A Water Filter and the Watershed Model
All of us expect the water we drink to be crystal clear and clean. Some people even go to the extent of using filtering methods to further clean our tap water. As the residents of Flint Michigan and other areas know, the use of a proper filtration method can be critical for clean water. Of the various filters available for cleaning water, the majority use activated carbon to remove chemicals, improve the taste, and give us that pure water look. Activated carbon is one of the most important aspects of the filter; analyzing each carbon particle and the spaces in between them with micro-CT imaging is now easier with the new watershed function in Bruker’s CTAnalyzer.
Carbon is used for filtering due to its multitude of pores and high absorption capability. There are various types of carbon used in filters, such as bituminous coal, wood, and coconut shell carbon. The best for use as a filter is coconut shell carbon. Regardless of the carbon chosen, they each provide the benefit of removing impurities such as chlorine, pharmaceutical compounds, pesticides, and cleaning by-products.
Micro-CT of the Carbon Granules
A carbon filter cartridge was scanned using our Bruker/SkyScan 1173 system, which is ideal for larger objects with its 6×6 inch scan area. This worked well for the filter which is five inches in height and a diameter of approximately two inches. Only a section of the filter in the center was scanned to highlight the carbon granules for analysis. After the process of scanning, NRecon was used to reconstruct the images into axial slices. These slices where then previewed in DataViewer and CTVox before being loaded into CTAnalyzer for the granule watershed separation.
The Watershed Analysis
Analysis of very fine granules can be tricky. Depending on the resolution, the space between the granules is very fine and, after binarization (selection) of the granules, the particles can appear to be one solid object (Figure 1, left). This can be improved though because our eyes can visually see the different granules. This is where the watershed separation comes into play.
By running the watershed separation found under the morphological operations in CTAn, the algorithm optimizes the gaps between the fine features (in this case carbon granules) and segments them into individual objects (Figure 1, right). This makes a drastic difference in the quantification of the total number of granules which changes from 34 before to 821 after the watershed separation (Table 1). Now the model is more accurate and any additional analysis such as surface area, volume, or porosity can be performed with more confidence.
Figure 1. Comparison of carbon granules before (left) and after (right) the watershed separation
Table 1. Number of Carbon Granules before and after watershed separation
This example is one of many where Micro Photonics has provided MicroCT testing to solve some of industries toughest problems. If you have an application you would like tested or are interested in providing a sample for the next image of the month feel free to e-mail: firstname.lastname@example.org.