Elsevier

Corrosion Science

Volume 73, August 2013, Pages 130-142
Corrosion Science

Characterisation and understanding of the corrosion behaviour of the nugget in a 2050 aluminium alloy Friction Stir Welding joint

https://doi.org/10.1016/j.corsci.2013.04.001Get rights and content

Highlights

  • The corrosion behaviour of the nugget of a FSW (Al–Cu–Li alloy) joint was studied.

  • A post-welding heat treatment modified the corrosion behaviour of the nugget.

  • For the nugget, intergranular/intragranular corrosion was linked to T1 precipitation.

  • A gradient in the nugget grain size led to galvanic coupling at a mesoscopic scale.

  • Onion rings structure explained the localisation of the corrosion in parallel bands.

Abstract

The corrosion behaviour of the nugget of a Friction Stir Welding joint employing a 2050 Al–Cu–Li alloy was investigated. The results showed that the nugget was susceptible to both intergranular and intragranular corrosion. Such corrosion behaviour was related to microstructural heterogeneities observed on a microscopic scale. Furthermore, heterogeneities in the corrosion behaviour of the nugget observed on a macroscopic scale were evidenced by a different corrosion behaviour from the top to the bottom of the nugget and by a localisation of the corrosion damage related to the “Onion ring structure”. Critical microstructural parameters were identified to explain the results.

Introduction

The reduction in the weight of aircraft metallic structures is a current problem that the aeronautic industry has to address. At a time where the role played by composite materials is becoming more significant, the use of the Friction Stir Welding (FSW) process in combination with the new generation aluminium-copper-lithium alloy presents an alternative solution. The FSW process was developed by The Welding Institute (TWI) and consists of using a non-consumable, cylindrical, rotating tool (usually hardened steel) that moves over the seam of two butted plates and stirs them together [1]. Both a strong plastic deformation of the material and a strong increase in temperature were observed below the tool, while gradients of deformation and temperature were recorded perpendicularly to the joint, leading to gradients of microstructural evolution that were also perpendicular to the joint. Therefore, a typical FSW joint consists of the unaffected base material (BM), a heat affected zone (HAZ), a thermo-mechanically affected zone (TMAZ) and a dynamically recrystallised zone (nugget) as observed in Fig. 1 for a FSW joint of an Al–Cu–Li alloy 2050 studied in a previous work [2]. This alloy, i.e., the base material, is a precipitation hardening alloy. The main phase responsible for the hardening process is T1 (Al2CuLi), but some other precipitates should be found, including θ′ (Al2Cu), T2 (Al5Li3Cu) and TB (Al7Cu4Li), which can also contribute to strengthening the alloy but to a lesser extent than the T1 precipitates [3], [4], [5]. Other intermetallic particles can be encountered in these alloys, such as Al3Zr particles, which prevent recrystallisation phenomena, and Al6Mn and Al20Cu2Mn3 particles, which help control the grain size [6]. The zones generated during the FSW process differ from one to another due to their grain size and morphology and because of their differences in hardening precipitation [2], [7], [8]. These microstructural differences lead to modification of the corrosion behaviour. For example, each zone of the FSW joint was characterised by its own corrosion potential value [2], [9], [10], which led to galvanic coupling phenomena, as shown in a previous work [2]. The results obtained in this previous work showed that the nugget and the HAZ acted as sacrificial anodes when the welded joint was not heat treated after welding [2]. Conversely, a post-welding heat treatment inverted the corrosion potential values, shifting the potential of the BM toward more cathodic values so that it acted as a sacrificial anode. Other authors corroborated these results [11]. Nevertheless, the most interesting part of the FSW welded joint is the centre, called the nugget. Previous studies have revealed that the nugget has typical corrosion behaviour, closely linked to its particular microstructure that is generated by a combined role of high strain rate and temperature. Jariyaboon et al. [9] and Mahoney et al. [12] have measured the temperature in the centre of the weld, i.e., the nugget, and the authors found that it was approximately 481 °C and 500 °C, respectively. Moreover, Mahoney et al. [12] have also estimated that there was a gradient of temperature in the thickness of the joint and, consequently, in the nugget. As mentioned before, the centre of the weld exhibits the most important deformation during the FSW process. Jata and Semiatin [13] have estimated a strain rate of 10 s−1, whereas Masaki et al. [14] assumed that it was approximately 2–3 s−1. It was assumed that the nugget was the place of dynamic recrystallisation and led to the formation of small equiaxed grains [2], [9], [13], [14]. Some authors have noticed that the grain size was not homogeneous in the nugget [15], [16] and instead varied along the thickness of the joint, with a maximum grain size in the top of the nugget. This evolution could be attributed to the thermal gradient between the top and bottom caused by the presence of the shoulder. Fonda and Bingert [17] have also noticed the existence of a hardness variation between the top and bottom of the nugget and related the result to both a gradient of temperature and the reduction of time for T1 precipitate nucleation in the bottom of the nugget. Moreover, the FSW process was found to lead to the formation of a typical microstructure called an “Onion Ring,” which consists of bands of different textures, as observed in Fig. 1. Kumar and Kailas [18] proposed a mechanism to explain such a microstructure: a wrenching of the material in front of the tool while the material was then put behind the probe. Following this proposed mechanism, some authors found that the distance between two bands in the “Onion Ring” was equal to the distance reached by the probe in one evolution [19], [20]. Microstructural heterogeneities at a finer scale were also found to have detrimental effects on the corrosion behaviour of the nugget. Birbilis and Buchheit [21] have studied the corrosion behaviour of intermetallic particles present in Al–Cu–Li alloys. They found that Al3Zr, Al6Mn, and Al20Cu2Mn3 particles have a corrosion potential equal to −0.752 V/SCE, −0.839 V/SCE and −0.550 V/SCE, respectively, in a 0.01 M NaCl solution whereas the matrix has a potential of −0.679 V/SCE in the same solution. Thus, some micro-galvanic coupling phenomena might appear to lead to the dissolution of the particles or the matrix. In addition, the hardening precipitates are very active particles, as mentioned by Li et al. [22], who evaluated the electrochemical behaviour of T1 and θ′ with respect to the matrix and brought to light that the T1 precipitates had a more cathodic corrosion potential than the matrix in a 4% NaCl solution. Li et al. [23] have also studied the evolution of the galvanic coupling between T1 or T2 precipitates with the matrix.

Therefore, numerous studies showed that the nugget of an aluminium alloy (AA) 2050 FSW joint presents a very heterogeneous microstructure, which suggests complex corrosion behaviour. In a previous study [2], the corrosion behaviour of an AA 2050 FSW joint has been studied, but the phenomena occurring in the nugget have not been elucidated. The aim of the present work is to investigate in detail the corrosion behaviour of the nugget of an AA 2050-T34 FSW joint and to correlate the results obtained to the microstructural characterisation to identify the critical microstructural parameters.

Section snippets

Material

The material studied in this work is a new generation aluminium–copper–lithium alloy AA 2050 (Al base, 3.5% Cu, 1% Li – weight per cent) provided by Constellium (Voreppe, France). It consists of 15 mm thick-rolled plates of the T34 metallurgical state, which corresponds to stretching before natural ageing. Two plates were joined together by Friction Stir Welding (FSW) in the EADS Innovation Works Laboratory. The welding process consists of firmly bridling the two plates edge to edge. A rotating

Results: heterogeneities in the corrosion behaviour of NHT and PWHT nuggets

Results obtained in a previous work showed that NHT and PWHT joints presented the same microstructure at the optical microscope scale with four different zones: base metal, HAZ, TMAZ and nugget [2]. These zones were found to have a different corrosion behaviour from one to another. In the present work, attention was paid to the corrosion behaviour of the nugget. Fig. 2 illustrates the results obtained from the gel visualisation technique after 24 h of exposure for the NHT (Fig. 2a and c) and

Understanding the susceptibility to intergranular and intragranular corrosion of the nuggets

As brought to light by the experiments performed, the morphology of the corrosion features observed in the corroded zones of the nuggets, i.e., in the whole surface of the NHT nugget and in some specific parallel bands for the PWHT nugget, depends on the metallurgical state of the nugget, with intergranular corrosion and both intergranular and intragranular corrosion observed for the NHT and PWHT nugget, respectively. TEM observations performed for the NHT nugget (Fig. 5) showed very few

Conclusions

This work focused on the corrosion behaviour of the nugget in an AA 2050 FSW joint. Two metallurgical states were considered, i.e., the as-welded joint called NHT and a joint submitted to a post-welding heat treatment called PWHT. The results showed that the post-welding heat treatment significantly modified the corrosion behaviour of the nugget. For the NHT nugget, a susceptibility to intergranular corrosion was shown while, for the PWHT nugget, both intergranular and intragranular corrosion

Acknowledgments

This work was financially supported by the ANR MatetPro program (ANR-08-MAPR-0020-05). The authors thank C. Henon (Constellium, Voreppe, France) for the material and P. de Parseval for the EPMA analyses.

References (39)

  • B.P. Huang et al.

    Scripta Mater.

    (1998)
  • B.M. Gable et al.

    J. Light Met.

    (2001)
  • A.K. Shukla et al.

    Scripta Mater.

    (2007)
  • M. Jariyaboon et al.

    Corros. Sci.

    (2007)
  • K.V. Jata et al.

    Scripta Mater.

    (2000)
  • K. Masaki et al.

    Scripta Mater.

    (2008)
  • M.A. Sutton et al.

    Mater. Sci. Eng.

    (2002)
  • K. Kumar et al.

    Mater. Sci. Eng. A

    (2008)
  • K.N. Krishnan

    Mater. Sci. Eng. A

    (2002)
  • G.R. Cui et al.

    Scripta Mater.

    (2008)
  • J.F. Li et al.

    J. Alloys Compd.

    (2008)
  • K.D. Ralston et al.

    Electrochim. Acta

    (2011)
  • P. Niskanen et al.

    Corros. Sci.

    (1982)
  • K.D. Ralston et al.

    Scripta Mater.

    (2010)
  • J.G. Brunner et al.

    Corros. Sci.

    (2012)
  • S. Gollapudi

    Corros. Sci.

    (2012)
  • S. Zhao et al.

    J. Stat. Planning Inf.

    (2007)
  • C. Luo et al.

    Corros. Sci.

    (2012)
  • S.H. Kim et al.

    Scripta Mater.

    (2001)
  • Cited by (104)

    View all citing articles on Scopus
    View full text