Practical Electron Microscopy and Database

An Online Book, Second Edition by Dr. Yougui Liao (2006)

Practical Electron Microscopy and Database - An Online Book

Chapter/Index: Introduction | A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z | Appendix

Reconstructing a Three-Dimensional (3D) Image from 2D Images in STEM Tomography

Tomography involves reconstructing a three-dimensional (3D) image or model from a series of two-dimensional (2D) images taken at different angles (a tilt or rotation series). For this reconstruction to be accurate, the intensities in the 2D images must represent a linear projection of the 3D object. This means that the image intensities should directly correspond to the integrated values of a property (such as mass or atomic number) along the projection direction.

In STEM high-angle annular dark-field (HAADF) imaging, the HAADF signal is largely insensitive to diffraction effects and is approximately proportional to the square of the atomic number (Z-contrast). This makes the HAADF signal nearly linear with respect to the material's thickness and composition, fulfilling the linear projection requirement. As a result, HAADF images can be accurately combined into a tomographic series to reconstruct the 3D structure of the material. Without satisfying the linear projection requirement, the reconstructed 3D model could be distorted or inaccurate, making it difficult to extract reliable structural or compositional information from the tomogram.