Microstructural properties of friction stir welded and post-weld heat-treated 7449 aluminium alloy thick plate

https://doi.org/10.1016/j.msea.2007.06.025Get rights and content

Abstract

In order to improve the efficiency of the friction stir welding (FSW) manufacturing process a recently developed 7xxx series aluminium alloy has been welded in an underaged temper for age forming (TAF) condition and subsequently heat treated to the required T7 in service temper. Several complementary microstructural measurement techniques (SEM, TEM and SAXS) have been employed to capture a detailed view of the precipitate distribution, mean particle size, phase identity and fraction. Each friction stir weld zone has been treated separately and compared with the original underaged structure to capture the effect of the local thermal history. A comparison is made between microstructural measurements and predictions of the equilibrium fraction and solvus temperature for each phase using a CALPHAD approach. In contrast with welding thin sections the cooling rates obtained after friction stir welding thick plate are slow enough to allow precipitation of η/η-phase within the nugget and TMAZ on cooling of the weld, at both grain boundaries and throughout the grain interior. Applying a subsequent post-weld heat treatment in effect replaces the fine η/η-phase with coarse overaged precipitates in the nugget/TMAZ and reduces the strength contrast between nugget/TMAZ and HAZ.

Introduction

Friction stir welding of aluminium alloys has been introduced to provide improved joint properties and performance especially for joining alloys that cannot be welded easily by fusion processes. Of particular interest is friction stir welding of high strength aluminium alloys for aerospace applications, where cost and weight savings are possible by replacing riveted joints with friction stir welds. In order to achieve an efficient manufacturing process friction stir welding has been carried out within the aluminium alloy heat treatment process, this allows welding while the material has a reduced strength. The welded joint receives a subsequent post-weld heat treatment in order to gain the required strength. However, the effect of this complex thermal and thermomechanical process route on the microstructure has yet to be determined.

The weld zone has distinct regions known as the nugget, thermo-mechanically affected zone (TMAZ), and heat-affected zone (HAZ) [1]. Each zone experiences a variety of thermal cycles (and deformation in the Nugget and TMAZ) which results in a complex mixture of microstructural processes. The nugget and TMAZ are typically characterised by coarse heterogeneous precipitation on weld cooling [2], [3] and post-weld natural ageing due to GP zone precipitation from retained solute. The grain structure in the nugget is generally dynamically recrystallised [4], [5] whereas the TMAZ generally contains deformed parent grains. The HAZ is characterised by a coarsening of the precipitate particle distribution caused by the thermal cycle alone [6], [7]. This in turn leads to a vast difference in mechanical properties across the weld and such variation is detrimental to the joint performance. Further changes in the microstructure and properties occur during the PWHT, which is used commercially to stabilise the parent material in the required final T7 overaged condition.

The microstructural changes that accompany the friction stir welding of thin section (6.5 mm) AA7449 have been studied by Dumont et al. [8]. They found that SAXS provided a correlation between the distribution of the heat input and the precipitate microstructure and the heat input due to slow weld speeds leads to an unfavourable coarsening regime and an extended heat-affected zone.

For the manufacture of wings by friction stir welding it will be necessary to join thick section material (40 mm). In this case, the thermal cycle and deformation behaviour can be expected to be very different to that of thin plate, essentially the thermal cycle duration is much longer with a significantly reduced cooling rate. In addition, much greater power is required to carry out friction stir welding of thick section plate. So in an attempt to reduce the energy required and load experienced by the tool the weld is produced in an underaged condition and then a post-weld heat treatment is performed. In this paper, the effects of both the different thermal and deformation conditions plus the influence of post-weld heat treatment on the microstructure of thick section AA7449 welds are investigated.

Section snippets

Experimental

Friction stir welding was carried out by Airbus UK on 40 mm thick AA7449 plate in a TAF temper supplied by Alcan, this material is solution heat treated and quenched (50 K/s) followed by a 2% stretch and ageing treatment. Half penetration welds were performed using an Airbus TriflatTM tool, with pin and shoulder diameters of 17 mm and 34 mm, respectively. A subsequent post-weld heat treatment was applied in order to achieve a standard Alcan T7 temper for this alloy.

Microstructural characterisation

As-received microstructure

The material prior to welding is in the TAF condition, this is essentially the first step in a two-step industrial T7 temper and TAF material is therefore in an underaged state. Using SEM in backscattered mode reveals the grain structure and grain boundary precipitates, shown in Fig. 1(a), while bright field imaging using TEM reveals the precipitation within the grain interior, shown in Fig. 1(d). To determine the phases present within the material, selected area diffraction (SAD) was used, and

Conclusions

Friction stir welding of thick gauge AA7449 in an TAF condition leads to a characteristic re-precipitation of η/η within the nugget and TMAZ regions, during the cooling portion of a weld cycle. This is due to the relatively slow cooling rate as heat dissipates through the thick section once the friction stir tool has translated along the weld line. Within the HAZ a regime of coarsening has been observed, this effect lessens from near the weld line towards the unaffected parent material. The

Acknowledgements

The authors gratefully acknowledge the support of Alcan, Airbus UK, EPSRC (EP/D029201/1) and very useful meetings with Phil Prangnell, Hugh Shercliff and Stewart Williams. We are also grateful to Dr. K. Geraki and Dr. J.G. Grossmann of station 2.1 CCLRC Daresbury for their assistance and discussions in regard to SAXS measurements.

References (30)

  • J.Q. Su et al.

    Acta Mater.

    (2003)
  • K.A.A. Hassan et al.

    Acta Mater.

    (2003)
  • C. Genevois et al.

    Acta Mater.

    (2005)
  • M. Dumont et al.

    Acta Mater.

    (2006)
  • S.K. Maloney et al.

    Scripta Mater.

    (1999)
  • X.Z. Li et al.

    Acta Metall.

    (1999)
  • K. Stiller et al.

    Mater. Sci. Eng. A

    (1999)
  • J.C. Werenskiold et al.

    Mater. Sci. Eng. A

    (2000)
  • L.K. Berg et al.

    Acta Mater.

    (2001)
  • T. Engdahl et al.

    Mater. Sci. Eng. A

    (2002)
  • G. Sha et al.

    Acta Mater.

    (2004)
  • A. Kverneland et al.

    Ultramicroscopy

    (2006)
  • J. Gjonnes et al.

    Acta Metall.

    (1970)
  • M. Mahoney et al.

    Metall. Mater. Trans. A

    (1998)
  • K.V. Jata et al.

    Metall. Mater. Trans. A

    (2000)
  • Cited by (115)

    View all citing articles on Scopus
    View full text