Abstract
Submicrometre x-ray computed tomography (CT), referred to as x-ray nanotomography, is a research area attracting much attention nowadays. The major limiting factors in observing compositional structures and features below the micrometre scale are signal-to-noise ratio and loss of information below the x-ray detector pixel size. Conventional image segmentation techniques, such as image thresholding, are not usually sufficient for accurately resolving such microscopic compositional distributions. In this work we carried out multi-energy x-ray CT simulations on a computer generated sample of nanoporous alumina with gold nanoparticles incorporated inside some of the pores. The multi-energy CT data sets served as inputs to our in-house developed data-constrained microstructure (DCM) modelling software, which can accurately predict the 3D chemical composition of a sample from CT data. Different levels of x-ray detector noise were also added to the CT simulations and the DCM chemical phase predictions were analysed under these conditions. The pixel resolution was 15 nm, and the x-ray projection images were re-sampled to lower resolutions to simulate the effect of features smaller than the CT pixel resolution. We found that despite the simulated sample having constituents that possess vastly different x-ray absorption properties, with one constituent having a total linear absorption coefficient up to two orders of magnitude greater than the other, the DCM showed a clear advantage over image thresholding, particularly in the presence of noise.
Original language | English |
---|---|
Article number | 015013 |
Journal | Modelling and Simulation in Materials Science and Engineering |
Volume | 20 |
Issue number | 1 |
DOIs | |
Publication status | Published - Jan 2012 |
Externally published | Yes |
ASJC Scopus subject areas
- Modelling and Simulation
- Condensed Matter Physics
- Materials Science(all)
- Mechanics of Materials
- Computer Science Applications