Some aspects of age-hardening, grain refinement, deformation and fracture in al-li based alloys
thesisposted on 2017-02-17, 01:25 authored by Moutsos, Stavroula
Aluminium-lithium alloys such as 8090 (Al-Li-Cu-Mg) have been developed with the aim of using them instead of conventional aluminium alloys such as 2024 (Al-Cu-Mg) and 7050 (Al-Zn-Mg-Cu) in order to save weight (by up to 10%), especially in aerospace structures. This aim has not, however, been achieved largely because the fracture resistance of Al-Li alloys is often inferior to conventional alloys, with Al-Li alloys being especially prone to low-energy intergranular fracture. Thus, Al-Li alloys have only been used in niche applications to date, such as for the external tank of the space shuttle and the fuselage skin and frames of a helicopter. Service experience for the latter showed that crashworthiness was poor due to the occurrence of brittle intergranular fracture, and the use of Al-Li alloys in this application has now been discontinued. Various hypotheses have been proposed to account for the poor intergranular fracture resistance of Al-Li alloys including: (i) a planar slip mode due to the presence of easily sheared δ' precipitates produced by age-hardening, which results in dislocation-pileups and high stress concentrations at grain boundaries, and (ii) lithium segregation at grain boundaries which weakens interatomic bonding across boundaries. One of the aims of the present work was to resolve controversies regarding the cause(s) of brittle intergranular fracture. Another aim was to determine if grain refinement by severe-plastic-deformation processes (friction-stir processing, high-pressure torsion, and equal-angle channel pressing) increased the fracture resistance and strength, without compromising other properties such as the age-hardening response. A wide range of techniques was used to characterise the microstructures, slip characteristics, and fracture-surface appearance, including: (i) optical microscopy (OM), (ii) scanning electron microscopy (SEM), (iii) transmission electron microscopy (TEM), and (iv) atom-probe tomography (for direct measurements of grain-boundary segregation). Other techniques used (to a lesser extent) included: (i) atomic force microscopy (AFM), (ii) scanning laser confocal microscopy (SLCM), and (iii) electron backscattered diffraction (EBSD). A range of Al-Li alloys with varying Li contents were tested and compared with other aluminium alloys, e.g. Al-Mg alloys with varying Mg contents, and with a nickel-base superalloy (Waspaloy) with easily sheared, coherent L12 γ' precipitates, analogous to the coherent δ' precipitates in Al-Li alloys. Since the work involved several aims and used a wide variety of techniques and alloys, separate results followed by a discussion have been written for (i) the effects of severe-plastic-deformation processing on microstructures, (ii) the effect of grain size on age-hardening, (iii) the fracture behaviour with respect to slip characteristics and solute content, and (iv) grain-boundary segregation. These topics are, of course, interconnected, but the above layout of the thesis makes it easier to follow than combining all the results into one section followed by one discussion section. The major aims of the work have largely been achieved, with the observations (and a critical review of the literature) showing that planar slip is not an important factor in promoting brittle intergranular fracture. Rather, it appears that lithium segregation (leading to a specific temperature-dependent 2-D structure) at grain boundaries, in conjunction with strain localisation in precipitate-free zones at grain boundaries, is mainly responsible for brittle intergranular fracture in commercial alloys (with low levels of alkali-metal impurities). For example, brittle intergranular fracture was observed at low testing temperatures for a mechanically alloyed 5091 alloy that exhibited homogeneous slip, and transitions to ductile transgranular fracture occurred with increasing temperature despite no changes in slip mode. For the nickel base superalloy, marked slip planarity was observed, but fractures were transgranular and ductile (at 20°C and -196°C). Numerous other observations also showed that there was no correlation between slip morphology and fracture behaviour. Direct evidence for lithium segregation at grain boundaries using atom-probe tomography was not obtained, probably because preferential field-evaporation of lithium (due to its low atomic number) occurred at grain boundaries. However, only a limited number of grain boundaries were examined since atom-probe tips often fractured (presumably at grain boundaries weakened by segregation) prior to collecting sufficient data for analysis. Magnesium segregation at grain boundaries was detected in some alloys (and has been reported in the literature for other alloys), and probably also results in grain-boundary embrittlement. Thus, results obtained in the present work, along with those from previous work, showed that the tendency for brittle intergranular fracture increased with (i) increase in Mg contents in Al-Mg alloys, and (ii) increasing Mg + Li contents in Al-Li-Mg alloys. Theoretical considerations discussed in the literature suggest that both Li and Mg segregation at grain boundaries in Al alloys should promote brittle intergranular fracture. The present work showed that brittle intergranular fracture was suppressed in 8090 alloys with ultra-fine grain sizes produced by friction stir processing work which involved establishing the processing conditions required for the production of ultra-fine grain sizes. The transition from brittle intergranular fracture in coarse-grained material to ductile dimpled fractures in ultra-fine grained material was probably due to the fragmentation of coarse constituent particles (resulting in numerous, small, dispersed particles that acted as a void-nucleation sites) along with an increase in the volume fraction of precipitate-free zones at grain boundaries. The age-hardening response of grain-refined material was significantly different from that in coarse-grained material for the 8090 alloy, and such differences would need to be borne in mind if ultra-fine grained material were to be used. For example, for the 8090 alloy processed by high-pressure torsion to a grain size ~170 nm, the initial hardness was high (VH ~200) due to grain-size strengthening according to the Hall-Petch relationship, but decreased with ageing time (at the high ageing temperatures of 170 to 200°C) because softening due to grain growth was more dominant than hardening due to precipitation. For friction-stir processed material, with a 600 nm grain size, precipitation-hardening was observed, but the peak-hardness was less than that for the coarse-grained material (despite some grain-size strengthening) due to a larger volume fraction of precipitate-free zone at grain boundaries. While most of the objectives of present work were achieved, there are quite a few ‘loose ends’, and further work is recommended in regard to (i) a more complete characterisation of microstructures after severe plastic deformation for different techniques and processing variables, (ii) further atom-probe and high-resolution transmission electron microscopy to try to detect lithium segregation at grain boundaries, and (ii) the effect of ultra-fine grain and nanocrystalline grain sizes on fracture behaviour. The incompleteness of some of the above topics is largely because it was decided it was better to obtain a broad overview of the issues, using a wide variety of experimental techniques, rather than concentrate on one particular material and processing technique.