Roger D. Doherty

Strain hardening of titanium: role of deformation twinning

Abstract The purpose of this study is to investigate the role of deformation twinning in the strain-hardening behavior of high purity, polycrystalline α-titanium in a number of different deformation modes. Constant strain rate tests were conducted on this material in simple compression, plane-strain compression and simple shear, and the true stress (σ)-true strain (e) responses were documented. From the measured data, the strain hardening rates were numerically computed, normalized by the shear modulus (G), and plotted against both normalized stress and e. These normalized strain hardening plots exhibited three distinct stages of strain hardening that were similar to those observed in previous studies on low stacking fault energy fcc metals (e.g. 70/30 brass) in which deformation twinning has been known to play an important role. Optical microscopy and Orientation imaging microscopy were conducted on samples deformed to different strain levels in the various deformation paths. It was found that the onset of deformation twinning correlated with a sudden increase in strain hardening rate in compression tests. The falling strain hardening rate correlated with saturation in the twin volume fraction. In shear testing a much lower rate of strain hardening was found, at all strains, and this correlated with a lower density of deformation twinning.

Grain boundary ductile fracture in precipitation hardened aluminum alloys

Abstract Published microstructural studies of grain boundary (gb) fracture in precipitation hardened aluminum alloys are reviewed with respect to the three main ideas that have been developed to explain the gb fracture surfaces. The ideas are 1. (1) microvoid growth at large gb precipitates, 2. (2) strain localization in the soft, and sometimes solute-free, gb precipitate free zones (pfz) and 3. (3) the influence of matrix precipitate shear giving rise to inhomogeneous “planar” slip that may apply large stress concentrations to the gb at the end of slip bands. Although the last two processes have a supporting role in many cases, the published evidence strongly suggests that the first process is of overwhelming importance. This conclusion has been tested by reversion experiments in model Al-Li alloys in which microstructures with increasing area fractions, A f , of large stable δ precipitates (Al-Li) were produced, but with equivalent matrix structures and yield strengths. The materials show marked falls of toughness and of fracture strain as A f was increased. Studies of surface slip markings in the Al-Li alloys suggested that slip was initiated at the large gb δ precipitates. Only very limited evidence for a role of planar slip in the fracture of the Al-Li alloys was found in contrast to observations on high purity Al-Zn-Mg-Cu alloys where planar slip seemed to show more importance. Brief studies on a nickel based alloy, MAR-M200, suggested that even in the absence of a pfz, strong room temperature embrittlement by gb precipitates was produced. The results of this study suggest that the marked problems of gb fracture in Al-Li alloys are associated with large gb δ precipitates. Jensrud and Ryum [ Mater. Sci. Engng 64 , 229 (1984)] have shown how gb precipitate growth is facilitated in this system as the gb δ phase is very much less soluble than the strengthening δ′ (Al 3 Li) phase.

Strain hardening regimes and microstructure evolution during large strain compression of high purity titanium

Abstract The sudden increase of strain hardening rates seen after small strains in titanium, was shown to correlate with the onset of deformation twinning. This result appears to match quantitatively with Hall–Petch grain size strengthening. The new twin boundaries appear to reduce the effective grain size.

Evolution of grain-scale microstructure during large strain simple compression of polycrystalline aluminum with quasi-columnar grains: OIM measurements and numerical simulations

Abstract Polycrystalline deformation and its modeling by currently used crystal plasticity models has been investigated by means of an experiment involving direct measurement of deformation induced orientation changes. The experiment used a polycrystalline aluminum sample with quasi-columnar grains, whose initial lattice orientations were mapped using the Orientation Imaging Microscopy (OIM) technique. The sample was then compressed 40% (along the axis of the columnar grains), and the lattice orientations after deformation were studied by OIM. It was found that most of the grains had significant in-grain misorientations in the form of deformation bands with two morphologies — either elongated on the grain scale or nearly equiaxed. In many, but not all cases, more than one similarly oriented deformation band was found in an individual grain. The deformation was then simulated using (i) a classical Taylor-type model, and (ii) a finite element model of the polycrystalline aggregate imposing equilibrium and compatibility between and within the constituent grains (in the weak numerical sense). A comparison of the predictions with the experimental results indicated that the Taylor-type model captured well the tendency to move towards a fiber texture but failed to predict correctly which pole was rotating towards the compression axis in the individual grains, and also by its implicit assumptions could not predict any in-grain misorientation. The finite element model predicted, reasonably well, grain rotations as well as the magnitude of the in-grain misorientations in most, but not all, of the individual grains, but failed completely to predict the morphology of the deformation bands that developed within the grains.

Recrystallization and texture

is, I hope, appropriate in a paper reviewing recent advances in the field of recrystallization at a meeting celebrating the contributions and achievements of a distinguished scientist to start with a few personal comments. My first contact with Robert Cahn was in my final year of graduate research at Oxford when he visited the Department of Metallurgy, then headed by Professor William Hume-Rothery. Robert Cahn was not the first person outside the department with whom I was able to discuss my research, but he was able to offer to me what I subsequently saw him offer to many other scientists: (i) real enthusiasm for new ideas, (ii) innovative suggestions for new experiments and (iii) very strong encouragement to a young investigator. He followed this by providing real help in that he obtained for me, from the publishers of a major conference on recrystallization, (I) the prepublication set of the conference papers that I critically needed for my doctorate thesis. (2’ Three years later he encouraged me to apply for, and he subsequently appointed me to, a lectureship at Sussex University where he was establishing the first undergraduate and graduate programme in Materials Science in the U.K. The opportunity this allowed, to be in at the start of an exciting if short-lived experience of founding a new subject in a new university, was perhaps the most stimulating teaching experience in my life. Robert Cahn and I wrote, in 1983, an accounV3’ of the brief history of Materials Science at Sussex University from when it was founded in 1965 to its closure in 1982/3. There is no need here to say more about that experience other than to quote the following from the introduction: “Creating and building up a group which achieved a worldwide reputation was deeply rewarding for all of us, in spite of the emotional stresses of the final period”. In the early years at Sussex while Robert Cahn was continuing his interest in recrystallization with two excellent research students, Dr Manu Bhatiac4) and Dr Edmund0 Chojnowski,‘5) and producing several scholarly reviews of the topic,‘“9’ I was following

Deformation texture transition in brass: critical role of micro-scale shear bands

Abstract The transition in deformation textures between low stacking fault energy f.c.c. metals (e.g. brass textures) and the medium to high stacking fault energy f.c.c. metals (e.g. copper textures) is addressed. A detailed microscopy investigation was conducted in parallel with texture measurements on deformed samples of copper and 70/30 brass to different strain levels in three different deformation paths, namely, plane strain compression, simple compression, and simple shear. The objective of the study was to identify the specific trends in the transition between the brass textures and the copper textures that correlated with the onset of deformation twinning and those that correlated with the onset of micro-scale shear banding. It was found that several important transitions in the evolution of the deformation textures, especially in the rolled samples, correlated not with the onset of deformation twinning but with the onset of micro-scale shear banding. These results strongly suggest that the critical feature in texture transition is not twinning directly, but the shear banding promoted by the high strain hardening rates of low stacking fault energy f.c.c. metal.

Inhibited coarsening of solid-liquid microstructures in spray casting at high volume fractions of solid

Abstract Experimental studies on coarsening in fine grained solid-liquid microstructures at high volume fractions of solid ( f s ) have been carried out to determine if inhibited coarsening under these circumstances could account for the anomalous fine cell sizes observed in spray castings. The materials investigated included a chill-cast dendritic binary alloy of Al-Cu, two spray cast alloys—AA2014 and Cu-Ti, whose grain size was the segregate spacing, and a d.c.-cast alloy Al-4.5 wt% Cu-1.5 wt% Mg in both coarse-grain and grain-refined conditions. The observed segregate spacings after coarsening were smaller than that predicted by empirical correlations of dendrite arm spacing and freezing time. In all cases, the coarsening was found to become slower as the temperature was reduced and f s increased. Conventional coarsening theories and experiments predict the opposite, i.e. faster coarsening at higher volume fractions of solid. Two additional coarsening models were developed for the grain growth at high volume fractions of solid by processes whose rates are limited by migration of liquid at grain boundaries as liquid films on 2-grain surfaces or liquid rods on 3-grain triple points. In both models, the conventional diffusion-limited t 13 coarsening law was reproduced, but the rate constant K contained the term 11−f s and so also predicted accelerated coarsening as f s → 1 . Three possible explanations for the observed lower K values at increasing f s are proposed. The first is the effect of the increasing difference between the solute contents of solid and liquid as the temperature is reduced. This produces a1/X 1 dependence of the coarsening rate constant K . The second inhibiting effect, specific to dendritic structures, is in-grain coalescence of dendrite arms at high f s which produces isolated liquid particles within the grains. The final possibility is particle-inhibition of grain boundary migration by minority (impurity) particles at the grain boundaries. Such particles were seen, however, for only two of the alloys, viz. the grain defined d.c. cast Al-4.5 wt% Cu-1.5 wt% Mg and the spray case AA2014, but they or gas-filled pores are proposed as strong possibilities to account for the fine grain sizes observed in all spray cast microstructures.

On the volume fraction dependence of particle limited grain growth

The purpose of this paper is to offer an analytical model that accounts for the different variation of the limiting grain radius with volume fraction of particles found in previous computer simulations as compared to the classic Zener predictions. The effect of particles, seen both in the simulations and in the predictions, is to cause grain growth to cease at some limiting grain size. The source of the discrepancy arises from Zeners assumption of random particle/grain boundary intersections. The computer simulation, on the other hand, shows a high correlation between boundaries and particles.

Detailed analyses of grain–scale plastic deformation in columnar polycrystalline aluminium using orientation image mapping and crystal plasticity models

Deformation studies at grain level have been performed in order to model how individual crystals in a polycrystalline material deform. The experiment was carried out by plane–strain compression of a high–purity polycrystalline aluminium with columnar grain structure with near ⟨100⟩ fibre texture parallel to the constrained direction in the channel die. This structure was chosen to allow a fully three–dimensional characterization of the grain structure. The grain orientations were mapped by orientation image microscopy, as the directionally solidified material was deformed in steps of 10% to a total height reduction of 40%. The grains were found either to show nearly uniform rotations or to split into two types of deformation bands, either with repeating orientation fields or with non–repeating orientation fields. The Taylor model and the finite–element method (FEM) were, as usual, quite successful in predicting the average deformation texture, but the Taylor model failed totally to predict the rotation of individual grains. The FEM was more successful in predicting the individual grain rotations but did not, as in a previous study, predict the morphology of the deformation bands. The significant discovery, made here, was that it appeared possible to model the local deformation at a grain scale, from the observed individual deviations of the grain rotations from those predicted if each grain underwent just the plane–strain conditions imposed on the sample. Plastic work rates were computed allowing four shears (two shears in each of the two contact planes) that are compatible with the channel–die geometry. It was found that in all the ‘hard’ grains (those with high Taylor factors), the additional shears (in type and magnitude) that minimized the plastic energy dissipation rate were the same shears that were needed to match the observed grain rotations. Adjacent Taylor ‘soft’ grains were found to have been subjected to the additional shears imposed by their neighbouring hard grains. This was true even when these shears raised the plastic work of the soft grains. This effect was most marked when the soft grains were small in size. These additional shears found by this plastic work analysis were consistent with the observed additional shear seen in the overall shape change of the sample. The grains forming non–repeating orientation fields had low initial Taylor factors and were surrounded by high–Taylor–factor grains, usually of larger size, but which had adopted somewhat different extra shears. The grains showing repeating orientation fields were found to have an orientation, near ‘cube’, (001) ⟨100⟩, which was initially unstable, leading to a break–up into different orientation fields when deformed. These differing deformation bands in the cube grains followed different strain paths, which also minimized their plastic work.

Direct observation of the development of recrystallization texture in commercial purity aluminum

Abstract To test if oriented nucleation or oriented growth causes recrystallization texture, orientations of new grains were measured in a study of cube texture in commercial-purity aluminum after warm, near plane-strain, extrusion. After full and partial recrystallization the frequency of near-cube grains (within 20° of cube) was very high, 40–75%, compared to the 4% frequency expected for random nucleation. The mean sizes of near and non-cube grains were however indistinguishable at all stages of recrystallization. Quantitative (ODF) studies of the textures of just new grains after partial recrystallization showed cube intensities of 48, 30 and 24 (times random) and values of 16 and 24 after 100% recrystallization. Brief studies of cold-rolled A1, which gave negligible cube texture showed near random frequency of cube nuclei on partial recrystallization. The present study, with that of Hjelen et al. [Acta metall. mater.39, 137, (1991)], shows that strong cube textures in polycrystalline aluminum come from oriented nucleation.