317 Modern imaging techniques for metals and alloys Jeff ThM De Hosson New developments have occurred in transmission electron microscopy and scanning electron microscopy, bearing in mind that the principal reason of doing microscopy in the first place, is to scrutinize the structure-property relationship. There still remain, however, fundamental and practical reasons which hamper a straightforward correlation between microscopic structural information with the properties of metals and alloys. It is argued that one should refocus more on the generic features of defects, using a mesoscopic approach including various length scale transitions, and on in situ rather than on postmortem observations of solely atomic structures. Addresses Department of Applied Physics, Netherlands Institute for Metals Research, Materials Science Center, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands; e-mail: hossonj@phys.rug.nl Current Opinion in Solid State & Materials Science 1998, 3:317-324 Electronic identifier: 1359-0286-003-00317 0 Current Chemistry ISSN 1359-0286 Abbreviations FEG field emission gun HRTEM high resolution transmission electron microscopy/microscope Introduction Generally speaking, microscopy in the field of materials sci- ence is devoted to linking microstructural observations to physical properties [l-lo]. In particular mechanical proper- ties of metals and alloys are structure-sensitive. It is often claimed that advanced microstructural investigations require a microscope with a resolving power in the order of a nanometre or even better [3,8-lo]. Although the microstructure-property relationship is in itself a truism, the actual linkage between structural aspects of defects in materials, studied by microscopy, and their physical proper- ties is almost elusive. The reason for this, is that various physical properties are actually determined by the collec- tive behaviour of defects rather than by the behaviour of one singular defect itself. For instance there exists a vast amount of electron microscopy analyses, in the literature on ex S&J deformed metals and alloys, which try to link observed dislocation patterns to the mechanical behaviour of these materials as characterised by their stress-strain curves. However, in spite of the enormous effort that has been put into both theoretical and experimental work, a clear physical picture that could predict the stress-strain curve on the basis of microscopy observations is still lag- ging. There are at least two reasons for the hindrance of a straightforward correlation between microscopic structural information and materials properties: one is fundamental and the other a practical reason. The latter will be explic- itly explained in the final section of this review. Of course it has been realised for a long time that in the field of dis- locations and interfaces in metals and alloys we are facing highly nonlinear and nonequilibrium effects. The defects determining many physical properties are in fact not in thermodynamic equilibrium and their behaviour is very much nonlinear. This presents a fundamental problem because there doesn’t exist an adequate physical and mathematical basis for a sound analysis of these effects. Nevertheless, the situation is not hopeless because nowa- days there are two approaches to circumvent these prob- lems and microscopy still contributes a crucial role. One approach has to do with numerical simulations of the evo- lution of defect structures, which incorporate the behav- iour of individual defects as known from both classical theory and from microscopy observations (of individual dislocations, interfaces and interactions between disloca- tions and interfaces) [ll-191. For example, to have a thor- ough understanding of the generation of cellular dislocation structures, vein structures, tangles, subgrain boundaries and persistent slip bands, important input for these numerical simulations on the behaviour of individual dislocations such as cross slip behaviour, climb and bipolar structures and so forth, are provided by microscopy research. Another approach, which is supplementary to these simulations and partly based upon them, is the con- tinuum mechanics approach. It provides a description of the global co-operative behaviour of defects and focuses on the instability transitions and accompanying structural transformations [ZO-281. Also, here, experimental knowl- edge provided by electron microscopy, in combination with complimentary techniques, is inevitable (e.g. disloca- tion pattern and cell formation) [29]. The situation to correlate the microstructural information obtained by microscopy of an interface to the macroscopic behaviour of polycrystalline solids is even more complex than in the case of dislocations. The reasons for this are numerous, for example there is a limited knowledge of the interface structure (i.e. both topological and chemical, at an atomic level of only a small number of special cases), there is the complexity due to the eight degrees-of-free- dom of an interface and a lack of mathematical-physical models to transfer information learned from bicrystals to the actual polycrystalline form. It has been shown that in some cases it is crucial to have, available, information about an atomic level provided by high resolution electron microscopy, but surely it is not always necessary and some- times it is rather more appropriate to image defects on a micrometre scale instead of correlating the structural infor- mation (obtained from high resolution microscopy) to physical properties. An interesting approach to solve this