AVS 51st International Symposium
    Thin Films Monday Sessions
       Session TF-MoP

Paper TF-MoP22
Ultra High Resolution Tomographic Reconstruction Using Scanning Electron Microscope and Focused Ion Beam

Monday, November 15, 2004, 5:00 pm, Room Exhibit Hall B

Session: Poster Session
Presenter: R.K. Bansal, University of Virginia
Authors: R.K. Bansal, University of Virginia
R.V. Hull, University of Virginia
J.M. Fitz-Gerald, University of Virginia
Correspondent: Click to Email
The ever shrinking device dimensions in the semiconductor industry and the advent of nanotechnology has necessitated the developments of new characterization techniques which can work at smaller length scales with high precision. The present study aims at demonstrating a tomographic reconstruction technique capable of reproducing the morphology and chemical structure of a material with a sub-10nm resolution. This has been achieved by serial sectioning of the material using focused ion beam (FIB) and subsequent imaging using a high resolution field emission scanning electron microscope (FE SEM). This process was repeated to obtain a series of two dimensional cross sectional images. These images are then concatenated in the computer and interpolated into three dimensional space to asses and visualize the structure of the material. Multilayered structure of alternating 14nm thick Si-Ge/Si layers was reconstructed using linear interpolation of slices while a slightly more complex shape based interpolation routine was utilized to reconstruct @theta@' Al2Cu precipitates in Al matrix. As expected, the precipitates were found to be 10-20nm thick plates oriented along the orthogonal (100) directions even though this was not apparent from individual slices. High resolution alignment procedures were applied using FIB milled trenches and their profile was verified using cross sectional TEM. These trenches were also used to accurately predict the distance between individual slices. This is especially important as drift during milling in FIB often results in unequally spaced slices. Interpolation and visualization was performed using routines developed in MATLAB environment.