Recently developed techniques to measure the structure of interfacial networks in three dimensions have the potential to revolutionize our ability to control the microstructures of polycrystals and interface-dominated materials properties. This paper reviews recent findings from two-and three-dimensional orientation mapping studies. The observations confirm a strong inverse correlation between the relative energies of grain boundaries and the frequency with which they occur in microstructures. The observations also show that during microstructure evolution, relatively higher energy grain boundaries are more likely to be shrinking while lower energy interfaces are more likely to be growing. These processes can lead to a steady-state distribution of grain boundaries that is influenced as much by the relative grain-boundary energies as by the exact processing conditions. Recent findings and emerging opportunities for grainboundary characterization are reviewed in the final section of the paper.Automated EBSD, which makes it possible to accumulate a map of orientations on a surface, has had a transformative effect on the study of grain boundaries in polycrystals (see Panel B). 10 It is Panel A. Description of multidimensional representations of grain boundary character distributions.Three different representation of the distribution of relative areas of different types of grain boundaries, using the same data. This example is for a hexagonal material, WC. (a) This is a single-parameter disorientation distribution. Each boundary is classified by its minimum misorientation angle, without consideration of the misorientation axis. Compared with the random distribution, there is an enhancement of grain boundaries with 301 and 901 disorientation. (b) When one considers the axis and angle of the misorientation, there are three independent parameters for each boundary, two for axis direction and one for the rotation angle. Therefore, each position in a three-dimensional space corresponds to a different grain boundary. In each layer of the axis angle space, all possible axis are represented; the rotation angle varies along the vertical direction. Note that the peak for 301 in (a) is concentrated at [0001], indicating that these are mostly 301 rotations about [0001], and the peak at 901 is concentrated at [10-10], indicating that this is the dominant misorientation axis. For any particular axis angle combination, there is a distribution of grain-boundary planes, as shown for the examples in (c) and (d). Note that when the misorientation axis, the misorientation angle, and the grain-boundary planes are specified, there are five independent parameters. The grainboundary plane distributions (c) and (d) are the relative areas of different grain-boundary planes at for the misorientation of 901 about [10-10] (c) 301 about [0001]. The plots are stereographic projections. The maxima show that the 901 [10-10] misorientation boundaries are concentrated on [10-10] planes, indicating that they are also pure twist boundaries. The local max...