Critical Assessment 25: Friction stir processing, potential and problems

ABSTRACT

This assessment considers recent work on friction stir processing (FSP), which has been demonstrated to be an effective method for grain refinement and synthesis of new alloys and composites. The grain refinement is attributed to high strain rates leading to recrystallisation, while external cooling suppresses grain growth during cooling. The technique is capable of producing nanocrystalline alloys, and also able to disperse nanoparticles into alloys. The mechanical properties of processed materials agree with a combination of existing models for grain refinement, and precipitate reinforcement theory. Further improvements in the technique may help deal with severe tool wear during the FSP of composites, and reduce the complexity of composite fabrication using novel processing methods and tooling.

KEYWORDS:

• Friction stir processing
• mechanical properties
• microstructures
• critical assessment

Introduction

Since the invention of friction stir welding in 1991, numerous process variants of this technology have been developed in the subsequent years. The technique involves using a rotating cylindrical tool which consists of a shoulder and a pin-shaped protrusion which plunges through the interface of two metal sheets, and is traversed across the material to join the pieces by consolidation. A notable feature of the resulting joint is that the weld stir zone consists of recrystallised grains that typically are finer than many base materials. Perhaps the most interesting and widely studied variant of this technology is friction stir processing (FSP), which involves traversing the rotating tool across only a single piece of material for the purpose of modifying the microstructure, properties, or composition.

The first journal publications discussing FSP emerged in 1999 [1], and there have since then been more than 1000 papers, hundreds of patents, and multiple review papers and textbooks covering the topic. The initial applications of FSP were reported on wrought aluminium alloys as a grain refinement method to provide superplasticity [1], and as a mechanical consolidation method to homogenise microstructures of castings [2], powder metallurgy materials [3,4], and composites [5], which have been reviewed in 2008 [6]. The present paper will discuss some of the more recent developments, in terms of the application of FSP to microstructural modification of alloys, synthesising new materials, and their corresponding properties. Future directions will be highlighted which remain under-represented in the literature, or hold potential for further research.

Potential for microstructure modification

The earliest application of FSP involved evaluating the potential for grain refinement of alloys, and this has continued to be a key focus, with studies involving cast Al, Mg, and Ti alloys [7,8], with methods for producing ultrafine (<1 µm) [9] or even nanograined structures in Mg alloys feasible [10], with compositions summarised in Table 1. The external cooling or quenching applied during FSP suppresses grain growth during the cooling period on the trailing edge of the tool. A comprehensive analysis has shown that when performing FSP on aluminium with purity levels from 99 to 99.999 wt% and applying liquid nitrogen cooling, the minimum grain size that can be achieved is correlated with Zener–Hollomon parameter, which is an empirical function of the (high) strain rates during deformation, and material constants that depend on alloy composition [11]. However, incipient melting has been reported when dealing with cast alloys such as AZ91 magnesium, during the FSP of this material which will limit the strain rates that can be achieved due to tool slippage [12]. Furthermore, FSP is normally applied as a surface treatment with a tool that is typically <25 mm wide with a depth of <5 mm; it is possible to stagger the passes and process thicker materials throughout their thickness to achieve bulk material synthesis. Processing of complex shapes with compound curves and components more than 1 m in each dimension is possible using robotic systems.

Table 1. Compositions of selected alloys used in FSP research discussed here.

Alloy	Composition, wt-%
AZ91	Mg–9Al–1Zn
AA6360	Al–0.7Si–0.36Fe–0.31Mg
AA6056	Al–1Si–0.9Mg–0.8Cu
AA5083	Al–4.5Mg–0.7Mn
AA5052	Al–2.5Mg–0.25Cr
AA7075	Al–5.6Zn–2.5Mg–1.6Cu
AA8026	Al–0.9Fe–0.7Mn–0.4Mg
Cast A356	Al–7Si–0.35Mg
Super austenitic stainless steel	Fe–0.02C–28Cr–33Ni–3.5Mo–0.6Mn
Commercially pure Cu	99.9Cu
IF steel	Fe–0.002C–0.424Ti–0.003N–0.005S–0.0465Al–0.0029Nb
Cast marine bronze	Cu–9Al–4.5Ni–1Mn
SKD11 (or D2 tool steel)	Fe–1.5C–12Cr–1Mo–0.6Mn–0.4Si–0.3V

FSP of aluminium alloys, such as AA6056, was demonstrated to nearly double the ductility to failure due to the break-up and redistribution of elongated dispersoids, coupled with closure of pre-existing porosity in the base material [13]. This mechanism helps to explain the enhanced grain refinement in alloys which were previously refined by deformation techniques, such as equal channel angular pressing [14] and constrained groove processing [15,16]. Break-up of insoluble dispersoids adds to further grain refinement when FSP is applied repeatedly during multi-pass processing; however, if the particles are soluble, they may dissolve with each pass, which leads to coarse grains and reduction in strength when the number of passes increases [17]. The application of external quenching at the trailing edge of the tool enables nano-scale grains (<500 nm) to form during FSP in multiple alloys, including Al-Mg alloys (such as AA5083 and AA5052) [18–20], super austenitic stainless steel [21], pure copper [22], and interstitial-free steel [23].

FSP for microstructure refinement has also been applied to coatings, including Al–Cr–O plasma-sprayed composites [24], Al–Al₂O₃ cold-sprayed composites [25], electro-spark-deposited TiB₂-TiC composite [26], flame thermal-sprayed Ni–Cr–Al alloy [27], and cemented carbide layers [28]. Other benefits also have been attributed to FSP, such as improved liquation cracking resistance in the heat-affected zones of nickel-based alloys [29], and consolidation of sintered microstructures in powder metallurgy-based metal matrix composites [30]. The technique has also been employed at a commercial level on cast bronze marine material for propellers to increase tensile strength from 420 to 700 MPa, with only a marginal drop in ductility from 20 to 14% [2,31].

There is interesting potential in combining FSP with other processes, such as laser surface melting, where it was found that nanostructured tool steels can be produced in SKD11 steel [32]. Preliminary studies have also shown that FSP can be used to refine materials such as a carbon nanotube/AlSi10Mg composite [33], and also copper/nickel 70/30 alloy [34] which were initially produced by additive manufacturing techniques. Other novel techniques involve conducting a large electrical current through the tool and work piece in order to extend the refined microstructure zone away from the tool shoulder and into the thickness of the plate; however, it was not verified if fine-grain structures could still be produced while applying resistive heating [35].

Fabrication of new alloys and composites

One of the unique features of the FSP is the ability to fabricate metal matrix composites in linear bands along the surface of a plate (with cross-sectional areas of up to approximately less than 15 × 10 mm) by consolidating seemingly arbitrary combinations of particles and alloys [36]. This can be achieved by processing metallic plates with either grooves or holes drilled into the surface which are later filled with the particles before processing the surface. Since fine-scale powders can be embedded into the plate, such as nanoparticles, this offers method to synthesise composites which benefit from an enhanced Orowan looping strengthening mechanism, while also refining grain structures by suppressing grain growth via grain boundary pinning [37]. However, care must be taken in selecting an alloy for composite fabrication, since peak-aged alloys may experience metal matrix softening in the stir zone which offsets the benefits from incorporating other particles [38]. In general, there has been a particular interest in fabricating composites using nanoparticles [39–41], since these may offer the greatest strengthening benefit, considering the Orowan looping and grain boundary pinning mechanisms since it is possible to distribute nano-scale particles which have an incoherent interface with the matrix, whereas traditional precipitation-aged aluminium alloys contain nano-scale particles with coherent or semi-coherent interfaces. Earlier methods of mechanical alloying (by mixing metallic powders with inert oxides) followed by extrusion can produce material in bulk, though ultra-fine grain structures cannot be readily achieved [42]. Also, FSP, in contrast, offers a method for applying selective mechanical alloying locally to a component, and achieves much higher strain rates which help to disperse the oxide layers on the metal powders and recrystallise the grain structures.

Although metal matrix composites incorporating carbon nanotubes into aluminium alloys offer potential, it is not easy to disperse severely agglomerated nanoparticles, such as tangled carbon nanotubes in a single FSP pass [43]. Attempts to improve the dispersion of carbon nanotubes using repeated passes result in damage and transformation of the nanotubes into other forms of carbon or aluminium carbide [44]. As a result, the addition of carbon nanotubes to aluminium alloys has not demonstrated significant advantages compared to other micro-scale or nano-scale particles based on inert oxides or carbides. Consequently, there has not been any report of an Al-based composite produced by FSP with substantially higher strength than some of the strongest conventional alloys (such as AA7075-T6 for example). Thus it may be preferred to utilise nano-particle-reinforced composites which incorporate phases which are thermally stable at high temperatures and stresses, such as ZrO₂ [45], or use powders which react with each other or the matrix alloy to nucleate nanoparticles in situ, by so-called reactive FSP. The reactive processing technique has been used to form nano-scale TiC particles (using K₂TiF₆ and graphite powders) [46], Al₃Ti and Al₂O₃ particles (using TiO₂ powder) [47], Al₃Ti particles (using Ti powder) [48], and TiB₂ particles (using K₂TiF₆ and KBF₄ powders) [49] for example. However, the in situ nucleation method requires some careful planning in order to identify starting materials and subsequent reactions, and this often leaves residual unreacted powders remaining in the stir zone with a non-uniform distribution.

A wide variety of other particles has been examined as reinforcing phases in FSP-fabricated composites, including Fe-based metallic glass particles which retain their amorphous structure [50], and NiTi particles which can impart some shape memory behaviour to the composite [51], and low-cost glass fibres [52]. Matrix materials other than aluminium have also been explored, including copper [53–55], polymer [56–58], magnesium [59,60], and steel [61]. These techniques may hold promise; however, tool wear is a major problem when making ceramic-reinforced composites, and this is greatly exacerbated when a high temperature material like steel is used as the matrix material during FSP, with tool wear that is even more severe than when conducting friction stir welding of steels (when no ceramic reinforcement is added).

Mechanical properties of friction stir-processed materials

A frequent application goal for FSP-modified or -synthesised materials involves enhancement of strength, hardness, or ductility as a result of grain refinement or improved dispersion of second phases. This generally follows from the fine-grain structures that lead to Hall–Petch strengthening, yet contain intermediate levels of dislocations which have not been saturated by work hardening which can occur during rolling or other conventional processing. Structures, which contain second phases, also derive increased ductility by the dispersion and refinement of secondary phases [62], and this suppresses the onset of microvoid coalescence and fracture when there is a mixture of hard and soft phases [63]. These mechanisms provide a structure which still can accommodate useful levels of strain while maintaining high strength. Detailed analysis of Al–Mg/Al₃Ti+MgO nano-particle-containing composite produced by reactive FSP revealed that the yield strength could be well predicted by combining models for Hall–Petch grain boundary strengthening and Orowan strengthening [19,64–67]. These reports are based on a modification of prior theories on aluminium- and magnesium-reinforced composites, although the contributions of mismatch in coefficient of thermal expansion, work hardening and misfit strain between the particle and matrix [68] have not been implemented in most models for FSP materials.

The enhanced ductility achieved following FSP has been widely recognised, and noted to enable high strain rate superplasticity in various Al and Mg alloys [6], including as-cast Al alloys [69,70]. A key goal of this has been to enable superplastic forming of components, and some progress has been demonstrated using AA5083-H116 alloy to enhance the ductility of local regions (typically <20 mm in width) which experience higher strains and avoid tearing during forming [71]. Using scribed marker lines, the high strains in the FSP microstructures have been directly shown to occur by grain boundary sliding [72], which is assisted by the fine-grain structure composed mainly of high angle boundaries [73].

Multiple authors have evaluated metal matrix composites that incorporate more than one type of particle, as so-called ‘hybrid’ composites, and observed that these tended to have enhanced wear resistance compared to when only one reinforcing particle is used with the same volume fraction. For example, the addition of MoS₂ and SiC particles to A356 cast alloy by FSP was found to decrease the wear rate during sliding tests [74], and likewise for AA6360/(TiC+B₄C) [75], AA8026-TiB₂-Al₂O₃ [76], and Al/(SiC+Al₂O₃) [77]. However, the wear rate is dependent on the specific ratio of particle fraction [77], and the exact mechanisms for the enhanced performance appear to be related to the formation of a superior thin tribo-film [75].

Outstanding problems

Some of the most significant challenges, in applying FSP to composite materials, relate to the excessive wear when composite materials are fabricated due to the abrasive effect of the hard ceramic particles. Although it has been noted in one case that the tool wear may slow down after the tool reaches a stable worn shape [78], this is not necessarily a general solution. The application of FSP to higher melting point materials, such as high temperature oxide dispersion strengthened steels [79], is also limited by severe tool wear. Another major issue that is often overlooked in the literature is the extremely time-consuming processing of composite fabrication, which typically involves cutting a groove or drilling holes, filling the cavities with powders, and multiple FSP passes. Although a significant amount of process development has occurred, there has been limited success in achieving a uniform distribution of particles in composites in a single pass. Despite a wide range of studies on composites using different particles, a clear correlation between the mechanical behaviours of composites with the particle properties (such as matrix/particle interface coherency) remains unclear. Although FSP services are available at a commercial scale and surface treatment has been demonstrated on components more than one metre in size, the application of the technology for manufacture of bulk composites is rather complicated and inefficient compared to competing processes such as mechanical milling followed by extrusion [42,80].

Summary and perspective

The use of FSP as a modification tool for refinement of alloy microstructures has become well established, and even used in limited commercial applications such as marine propellers [81]. Grain refinement has been shown to be a function of the high strain rate deformation Zener–Hollomon parameter, while external cooling suppresses grain growth during cooling. FSP has also been demonstrated as a useful tool for synthesising new materials, including composites, offering a particular though not unique capability to disperse nanoparticles into a wide range of matrix metals. Although severe tool wear is an issue during FSP of composites, tool materials adopted from friction stir welding work (such as polycrystalline boron nitride, and tungsten alloy/carbide composites) likely could help to address tool wear, but any selections need to be cost-effective, and cost cannot be assessed without a specific component in mind, whereas most research has focused on basic studies. Future work may also consider methods to reduce the required time and complexity of FSP using alternative technologies, such as a novel friction stir additive manufacturing method [82], although the associated equipment is rather more specialised than FSP techniques, utilising standard friction stir welding equipment.

Disclosure statement

No potential conflict of interest was reported by the author.

ORCID

Adrian P. Gerlich http://orcid.org/0000-0003-0270-7475