The available technologies for the direct three-dimensional characterization of microstructural features at a millimeter length scale are based on automated metallographic serial sectioning techniques and optical microscopy. Prior to offering commercial sales of the Robo-Met.3D, UES demonstrated the technology on applications ranging from pitch-based carbon foams in a study correlating microstructure with the mechanical and heat transfer properties  to dendritic solidification to porosity and phase fractions in sintered Fe-Cu materials .
Fig. 1 shows an example of a small volume (20 slices approximately 2 mm apart) of reconstructed manually serial sectioned optical images showing the outlines of the prior b grains. The scale in the “stack” direction is not to scale to better illustrate what could be observed once the installation of the Robo-Met.3D is complete. It is clear that the true shape of the grains can be determined. This image also shows grain boundary curvatures clearly (see arrows), as well as a distribution in the sizes of the prior b grains. Clearly, such reconstructions would provide invaluable contributions to the modeling of the full three-dimensional nature of materials.
The Robo.Met-3D Unit at OSU represents the first commercial sale and delivery by UES. Installation was completed in December 2006.
There are three principal components to the Robo.Met-3D, namely an Allied High Tech Multi-Prep polisher, a Rixan robot, and a Zeiss Axiovert 200 optical microscope. All three components are colocated on a 1500 lb artificial granite tabletop in an enclosed system. The software simultaneously controls each of the three components so that a small mount can be polished and have a certain amount of material removed in a controlled, preprogrammed fashion. There are then three stations where cleaning, etching, and neutralizing can occur, prior to imaging, including an autofocusing routine. There is the capability of acquiring mosaic images using the Zeiss Axiovert optical microscope with motorized stage.
Slice and View™ is commercially available software that automates the successive milling ("slice") of a defined volume of material from the bulk as well as imaging the milled surface ("view"). Slice and View allows the user to specify electron and ion beam settings for imaging and milling in order to optimize the procedure. The maximum sectioning depth of the magnum column in the DB235 is 8 microns into the material before manual realignment is necessary. The Nova 600 NanoLab is capable of sectioning 13-15 microns into the material before manual realignment is required. Both FIB systems also have the software RunScript which executes custom scripts that have been written to extend the sectioning depth of both FIBs to the point that time and monetary constraints are the major limitations for data collection.
The Nova 600 NanoLab comes with the newly developed EBS3 software, which was developed in collaboration with CAMM and HKL Technology (Oxford Instruments) as a strategic partner. EBS3 is an optimized extension to Slice and View technology pioneered by FEI. While Slice and View generates serial images for 3D image reconstruction, EBS3 also provides crystallographic information, including serial EBSD orientation maps for quantitative 3D grain reconstruction, using the crystal orientation of each individual voxel within the volume of interest. The automated data collection procedure includes accurate (re)positioning between the milling location and the EBSD location to accommodate the geometry requirements for both operations.
Transmission electron and scanning transmission electron tomography, techniques for the three-dimensional characterization of microstructural features at a nanometer length scale, make use of the fact that, despite their limited thickness, thin foils do have a third dimension. This limited thickness has traditionally caused difficulties in quantifying features observed in TEM micrographs which are projections of the three-dimensions. Such difficulties, including well-known phenomena such as the Holmes effect (i.e., overprojection) and particle truncation [1-3], make traditional 2-D stereological procedures such as volume fraction, difficult or impossible. However, the same third dimension (i.e., foil thickness) which renders traditional stereological methods impractical makes electron tomography possible.
Tomography is a technique where images are recorded from a specimen in as many viewing directions as possible. Thus, tomography is the integration of data collected from a series of images in which the orientation of the specimen relative to the incident beam is progressively varied, and the series then reconstructed. Traditionally in the TEM this has meant recording a series of TEM images at different tilt-angles, typically with a CCD camera and over the range of -70° to +70° at regular tilt intervals of approximately 1° or 2°. This approach has been applied successfully in life science for already more than decade [4-7]. In many cases it is possible to record data series by hand – though it is a very tedious process involving precise tilting, image shifting (one of the most important aspects of tomography to minimize information loss through variation in imaging area), focusing, imaging, and tilting. Researchers performing this manually can only process a very limited number of samples in a given timeframe, resulting in an increase in cost, both operationally (e.g., microscope time) and from an opportunity cost perspective.