Image analysis to understand and control sintering processes

One of the main challenges in the science of sintering is to be able to mon itor the microstructure of sintered materials. This needs a detailed unders tanding of microstructural evolution during sintering, from the particle pa cking to the full density material. Image analysis is a classical tool to g et quantitative data on materials microstructure. it has been extensively u sed on sintered materials. The present paper reviews the recent literature in the field of the application of image analysis to study sintering proces ses from microstructure. It does not account for the large amount of work w hich aims to analyse the relationships between microstructure and physical properties. The main phenomena which occur during sintering are used as a g uideline for the present discussion the various parameters and methods. The simplest parameter is volume fraction of phases, including porosity Its measurement by image analysis is mainly useful in multiphase systems for l ocal measurements in inhomogeneous materials, or in the purpose of quality control of images for further measurements. Mean size and size distributions are measured in order to analyse particle growth and to test growth laws. Chord lengths and equivalent diameter are t he most commonly used parameters, through average values or distributions. Measurements are performed on 2D images, while sintered materials are three dimensional. Either 3D evaluation (unfolding) from experimental data or 2D translation of 3D theoretical distributions are performed to enable compar isons. The use of computers enables to fake into account particle shapes wh ich are more realistic than the classical sphere, and can be adapted to eac h studied material, for unfolding methods (iterative methods) as well as fo r getting 2D cuts from model distributions (computer simulation). Abnormal grain growth, which often concerns anisotropic and facetted grains, has rec ently given rise to the developments of methods based on shape analysis. Sh ape parameters are obviously used to characterize the shape of objects, but also to improve the 3D analysis of size from 2D sections: measuring severa l parameters, such as larger diameter or number of edges) enables multipara meter unfolding methods. The measurement of contact numbers between particles needs particular hypot heses and its application is limited to very simple structures Such paramet ers, which cannot be directly measured on random sets of cross sections, ne ed << 30 images >>. Serial cuts have been used for large structures, as the distance between cuts must be small with respect to microstructural elemen ts; recent developments in X-ray tomography are promizing. Microstructure m odelling, from random models for instance, is a powerful way to evaluate su ch parameters, even on rather complex microstructures. Contiguity evaluatio ns, based on direct contact area measurements, is of more general use, and have been widely applied in various fields of sintering. Several methods ha ve been used to characterize homogeneity of phase repartition in sintered m aterials and analyse the consequences on sintering processes. Classic stereological parameters, mainly based on Volume fractions and surf ace areas of different phases and interfaces, which can be applied to any t ype of microstructure, have been used for a long time, although by a limite d number of authors. The interest of these parameters is renewed by recent works which evidence that classic models of sintering do not predict a corr ect evolution of microstructure for actual materials. This also evidences t he need for other sintering models The use of a wide set of complementary p arameters, for instance in the case of sintering of composite materials, le ads to a more detailed description of microstructure evolution, In recent s intering models, the local image analysis data on actual materials are expl icit inputs for the model, so that microstructure evolution can be connecte d to overal densification for instance. This may be the basis of future ana lysis of sintering processes. The tendencies which rises from the examples analysed in this paper are of two natures : from the point of view of measurements, the improvement of the quality of d ata in recent works shows that the image analysis technique has become a cl assic tool in the science of sintering. Even if volume fraction and size me asurement constitute the most frequently used parameters, more detailed des criptions of microstructure are now used. The systematic use of computers e nables more accurate description, and microstructure modelling is a useful complementary tool to get data; from the point of view of sintering science, i.e., of the knowledge of sint ering mechanisms and their effects on microstructure, image analysis has be come the quantitative complement of classic metallography It enables to evi dence the limits of sintering models. Furthermore, quantitative data are th e basis of new models which account for microstructure evolution and enable to analyse the mechanisms in complex situations. Future developments of ne w materials should benefit from these techniques.