Chip Morphology and Features of Interest

Below is a cross section of a machined chip. Three regions are highlighted.

Note that this is all a single phase. In other words we cannot define a binary microstructure as we are usually accustomed. The approach now is to capture the location of boundaries and generate statistics.

Image Processing Follow Up

We feel pretty confident with our developed methodology to accurately capture boundary features. Once boundaries are detected the boundary thickness is reduced to 1 pixel. In this way we eliminate any bias that may be introduced from the perceived boundary thickness.

Preliminary Autocorrelation Results

Left Shear Zone

Middle ‘Undeformed’ Region

Right Shear Zone

Note that the volume fraction of boundaries is roughly 10% which agrees with visual observation. Also note that this volume fraction is not necessarily physically accurate as we reduced boundary thickness to 1 pixel earlier…

You may be able to see some faint trends however it is difficult because the overall boundary volume fraction is rather low…

Taking the log of the autocorrelation results allows for a better visualization relative to the image.

Ln() Left Shear Zone

Ln() Middle ‘Undeformed’ Region

Ln() Right Shear Zone