Deconvoluting error in measurement of low angle misorientation distribution.

Misorientation angle distribution gives information about the type and fraction of grain boundaries present in a material. Since grain boundaries affect various mechanical and functional properties of the material, the distribution of grain boundary misorientation is important in order to evaluate these properties. This becomes particularly important when we want to study the microstructure in finer detail, such as understanding the average misorientation within a grain. One of the techniques increasingly used in past two decades for characterization of grain boundary misorientation is electron back scatter diffraction (EBSD). Reliable detection of very small misorientation angles using conventional EBSD system is quite challenging due to the presence of measurement error. This makes the comprehensive characterization of microstructures difficult and prone to error. In order to prevent such problems, it is important to understand the nature of measurement error and find ways to minimize it. The present work aims to elucidate the effect of measurement error on the observed misorientation angle and its statistical distribution in low misorientation angle regime. A true strain of 0.3 was imposed during cold-rolling of Cu-5%Zn alloy sample. The rolled sample was then subjected to in-situ heating from room temperature to 500 °C (∼0.58 Tm). It was found that the overall measurement error in misorientation distribution consists of random error caused by limited angular precision and systematic error which manifests primarily in the statistical distribution of low angle misorientation. In this work, we show a way to deconvolute this overall error based on the measurement technique. We further show that this systematic error is not limited to any particular measurement technique, rather related to the presence of a lower bound in the measurement.