EFFECTS OF Zn ADDITIONS ON THE GRAIN BOUNDARY PRECIPITATION AND CORROSION OF Al-5083 M.C. Carroll, P.I. Gouma, M.J. Mills, G.S. Daehn and B.R. Dunbar* The Ohio State University, Department of Materials Science and Engineering, Columbus, OH *Century Aluminum Corporation, Ravenswood, W.V. (Received April 27, 1999) (Accepted in revised form October 12, 1999) Keywords: Electron diffraction; Transmission electron microscopy; Aluminum alloys; Grain boundaries; Corrosion Introduction Stress corrosion cracking (SCC) concerns in aluminum alloys containing Mg levels greater than 3.5% have been largely attributed to the formation of the -phase (Al 3 Mg 2 ) at grain boundaries [1]. It has been demonstrated that the -phase need not be continuous in order to provide a path for crack propagation, but aging treatments, exposure to intermediate to high temperatures, and excessively corrosive environments can all contribute to early failure of Al-Mg alloys due to SCC. Proof of the presence of a corrosion-prone secondary phase can be demonstrated easily through exfoliation testing and the associated lining of grain boundaries, which can be confirmed optically. Additions of Zn to these Al-Mg alloys in levels of 1–2wt% have been shown to be more SCC resistant due to the formation of a stable ternary Al-Mg-Zn phase [2], the  phase. Recent studies have shown that Al-5083 variants which contain even minor levels of Zn (0.68 – 0.70wt%) perform much better during exfoliation testing [3]. Transmission electron microscopy (TEM) observations, coupled with energy dispersive spectros- copy (EDS) investigations, demonstrate the formation of a ternary Al-Mg-Zn phase along grain boundaries when Zn is added in levels of between 0.68 and 0.70% to an otherwise standard Al-5083 alloy composition. Investigations of the microstructure as well as the crystal structure of the binary Al-Mg phase and the ternary Al-Mg-Zn phase, along with the pitfalls associated with classifying these phases, are compared and discussed. Experimental Castings of 5000-series alloys, which fell into the normal window of the common Al-5083 alloy (4.7wt% Mg), were modified by the addition of Zn in levels between 0.68 and 0.70%. All samples were processed in a manner that produced an H31-type temper, which included a small amount of cold work (18% reduction) followed by stabilization at 110°C for two hours. For comparison, several samples were also studied which were either of a standard Al-5083 chemistry or contained elevated levels of Mg (5.4%) in order to maximize sensitization effects and -phase formation. It should be noted that no samples based on the 5083 chemistry exhibit the formation of a secondary phase without an additional “sensitizing” artificial aging treatment at an elevated temperature. Samples for this study underwent a sensitizing treatment of 24 hours at 200°C in order to allow precipitation of secondary particles to a Scripta mater. 42 (2000) 335–340 www.elsevier.com/locate/scriptamat 1359-6462/00/$–see front matter. © 2000 Acta Metallurgica Inc. Published by Elsevier Science Ltd. All rights reserved. PII: S1359-6462(99)00349-8