TIG and MIG melt aluminium; friction stir welding does not. Because the joint is forged below the melting point, the defects that come from a melt pool cannot form: no hydrogen porosity, no solidification cracking, no shrinkage. For repeating production welds in aluminium, friction stir welding typically delivers higher joint strength, far less distortion, several times the travel speed, and a machine-run process that does not depend on a scarce certified aluminium welder. TIG and MIG remain the better choice for one-offs, complex three-dimensional access, on-site repair, and joints that cannot react the process forces.
Why do TIG and MIG struggle with aluminium?
Aluminium is one of the harder metals to fusion weld well. It carries a tenacious oxide, conducts heat roughly five times faster than steel, and dissolves hydrogen when molten, which comes back out as porosity when the weld solidifies. In production that shows up as a familiar set of symptoms:
- Porosity and solidification cracking found at inspection, driving rework that was never in the schedule.
- Distortion and shrinkage from the high, localised heat of the arc, often requiring straightening, re-rolling or post-weld machining.
- A skills bottleneck. A good aluminium TIG welder is a craftsman, and every weld carries their day-to-day variability. Recruiting and certifying them is getting harder.
- Consumables: shielding gas, filler wire, fume extraction, UV protection — recurring cost the weld itself never shows.
None of these are execution failures. They are properties of any process that takes aluminium through liquid and back.
How does friction stir welding avoid these problems?
Friction stir welding, invented at The Welding Institute (TWI) in 1991, uses a rotating tool that generates frictional heat, softens but never melts the material, and forges the two sides together as it traverses the joint. The consequences follow from the physics:
- No melt pool, so no melt-pool defects. Shrinkage cracking, dendritic growth, elemental segregation and hydrogen porosity cannot occur in a weld that never melts.
- A forged, fine-grained weld. Joint strengths above 90% of parent ultimate tensile strength are typical in thin-section aluminium, with fatigue performance generally superior to fusion welds because there is no cast microstructure to initiate cracks.
- Low, localised heat input — a fraction of the distortion, which in many products removes the straightening operation entirely.
- A machine process, not a craft process. Once parameters are locked, the machine repeats them.
- Speed. Production TIG on aluminium typically runs at roughly 100-200 mm/min (around 130 mm/min is a useful mid-range figure). Friction stir welding in comparable thin aluminium commonly runs 750-1,500 mm/min, so three to ten times TIG; in favourable alloys and thicknesses, speeds above 1 m/min are routine.
- Single-pass thick sections. Arc processes need multiple passes, with inter-pass cleaning, as thickness grows. Friction stir welding welds thick aluminium in one pass; sections of tens of millimetres are established industrial practice, and single-pass welds up to 75 mm have been demonstrated.
- No consumables, low energy. No gas, no wire, no fume, no arc — and correspondingly lower energy per metre of weld.
| Criterion | TIG (GTAW) | MIG (GMAW) | Friction stir welding |
|---|---|---|---|
| Physical state | Fusion (melts) | Fusion (melts) | Solid state (no melting) |
| Porosity / hot cracking | Present | Present | Eliminated by mechanism |
| Typical joint efficiency (Al, thin section) | Roughly 60-80% | Roughly 60-80% | Above 90% of parent UTS |
| Distortion | High; usually needs correction | High; usually needs correction | Low; correction often eliminated |
| Travel speed (Al, production) | Slow, typically 100-200 mm/min | Moderate | Typically 3-10 times TIG |
| Thick sections | Multi-pass | Multi-pass | Single pass; tens of millimetres established |
| Skill dependency | Certified welder | Certified welder | Machine operator |
| Consumables | Gas plus filler | Gas plus wire | None (tool wear only) |
| Fume, UV, spatter | Yes | Yes | None |
| Best suited to | One-offs, repair, 3D access | General fabrication | Repeating production joints |
When is TIG or MIG still the right choice?
An honest comparison cuts both ways. Arc welding remains the better tool when:
- You are welding one-offs or prototypes with constantly changing geometry. Fixturing and parameter development pay back over repeating parts, not single pieces.
- Access is complex and three-dimensional in ways a rotating tool cannot reach. A hand-held torch goes places a machine spindle cannot.
- You need field or in-situ repair away from a machine.
- The joint cannot react the process forces. Friction stir welding needs rigid clamping and, for single-sided welds, backing support behind the joint line.
- The exit keyhole matters and cannot be designed out. Every weld ends with a keyhole where the tool retracts. It can be placed in scrap tabs, machined regions or fastener locations, or closed with retracting-pin tooling, but it must be planned for.
- Steels at low capital, where an existing TIG or MIG cell is already amortised and the metallurgy is not the problem.
Where has this substitution already happened?
This is not a new idea. It has two decades of production history, and the highest-profile case is public.
Rail carbodies
European and Japanese carbody manufacturers moved long aluminium panel welds from mechanised MIG to friction stir welding from the late 1990s (Hitachi, Alstom LHB, Sapa), eliminating the straightening and grinding that arc distortion forced on them. The substitution is documented in TWI’s published account of railway rolling-stock manufacturers implementing friction stir welding (2002).
NASA, then SpaceX — launch-vehicle tanks
NASA Marshall Space Flight Center began developing friction stir welding for aluminium-lithium cryogenic tanks in the mid-1990s, after fusion welding of Al-Li 2195 proved difficult. The process was implemented on the Space Shuttle Super Lightweight Tank: NASA’s own overview describes friction stir welds as stronger and easier to make than the earlier fusion-arc joints, with fewer process elements to control (NASA, Friction Stir Welding; NASA NTRS, FSW implementation on the External Tank). The first External Tank to fly with friction stir welds was ET-132 in 2009.
SpaceX took the same process into commercial production and has kept it there. The company’s own Falcon 9 description stated that the tank walls and domes are aluminium-lithium and that “SpaceX uses an all friction stir welded tank, the highest strength and most reliable welding technique available” (SpaceX Falcon 9 vehicle page, archived 2012). Independent shop-floor reporting in 2008 described SpaceX friction stir welding aluminium-lithium tank material from 1.6 mm to more than 12 mm, using a retractable-pin head of the same family used on the Shuttle external tank (The Fabricator, Rocket science, entrepreneur-style). NASA’s 2009 Spinoff account confirms SpaceX performing circumferential welds on its rockets with that equipment (NASA Spinoff 2009).
Falcon 9 first flew in 2010 and remains in operational service. That is more than fifteen years of production friction stir welding on flight hardware, which is the longest publicly documented commercial campaign of the process. StirLight has no commercial relationship with SpaceX; this is a public manufacturing case study, cited because it is the case a sceptical production engineer will already have heard of, and because it answers the “has anyone actually productionised this?” question without relying on unnamed customers.
Shipbuilding is the third established substitution: prefabricated friction stir welded aluminium panels have been supplied to yards since 1997 (Hydro/Sapa).
A StirLight-measured flatness figure
On an automotive battery-tray panel built from extrusions, StirLight achieved 0.5 mm flatness per square metre over a panel containing 4 metres of weld, with no post-weld straightening and no post-weld heat treatment. That is a StirLight measurement on that geometry, not a universal promise: thin cold plates, tanks and structurals each move differently, which is why a feasibility trial measures distortion on your actual part.
Frequently asked questions
Is a friction stir weld as strong as the parent metal?
In thin-section aluminium, joint efficiencies above 90% of parent ultimate tensile strength are typical. Fatigue performance is generally superior to fusion welds because there is no porosity or cast microstructure to initiate cracks.
Can friction stir welding join the unweldable aluminium grades?
Yes. The 2xxx and 7xxx aerospace families, generally regarded as unweldable by fusion processes, are routinely friction stir welded. That is why launch-vehicle tanks could move off arc welding.
Does friction stir welding need shielding gas or filler wire?
Not for aluminium. That removes the cost of buying, storing and administering gas and wire, and the fume extraction that goes with arc welding.
Did SpaceX invent friction stir welding?
No. The Welding Institute patented the process in 1991. NASA developed it for cryogenic tanks in the 1990s. SpaceX adopted it for Falcon 9 tank production and has run it in series since the vehicle entered service. The credit is for long-term production use, not invention.
What is a sensible first step?
A structured feasibility trial on your geometry: parameter development plus welded samples you can test.
Next step A structured feasibility trial answers the question on your geometry, with measurements rather than argument.