A TEM Kikuchi pattern study of ECAP AA1200 via routes A, C, B_C

Abstract

Shear plastic deformation in ECAP occurs through the formation, movement and storage of dislocations. The differences which exist between shearing characteristics lead to important implications concerning the optimum processing route. It is known that the effectiveness of cell evolution into an array of high-angle boundaries (HABs) is in the order B_C > C > A, or A > B_C > C, depending on the two-channel-angle intersection of the ECAP die. In route A, large portions of HABs are continuously and progressively generated, while routes C and B_C cause fully redundant deformation at each 2n passes, and each 4n passes, respectively. This study is focused on dislocation generation, storage and recombination during ECAP for routes A, C and B_C. Kikuchi bands identified with TEM were used to quantitatively measure the cell and grain boundary misorientation. ECAP was performed on an AA1200 commercially pure aluminium alloy up to ε = 8.64. A different hierarchy for HAB generation efficiency was found. Thermal stability was studied by annealing the alloy at 0.5, 0.6, 0.7 T_M (where T_M is the alloy melting point) for 2 h after the severe plastic deformation. It appeared that, even if route B_C involves the fastest microstructure grain refining, route C is likely to be the most stable upon reheating.

Introduction

Due to their superior mechanical and physical properties, ultra-fine-grained materials are currently attracting a great deal of attention. These properties include increased strength, low temperature superplasticity, and the possibility to produce new classes of materials, like high-strength aluminium alloys, that do not require age hardening [1], [2], [3]. Ultra-fine-grained materials may be prepared using various techniques including gas condensation and subsequent in situ consolidation under high vacuum or high energy ball milling and they may be produced in a wide range of materials (e.g. pure metals, metallic alloys including superalloys, intermetallics, and semiconductors) by subjecting these materials to a very high plastic strain [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19]. Cost-effective methods that can be used to produce bulk materials are of particular importance in this context. Severe plastic deformation (SPD) techniques are able to produce bulk materials with fine-grained structure. Among the various SPD techniques, equal-channel angular pressing (ECAP) [18] is able, in certain circumstances, to give grain sizes in the submicrometre range of 400–800 nm [20], [21], [22]. In particular, ECAP is an especially attractive processing method because it allows large bulk samples to be produced which are free from any residual porosity, inducing small shape changes (notably almost no cross-sectional change).

The evolution of the microstructure during ECAP, is closely driven by the specific pressing conditions, which means by the shearing deformation induced in the material at each pass through the die [23]. Plastic deformation of metals occurs as a result of the formation, movement and storage of dislocations. In fact, microstructure evolution during ECAP is directly linked to a complex evolution of dislocation networks and their recombination and annihilation phenomena. Imbalance of dislocations of opposite signs moving in opposite directions (or having some Burgers vector components of opposite sign and sliding direction) leads to accumulation, generally at the grain boundary, or at high-angle boundaries (HABs), of a surplus of dislocations of the same sign. The groups of excess dislocations in fact stimulate crystallographic slip. Therefore, the evolution and accumulation of misorientation across both low-angle boundaries (LABs) and HABs is closely linked to the crystallographic accommodation of each cell, grain and the neighbouring crystallites during the shearing deformation. In deformation route A the billet is extruded without rotation between passes (unidirectional straining path); route C is defined as a rotation of 180° between passes and route B_C as a rotation of 90° in the same direction between passes [4], [13], [24], [25], [26], [27], [28], [29].

In a study of pure aluminium, deformed at room temperature by a 90° equal-channel die, at a strain ε = 4, Iwahashi et al. [11], [20] reported the effectiveness of cell evolution into an array of HABs as B_C > C > A. In accordance with this observation, they found route B_C to be the most effective route for producing equiaxed fine grain structure with HABs. Segal [30] arrived at the same conclusion showing that each pass of ECAP is likely to develop thin shear bands oriented along the shear plane. The above mentioned differences in shearing characteristics lead to important implications concerning the optimum processing route. In fact, grain refining is expected to be less effective in route A compared to routes B_C and C. Route B_C appears to be preferable to route C, which lacks any deformation in the plane normal to the pressing direction [4], [24]. In route A, the shear bands that form during the second pass intersect the billet axis at an angle of 45°, while the shear bands that form in the first pass change orientation to an angle of about 20° to the billet axis. The same procedure is repeated at the following passes and this continuously transforms the previous shear bands into an elongated shape along the flow direction. Route B_C is effective in refining the microstructure, because four sets of shear bands are created after a fully redundant cycle (4n passes). Gholinia et al. [13] offered an alternative to Segal's conclusions. They suggested that shear strain is accumulated in the same direction by route A, making this route the most effective method for generating HABs. Both routes B_C and C are redundant strain processes, and consequently are less efficient in creating HABs. However, Zhu and Li [4], [31] suggested that combinations of shear plane with texture and crystal structure play a primary role in grain refinement, while the accumulative strain plays a secondary role. These authors also claimed that their theory could explain why B_C is the most effective deformation route in grain refinement for the 90° die while route A is the most effective for the 120° die.

Under severe plastic deformation (SPD), and for every possible route in ECAP, dislocation boundaries evolve into a regular pattern of grain subdivisions belonging to two scales [32], [33], [34], [35], [36], [37], [38]: large-scale long and continuous dislocation boundaries, called geometrically necessary boundaries (GNBs), and small-scale incidental dislocation boundaries (IDBs) [33]. The former include micro-bands (MBs) and dense dislocation walls (DDWs) surrounding groups of equiaxed cells. The group of cells enclosed by GNBs are called cell blocks (CBs), while IDBs include ordinary cell boundaries. The boundary spacing and misorientation angle distribution of GNBs and IDBs evolve differently as a function of strain; in particular, they exhibit different morphologies at small to medium strains, but similar at high strains [32], [33], [34], [35], [36], [37], [38]. Generally, misorientation axes for IDBs are randomly distributed, whereas GNB orientation distribution clusters on preferred axes and some GNBs show a pattern of rotation around the transverse direction [39], [40], [41]. GNBs play an important part in the deformation process. In fact, to accommodate the strain imposed during deformation, additional dislocations arrange to form geometrically necessary high-dislocation regions associated with dislocation interactions and slip patterns that develop within a deforming grain [40]. Thus, the distance between neighbouring GNBs decreases with strain, approaching, as a limit value, the cell spacing, which decreases much less with strain. Therefore, whatever the route used, grains subdivide under the effect of newly introduced HABs on finer and finer length-scales, until ultimately a limit is reached where the HAB spacing converges with the cell size [42], [43]. The number of cells between GNBs therefore decreases with increasing strain, with a consequent evolution from cells that have only other cells as nearest neighbours towards cells that have both GNBs and cells as nearest neighbours [39]. The IDB mechanism of formation and evolution is the statistical trapping of glide dislocations, whereas the formation of GNBs is associated with differences in the slip systems operating in neighbouring crystallites or, alternatively, involving similar slip systems but having different shear amplitude [44]. In this study, the Kikuchi pattern approach was followed to determine the boundary character. Thus, it was possible to directly identify cells as IDBs, and HABs as GNBs.

The average spacing between dislocation boundaries (LABs and HABs) is a common microstructure parameter in deformed materials due to its inverse relationship with flow stress, as amply documented in literature [40], [45], [46]. Studies of misorientation angles and of the early cell formation process are less investigated (e.g. [2], [7], [12], [30], [47], [48]). Experimental studies on the correlations between the deformation pattern and the resulting microstructure have been carried out, especially by transmission electron microscopy (TEM) inspections [21], [33], [48], [49], [50], [51]. These observations have shown that a structural description, based on individual dislocations or dislocations arranged in smaller groups (tangles and cell walls), is valid at low strains. However with increasing strain the characteristic structural features are dislocations with relatively large misorientation across the boundaries.

Thermal stability is an important feature for materials subjected to severe plastic deformation. Refined microstructures must maintain a reasonably refined structure upon heating at least up to 0.5 T_M (where T_M is the alloy melting temperature). However, their high defect content in the form of crystal boundaries makes them susceptible to structural coarsening.

Several studies have reported the superplasticity behaviour of ECA pressed aluminium and other alloys, which ultimately implies a rather good thermal stability of the severely deformed microstructure [52], [53], [54], [55]. Langdon et al., in several studies [56], [57], [58], [59], [60], [61], have reported that a refined microstructure has a meaningful mechanical application only if this refined structure is not completely destroyed (i.e. recrystallized) upon heating. Of course, there will be temperature threshold-like limits that will depend on the specific aluminium alloy (in terms of secondary phases and grain refining dispersoids), the thermo-mechanical status of the alloy and the level of SPD.

Microstructure inspections of the deformation structures in ECAP materials have been studied by transmission electron microscopy (TEM) and by electron backscattered diffraction (EBSD) using FEGSEM. For example, Iwahashi et al. [10], [11], Liu et al. [32] studied microstructure using TEM, and Gholinia et al. [13], El-Danaf [62], Reihanian et al. [63] used high-resolution EBSD to measure boundary misorientation. However, the latter procedure involves a systematic overestimation of the HAB fraction. In fact, to date, the lower limit of boundary misorientation accuracy of the FEGSEM is typically of ϕ ≥ 2°. In practice this limitation requires the use of more reliable techniques, such as analysis of Kikuchi patterns in TEM, in order to achieve meaningful quantitative measurements [32], [48], [64], [65], [66]. For this reason, in this study, the Kikuchi pattern method was used for all the misorientation and size measurements of both cells and grains (LABs and HABs). The microstructure evolution and refining process, in the three routes (A, C, and B_C), was studied up to the strain ε = 8.64. Thermal stability studies were carried out for three different temperatures (0.5 T_M, 0.6 T_M, and 0.7 T_M) for all three routes: A, C and B_C.