstom), and Kawasaki have all industrialised FSW rail floor panel production. The Shinkansen (Japanese bullet train) N700 series uses FSW floor panels.
Pressure Vessels and Piping
FSW has been qualified for pressure vessel manufacture under ASME standards (ASME Section IX QW-290 for friction welding covers FSW-related procedures). Applications include aluminium alloy pressure vessels for aerospace cryogenic service, copper heat exchanger tube sheets, and titanium alloy components for chemical plant. The absence of fusion zone porosity and hot cracking is particularly valuable for leak-tight pressure boundaries in cryogenic service.
FSW vs Fusion Welding: Quantitative Comparison
| Property / Characteristic | FSW | GMAW (MIG) | GTAW (TIG) |
|---|---|---|---|
| Peak process temperature | 70–90% Tsolidus (solid state) | >Tliquidus (liquid pool) | >Tliquidus (liquid pool) |
| Hot cracking risk (Al 2xxx, 7xxx) | None (no solidification) | High (solidification cracking) | High (solidification cracking) |
| Porosity | Essentially none in optimised process | Significant in Al; controlled by shielding | Low; requires gas purity control |
| Joint efficiency (7075-T6) | 70–88% UTS | Not applicable (unweldable) | Not applicable (unweldable) |
| Residual stress (longitudinal) | 30–80% YS (tension in nugget) | 80–100% YS (tension in HAZ) | 60–90% YS (tension in HAZ) |
| Distortion | Low (lower heat input) | Moderate–high | Moderate |
| Filler metal required | No (except some lap joint variants) | Yes | Usually yes |
| Shielding gas required | No (for Al alloys) | Yes | Yes |
| Skill requirement | Machine operator; high fixturing precision | Skilled welder or robot | Highly skilled welder or precision robot |
| Capital cost | High (purpose-built FSW machine) | Low–moderate | Moderate |
| Applicable joint types | Butt, lap, T-joint, corner (tooling-dependent); closed profiles with bobbin tool | Butt, lap, T, fillet, corner; all positions | Butt, lap, T, fillet; all positions |
| Position flexibility | Limited (typically flat; 5-axis machines expand range) | All positions | All positions |
FSW provides the clearest advantage when one or more of the following apply: the base material is a fusion-unweldable high-strength aluminium alloy (2024, 7075, 7050); the joint is a long straight butt or lap joint in flat or cylindrical geometry; fatigue performance is critical (lower residual stress, no porosity); or joint quality consistency must be demonstrated without 100% weld inspection (FSW is inherently more reproducible when machine parameters are held constant). For short welds, complex geometry, out-of-position joints, or non-aluminium materials where tool cost and wear are prohibitive, fusion welding remains the preferred process.
Standards and Qualification
FSW welding procedures are qualified under the following principal standards frameworks:
- ASME BPVC Section IX, QW-290: Covers friction welding (including FSW) procedure qualification for pressure vessels and piping. Essential variables include rotation speed, axial force, upset distance, and heat input parameters.
- AWS D17.3/D17.3M: Specification for friction stir welding of aluminium alloys for aerospace applications. Specifies weld class, inspection requirements, and procedure qualification testing including tensile, bend, and macrosection examination.
- ISO 25239 (Parts 1–5): International standard for FSW of aluminium, covering terms and definitions (Part 1), design of weld joints (Part 2), qualification of welding operators (Part 3), specification and qualification of welding procedures (Part 4), and quality and inspection requirements (Part 5).
- EN ISO 15614-14: European standard for welding procedure qualification for friction welding including FSW, applicable to pressure equipment under the PED.
NDT of FSW joints typically requires phased-array ultrasonic testing (PAUT) capable of detecting the root flaw and kissing bond defects that are not reliably detected by conventional radiography. ASTM E2376 and EN ISO 11666 provide frameworks for UT qualification of FSW joints.
Frequently Asked Questions
What is friction stir welding and how does it differ from fusion welding?
Friction stir welding (FSW) is a solid-state joining process invented at TWI in 1991. A non-consumable rotating tool — consisting of a shoulder and a pin — is plunged into the joint line between two workpieces. Frictional and viscoplastic heat (typically 70–90% of the solidus temperature in absolute Kelvin) softens the material without melting it. The rotating pin mechanically stirs the softened material across the joint line, forming a solid-state weld on cooling. Unlike fusion welding, FSW avoids all solidification-related problems: hot cracking, porosity, loss of alloying elements by evaporation, and the need for filler metal or shielding gas in most aluminium applications. These advantages are most pronounced for high-strength 2xxx and 7xxx aluminium alloys that are classified as non-weldable by fusion processes.
What are the four microstructural zones in a friction stir weld?
The four zones (Threadgill classification, 1997) are: (1) Weld Nugget (WN) / Stir Zone (SZ) — directly processed by the rotating pin; severe plastic deformation drives dynamic recrystallisation, producing fine equiaxed grains (typically 2–10 µm in aluminium alloys); in precipitation-hardened alloys, strengthening precipitates are dissolved. (2) Thermomechanically Affected Zone (TMAZ) — immediately outside the nugget; subjected to both heat and some deformation; grains are elongated and rotated but not fully recrystallised. (3) Heat-Affected Zone (HAZ) — beyond the TMAZ; thermally affected without plastic deformation; in precipitation-hardened alloys, strengthening precipitates coarsen and over-age, creating the hardness minimum (failure locus). (4) Unaffected Base Metal — properties, grain structure, and precipitate state unchanged from original temper.
What materials can be friction stir welded?
FSW was developed for aluminium alloys and is most widely deployed for all 2xxx, 5xxx, 6xxx, and 7xxx series alloys, including fusion-unweldable grades such as 2024 and 7075. The process has been extended to magnesium alloys, copper, copper alloys, titanium alloys (requiring PCBN or W-Re tools), stainless steel, structural steel, and dissimilar combinations (aluminium-to-copper for EV battery applications; aluminium-to-steel for automotive mixed-material structures). For harder materials, tool cost and wear are the primary limiting factors: PCBN tools for steel FSW currently last only a few metres of weld, making process economics challenging outside specialised aerospace and defence applications.
What defects can occur in friction stir welds and how are they prevented?
FSW defects include: (1) Wormhole (tunnel defect) — continuous void on the advancing side from insufficient material flow; prevented by reducing traverse speed or increasing rotation speed. (2) Kissing bond (cold lap) — oxide-contaminated partially bonded interface; the most dangerous defect for fatigue performance; prevented by pre-weld surface preparation and optimised N/v ratio. (3) Root flaw — incomplete penetration at weld root from under-length pin; controlled by accurate tool length setting and position control. (4) Flash — expelled material from excessive heat or force; controlled by reducing axial force or heat input. (5) Joint line remnant (Lazy S) — dispersed oxide particles from original faying surface; reduced by pre-weld cleaning and optimised stirring parameters. Phased-array UT (PAUT) is required to reliably detect kissing bonds and root flaws that are not visible on radiography.
How does FSW affect the strength of precipitation-hardened aluminium alloys?
In precipitation-hardened alloys (2xxx and 7xxx series), FSW creates a characteristic W-shaped hardness profile. The nugget loses coherent precipitates by dissolution at peak temperature but partially compensates through fine grain Hall-Petch strengthening (nugget hardness typically 160–175 HV for 7075-T6). The HAZ, experiencing moderate temperatures (150–350 °C) without deformation, shows over-ageing of MgZn2 (η′) precipitates to coarser incoherent η phase, reducing local hardness to approximately 140–155 HV. The HAZ minimum is the tensile failure locus, giving joint efficiencies of 70–88% UTS for 7075-T6. Post-weld artificial ageing (PWAA) at the standard alloy ageing temperature re-precipitates fine strengthening phases in the nugget and partially recovers HAZ strength, raising joint efficiency to 85–92%.
What is the difference between the advancing side and retreating side in FSW?
The advancing side (AS) is where tool rotation direction adds to the traverse velocity vector; the retreating side (RS) is where tool rotation opposes traverse. This creates an asymmetric thermomechanical condition: the AS experiences higher local strain rates and a sharper nugget/TMAZ microstructural boundary. Defects (wormhole, kissing bond) preferentially form on the AS because material transport is more critical there. The joint line oxide is swept predominantly to the AS. Tool rotation direction (clockwise vs counterclockwise) therefore determines which side of the joint bears the higher defect risk, and is an important variable in FSW procedure qualification for asymmetric joints (e.g., different alloys on each side, or when one side has restricted access for inspection).
What are the key process parameters in friction stir welding?
The five primary process parameters are: (1) Tool rotation speed (N, RPM) — governs frictional heat generation; higher N increases heat and material softening. (2) Traverse speed (v, mm/min) — governs heat input per unit length; higher v reduces heat input. (3) Axial force or plunge depth — determines shoulder contact pressure and forging consolidation of the nugget. (4) Tilt angle (1–3° trailing) — improves shoulder contact and material containment behind the pin. (5) Pin geometry and length — determines material flow pattern and root penetration. The composite N/v ratio (heat index) is the most useful single parameter for process development: optimal N/v must be established from macro/hardness/tensile testing for each alloy and thickness combination. Pseudo-heat-input (PHI = torque × omega / traverse speed) can be monitored from machine spindle data for production process control.
Why is FSW considered advantageous for aerospace aluminium structures?
FSW offers decisive advantages for aerospace aluminium structures: the primary structural alloys 2024-T3 and 7075-T6 are classified as fusion-unweldable due to solidification hot cracking, but can be joined by FSW with joint efficiencies of 70–90%. The absence of solidification eliminates porosity, hot cracking, and filler metal requirements; joint properties are highly repeatable from a controlled machine process. Residual stresses are 30–60% lower than fusion welds, directly improving fatigue life in cyclic-loaded primary structures. No consumables (filler, shielding gas) are needed for most Al FSW, reducing recurring cost. These advantages have driven adoption in Eclipse 500 fuselage panels (replacing 7,300 rivets), Space Shuttle and SpaceX Falcon launch vehicle cryogenic tanks, and numerous military aircraft and missile structures.
How is heat input calculated and controlled in friction stir welding?
FSW heat generation comes from shoulder friction (Q = 2/3 × π × μ × P × ω × (Rs3 − Rp3)) and viscoplastic dissipation in the plasticised material. In production, the pseudo-heat-input (PHI = measured torque × angular velocity / traverse speed, in J/m) is the preferred process control metric because it is directly measurable from machine spindle data without requiring friction coefficient or axial pressure inputs. Higher N/v ratio (RPM/mm/min) correlates with higher PHI and higher peak temperature. Optimal PHI is established during procedure qualification by mapping PHI against nugget soundness, HAZ softening width, and tensile properties. Process control in production monitors spindle torque (lower torque = higher temperature = potential overheating) and axial force or plunge depth, with alarm limits set at ±10% of qualified values.
Recommended Reference Books
Friction Stir Welding and Processing — Mishra & Mahoney (Eds.)
The definitive reference on FSW science and technology: process mechanics, microstructure, mechanical properties, modelling, and applications across aerospace, automotive, and marine sectors. Essential reading for FSW practitioners.
View on AmazonThe Friction Stir Welding of Aluminium Alloys — Leal, Loureiro & Roldo
Focused technical treatment of FSW microstructure and mechanical property development in 2xxx, 5xxx, 6xxx, and 7xxx series aluminium alloys, with extensive fractography, EBSD, and hardness mapping data.
View on AmazonASM Handbook Vol. 6A: Welding Fundamentals and Processes
Comprehensive ASM reference covering all welding processes including FSW, with chapters on solid-state welding, process selection, microstructure, NDT, and qualification standards for aerospace and industrial applications.
View on AmazonAluminium Alloys: Structure and Properties — Mondolfo
The authoritative encyclopaedic reference on aluminium alloy metallurgy: composition, phase diagrams, precipitation sequences, mechanical properties, and thermomechanical processing. Essential companion for understanding FSW microstructure in all alloy series.
View on AmazonDisclosure: MetallurgyZone participates in the Amazon Associates programme. If you purchase through these links, we may earn a small commission at no extra cost to you. This helps support free technical content on this site.
Further Reading & Related Topics
References
- Thomas, W.M. et al., “Friction Stir Butt Welding,” International Patent Application No. PCT/GB92/02203, 1991.
- Mishra, R.S. and Mahoney, M.W. (Eds.), Friction Stir Welding and Processing. ASM International, 2007.
- Threadgill, P.L., “Terminology in friction stir welding,” Science and Technology of Welding and Joining, 12(4), 357–360, 2007.
- Mishra, R.S. and Ma, Z.Y., “Friction stir welding and processing,” Materials Science and Engineering R, 50(1–2), 1–78, 2005.
- Nandan, R., DebRoy, T. and Bhadeshia, H.K.D.H., “Recent advances in friction-stir welding — process, weldment structure and properties,” Progress in Materials Science, 53(6), 980–1023, 2008.
- AWS D17.3/D17.3M: Specification for Friction Stir Welding of Aluminium Alloys for Aerospace Applications. American Welding Society.
- ISO 25239-1 through 25239-5: Friction Stir Welding — Aluminium. ISO Geneva.
- Threadgill, P.L. et al., “Friction stir welding of aluminium alloys,” International Materials Reviews, 54(2), 49–93, 2009.
