An Overview of Accomplishments and Challenges in Recrystallization and Grain Growth

The study of microstructural evolution in polycrystalline materials has been active for many decades so it is interesting to illustrate the progress that has been made and to point out some remaining challenges. Grain boundaries are important because their long-range motion controls evolution in many cases. We have some understanding of the essential features of grain boundary properties over the five macroscopic degrees of freedom. Excess free energy, for example, is dominated by the two surfaces that comprise the boundary although the twist component also has a non-negligible influence. Mobility is less well defined although there are some clear trends for certain classes of materials such as fcc metals. Computer simulation has made a critical contribution by showing, for example, that mobility exhibits an intrinsic crystallographic anisotropy even in the absence of impurities. At the mesoscopic level, we now have rigorous relationships between geometry and growth rates for individual grains in three dimensions. We are in the process of validating computer models of grain growth against 3D non-destructive measurements. Quantitative modeling of recrystallization that includes texture development has been accomplished in several groups. Other properties such as corrosion resistance are being related quantitatively to microstructure. There remain, however, numerous challenges. Despite decades of study, we still do not have complete cause-and-effect descriptions of most cases of abnormal grain growth. The response of nanostructured materials to annealing can lead to either unexpected resistance to coarsening, or, coarsening at unexpectedly low temperatures. General process models for recrystallization that can be applied to industrial alloys remain elusive although significant progress has been made for the specific case of aluminum alloy processing. Thin films often exhibit stagnation of grain growth that we do not fully understand, as well as abnormal grain growth. Grain boundaries respond to driving forces in more complicated ways than we understood. Clearly many exciting challenges remain in grain growth and recrystallization. Grain Boundary Properties: Energy It is not difficult to appreciate that the energy of grain boundaries varies with type even in the most symmetric crystal classes. Dihedral angles, a direct measure of anisotropy of interfacial energy, vary widely from the 120° value expected for isotropy; even in strongly <111> fiber textured thin films of aluminum, the range is from 90-180°. In fact, the dihedral angles can be used to extract the variation in grain boundary over the whole range of five (macroscopic) degrees of freedom required to describe the crystallographic type [1]. Experiments in which a liquid is allowed to penetrate down grain boundaries show that certain boundary types are much more likely to be penetrated than Materials Science Forum Vols. 558-559 (2007) pp. 33-42 online at http://www.scientific.net © (2007) Trans Tech Publications, Switzerland All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of the publisher: Trans Tech Publications Ltd, Switzerland, www.ttp.net. (ID: 128.2.112.172-20/07/07,19:37:25) others [2]. The resistant boundaries are low energy and the penetrable boundaries are high energy types. The received wisdom on grain boundary energy has been that grain boundaries have disordered structures with high energy except for boundaries with misorientations close to coincidence relationships (CSLs). Near-CSL boundaries were assumed to be special, i.e. low energy, because of better atomic fit. However, theory and simulation [3, 4], combined with recent experimental results [1] suggest that special boundaries are those that combine low surface energies, placing the emphasis on boundary normal as opposed to lattice misorientation. Moreover there is a strong inverse correlation evident between grain boundary energy and the relative frequency of boundaries of different types [5, 6]. To choose a simple example in fcc metals, the Σ3 and Σ5 misorientation relationships might be expected to exhibit good atomic fit and low energies: this, however, is not borne out by the experimental evidence and, even for the Σ3, only the pure twist boundary type, with the normal parallel to the misorientation axis to give a coherent twin, gives a low energy configuration. This has been followed up by simulation of grain growth, which has demonstrated that anisotropies on the order of a few percent are sufficient to induce a corresponding anisotropy in the grain boundary population [7]. Some experimental support for an evolution in GBCD that mirrors the energy anisotropy has been reported by Piazolo et al. for grain growth in an aluminum foil with a columnar grain structure [8]. Both an increase in the number of low angle boundaries, and an increase in the number of low-index planes in boundaries was reported. This correspondence offers the tantalizing possibility that it may be possible to infer the anisotropy of grain boundary energy from the anisotropy of the distribution of boundary types, i.e. the grain boundary character distribution (GBCD). Such an approach is likely to require near random texture and sufficient (normal) grain growth, however, to ensure that the GBCD is a product solely of anisotropy in the energy function. Other applications in which the anisotropy of grain boundary energy may play a role include intergranular corrosion which has received considerable attention with respect to the importance of the way in which boundaries are connected together in networks [9, 10]. Grain Boundary Properties: Mobility We define mobility as the (linear) material property that multiplies a driving force for boundary motion to give its velocity [11]. Most of the knowledge on grain boundary mobility has been determined only for cubic metals, and, amongst that class, nearly all for (fcc) aluminum of various compositions. The general characteristics of mobility are that it is strongly dependent on temperature, grain boundary character and purity. The temperature dependence can be understood in terms of thermal activation of atom transfer across the grain boundary [12]. Although some attempts have been made to develop this approach and account for grain boundary character, the results have not been to explain the essential features of the anisotropy of mobility [13, 14]. The experimental results for Al show that the enthalpy of mobility varies strongly with boundary type over a range that is difficult to reconcile with the existence of a single type of atomic transfer mechanism. Some insight into the complexity of boundary motion has been found in detailed analysis of the complex individual atomic motions that occur during migration in molecular dynamics simulations [15]. Overall, for pure aluminum and grain boundary character characterized by low-index misorientation axes, tilt boundaries move more rapidly than twist boundaries and the ranking goes as 111>100>110. The relative activation energies, Q, are in the opposite order with Q111 < Q100 < Q110 [16]. Given the need for constitutive descriptions of grain boundary properties for use in modeling microstructural evolution, one interesting possibility is to attempt to generate a master curve for various different materials. The discussion is once again confined to fcc metals. One approach is to scale both the pre-factor and the activation energy for mobility as a function of the solute content but assume that the variation of mobility with crystallographic type is independent of solute level. As an example, activation energies for several different experimental and simulation results by Recrystallization and Grain Growth III 34