Precipitates and intermetallic phases in precipitation hardening Al–Cu–Mg–(Li) based alloys

Abstract

The present study contains a critical review of work on the formation of precipitates and intermetallic phases in dilute precipitation hardening Al–Cu–Mg based alloys with and without Li additions. Although many suggestions for the existence of pre-precipitates in Al–Cu–Mg alloys with a Cu/Mg atomic ratio close to 1 have been made, a critical review reveals that evidence exists for only two truly distinct ones. The precipitation sequence is best represented as: supersaturated solid solution→co-clusters→GPB2/S"→S where clusters are predominantly Cu–Mg co-clusters (also termed GPB or GPB I zones), GPB2/S" is an orthorhombic phase that is coherent with the matrix (probable composition Al10Cu3Mg3) for which both the term GPB2 and S" have been used, and S phase is the equilibrium Al2CuMg phase. GPB2/S" can co-exist with S phase before the completion of the formation of S phase. It is further mostly accepted that the crystal structure of S' (Al2CuMg) is identical to the equilibrium S phase (Al2CuMg). The Perlitz and Westgren model for S phase is viewed to be the most accepted structure. 3DAP analysis showed that Cu–Mg clusters form within a short time of natural and artificial aging. Cu–Mg clusters and S phase contribute to the first and second stage hardening during aging. In Al–Cu alloys, the θ phase (Al2Cu) has I4/mcm structure with a=0.607 nm and c=0.487 nm, and θ' phase has a tetragonal structure and a=0.404 nm, c=0.58 nm, the space group is I4m2. Gerold's model for θ" (or GPII) appears to be favourable in terms of free energy, and is consistent with most experimental data. The transformation from GPI to GPII (or θ") seems continuous, and as Cu atoms will not tend to cluster together or cluster with vacancies, the precipitation sequence can thus be captured as: supersaturated solid solution→θ" (Al3Cu)→θ' (Al2Cu)→θ(Al2Cu). The Ω phase (Al2Cu) has been variously proposed as monoclinic, orthorhombic, hexagonal and tetragonal distorted θphase structures. It has been shown that Ω phase forms initially on {111}Al with c=0.935 nm and on further aging, the c lattice parameter changes continuously to 0.848 nm, to become identical to the orthorhombic structure proposed by Knowles and Stobbs (a=0.496 nm, b=0.858 nm and c=0.848 nm). Other models are either wrong (for example, monoclinic and hexagonal) or refer to a transition phase (for example, the Garg and Howe model with c=0.858 in a converted orthorhombic structure). For Al–Li–Cu–Mg alloys, the L12 ordered metastable δ' (Al3Li) phase has been observed by many researchers. The Huang and Ardell model for T1 phase (space group P6/mmm, a=0.496 nm and c=0.935 nm), appears more likely than other proposed structures. Other proposed structures are perhaps due to the T1 phase forming by the dissociation of ½<110> dislocations into ⅙<211> Shockley partials bounding a region of intrinsic stacking fault, in which copper and lithium enrichment of the fault produces a thin layer of the T1 phase.