Thermodynamics-based design of creep resistant Mg solid solutions using the Miedema scheme

Abstract

Most current creep resistant alloy design methods for Mg alloys aim at incremental strengthening of existing alloys such as AZ91 thorough addition of ternary and quaternary elements, or alternatively through RE’s additions. The most remarkable improvements has been through either Mg-Y or recently developed Mg-Gd-Y-Zr and Mg-Y-Nd-Zr alloys, although still below the strength of Mg-Th alloys, which are being phase out due to Th’s radioactivity. Many micromechanisms proposed to account for the observed improvements in either hardness or creep strength of Mg alloys including solid solution and/or precipitation hardening, dynamic precipitation, solute dragging/solute clustering and more importantly atomic size controlled diffusivity more often than not are little more than ad-hoc, after-the-fact explanations, valid at best for the particular system under consideration. Hence, extrapolation to other alloy-systems is either not possible or contradictory. In practice, a feasible path for systematic, either precipitation-hardened or creep-resistant Mg alloys, selection and design is still lacking. The aim of the present work was to use atomic level thermodynamical arguments that suggest that extending the athermal regime through short range order (SRO) is a most feasible path to increasing the creep strength of Mg alloys. Potential solutes were sorted out and ranked based on their tendency to developing SRO using the Miedema’s phenomenological scheme. The predictions were corroborated by experiments involving dilute binary alloys as well as through a collection of data from the literature. Monotonic compression and stress relaxation tests were carried out on specimens of 6 cast binary alloys with (at.%) 2.5 Al, 0.6 Sn, 2.2 Zn, 0.8 Gd, 1.3 Y and 0.9 Nd, and of a similarly cast AZ91D alloy for reference. The solute concentration in the binary alloys was kept deliberately low to limit precipitation hardening effects during the testing. Compression testing was carried out at 298K (25°C), 373K (100°C) and 453K (180°C) for all of the alloys and at 493K (220°C) and 523K (250°C) for the Gd, Y and Nd containing ones. Stress relaxation testing was done at 453K (180°C) at either a predetermined strain (0.05) or stress (150 MPa). The Mg-Al and the AZ91 alloys softened considerably above 373K (100°C), consistently with random solid solution by Al. The rest of the alloys exhibited increasing strength and reduced relaxation, in the order Sn, Zn, Nd, Gd and Y, an indication of a progressively extended athermal regime in the strength-temperature relationship of these alloys. The overall behaviour can be accounted for through the respective solutes’ tendency to develop short range order, lowest for Al and highest for Y, Gd and Nd, in agreement with what the respective phase diagrams suggest. The relative tendency or lack of it, of most solutes to develop SRO rationalizes a number of observations in current multicomponent Mg alloys while it disputes the viability of several other micromechanisms often considered. The feasibility of strengthening HPDC Mg alloys through a combination of percolating eutectics and residual solute in solution, or, more generally, age hardenable alloys through either homogeneously nucleated, coherent with the Mg matrix, precipitates, or via internally ordered precipitates mimicking Mg-Th alloy system, is also considered by making parallels with either the Al-Zn or the Al-Cu alloy systems. Examples of these approaches were discussed using data from the literature, such as experiments on alloys combining Mn and Sc to introduce order in the soft Mn precipitates, or the use of Na to provide finely dispersed, homogeneously nucleated clusters with the aim of refining the subsequent precipitation in alloys such as Mg-Sn.