Since inventing friction stir welding (FSW) in 1991, TWI has continued to develop its potential. The technique continues to be advanced and refined for the joining of a various materials, as detailed on this page.
Friction stir welding was originally invented for welding aluminium, as some grades of aluminium are considered difficult to weld by existing arc welding techniques, and a few, such as the very high-strength 2XXX and 7XXX series of alloys, unweldable.
A significant problem arises from the volatility of many of the alloying elements used to strengthen the aluminium: elements such as lithium are easily oxidised or even burned off during fusion welding and thus upon solidification the weld zone is no longer the same grade of material as the parent.
Aluminium and its alloys also undergo significant volume changes (up to 4%) during the melting and resolidification processes. This can cause very significant distortion of the components being welded, and give rise to solidification cracking due to the stresses involved. Friction stir welding, being a solid-state, mechanical process, does not encounter these problems to the same degree and can produce sound, low-distortion welds in all aluminium alloys, wrought and cast.
There is a widespread perception that magnesium is difficult to weld using fusion techniques, so once FSW had been demonstrated to produce excellent, low-cost, high-integrity welds in aluminium, there was a great demand for the process to be adapted for welding magnesium. A development programme duly established that magnesium alloys can be welded with relative ease in both their wrought and cast forms.
TWI’s Yorkshire Technology Centre has developed the welding of 75mm-thick magnesium plates, 1.3m by 3m, to form the slip bed of large vibration test machines for TWI member company LDS Test and Measurement Ltd.
Copper and copper alloys
The high thermal and electrical conductivity of copper have long made it a difficult material to weld, particularly in thick sections. As FSW does not melt the workpiece and applies mechanically generated heating very locally, it was seen as a potentially useful process for welding copper. Early results on flat plate copper material were sufficiently encouraging for SKB, a Swedish organisation charged with developing containment canisters for nuclear waste, to commission the design and build of an experimental prototype FSW machine at TWI to weld circumferential parts representative of full-sized canisters.
This machine proved capable of welding 50mm-thick material and the experience gained was used to specify a bespoke machine to be installed in Oskarshamn for further tests. Twenty years of joint development with SKB has produced a FSW process that can meet the highly demanding requirements for the closure and sealing of the 5m-long, 1m-diameter, 50mm-wall-thickness, copper canisters that will provide the main physical barrier in the waste storage system.
Other applications in copper alloys include the welding of copper cooling plates for X-ray sputter targets by Hitachi and the fabrication of brass components, including propellers and impellers, for marine applications.
A number of nickel aluminium bronze propellers for naval applications have had their surfaces friction stir processed using robotic FSW systems. This was undertaken to reduce the incidence of fatigue cracking at the blade root but was found additionally to bring about desirable improvements in surface properties that reduced cavitation and noise.
Hafnium and zirconium
Hafnium and zirconium are metals with specialist applications in the petro-chemical, oil refining and power generation industries. Research has shown that both can be friction stir welded.
Inconel and superalloys
Inconel and other high-temperature allows were developed for use in the demanding hot stage applications of gas turbine engines. Welding Inconel alloys is difficult due to cracking and microstructural segregation of alloying elements in the HAZ, a problem that can potentially be overcome by FSW. Development of FSW tool materials and process parameters for superalloys such as Inconel is ongoing, but faces difficulties due to the high plasticisation point of superalloys and, in the case of Inconel, its property of work hardening whilst being stirred.
Steel and ferrous alloys
Friction stir welding of steels has now reached a level of technical maturity where weld lengths up to 30m can be attained in a wide range of engineering steels. These welds demonstrate excellent mechanical properties and there are good indications that their corrosion and fatigue properties will exceed those of fusion welds. The process shows a degree of robustness, suiting it to industrial application and can weld some steels, for example ODS steels, that are currently considered unweldable.
The benefits associated with the FSW of steel include:
- Reduced distortion, and thus reduced rectification costs
- Lower defect rates, therefore reduced re-work and, possibly, lower NDT costs
- Enhanced weld properties (mechanical and corrosion)
- The ability to weld steels that are difficult or impossible to weld by other techniques
- Reduced labour, training and certification costs
- Reduced energy consumption
- No exposure to weld fume, hexavalent chromium, ultraviolet light, radiofrequency radiation or molten metal.
When these considerations are taken into account, the economic case for FSW begins to shift to the extent that several applications of FSW in steel are currently entering commercial use.
Although the majority of common titanium alloys are generally weldable by conventional means, problems with workpiece distortion, and poor weld quality, can occur. In addition, some of the more advanced titanium alloys (such as Ti-6246 and Ti-17) can be difficult to weld by fusion processes. The development of FSW offers the possibility of a new, cost-effective method of producing high-quality, low-distortion, welds in Ti sheet and plate.
The first trials on FSW of Ti were carried out as early as 1995, as part of TWI's internal research programme. These initial welds were conducted on commercially pure (grade 2) titanium, and proved the potential of applying FSW to Ti alloys.
The current state of the art with respect to FSW of Ti alloys can be defined as follows:
- The feasibility of joining Ti alloys by FSW has been proven
- Welds have been successfully made in CP Ti, Ti-6Al-4V, and Ti-15V-3Al-3Cr-3Sn
- Weld tensile strengths similar to those of the parent alloy have been achieved in 6.35mm thickness Ti-6Al-4V.
Friction stir welding of titanium alloys is still an emerging technology, and the following limitations exist at this time:
- The FSW process has not yet been fully optimised for this application, and the joining of Ti alloys by FSW remains a challenging undertaking
- Limited experience exists of the application of this technology to real components
- The advanced tool technology and supporting systems required for FSW of Ti are currently relatively expensive.
TWI is currently in the process of conducting further work aimed at improving the quality and repeatability of FSW in Ti alloys. It is believed that the results of this work will be of significant benefit to the commercial application of this technology.
As FSW is a solid-state process that mechanically mixes metals together to form a bond between them, it can be used to join dissimilar metals. This is most easily achieved when the metals to be joined have similar thermal properties and plasticisation temperatures. As an example, TWI has successfully joined copper, silver and gold, replicating the ancient Japanese mokume-gane technique for making jewellery, but with far less waste than the traditional handcraft method.
Thermoplastics soften and flow as they are heated, and then regain their stiffness as they cool. They can therefore be friction stir welded. Friction stir welding of thermoplastics is more complex than the FSW of metals as thermoplastics comprise long chain molecules rather than individual atoms and so have very different flow characteristics. Process parameters and tool designs for the FSW of thermoplastics are therefore quite different from those required for metals.