Deconvolved Images

Data from the deconvolved 3D data set are shown in Figure 13.12. To visualize the 3D sensitivity we have zoomed-in on the data set at the position of about the middle of the Golgi stack as shown in Figure 13.11. Figure 13.12a shows an image from the original (before deconvolution) data set at the same focus position as Figure 13.12c. Three of the deconvolved images are shown in Figure 13.12b-d; each image differing 50 nm in the focus position. The deblurring effect of the deconvolution is clearly visible, similar to that found with deconvolution in wide field light microscopy [14]. Between the images of Figure 13.12b-d, numerous changes in the structures of the membranes and filamentous structures within the Golgi stack can be seen. Several oval dashed lines with different colors are added as a guide to the eye to compare the same structures in each image. The positions were chosen arbitrarily to provide a few examples of structural changes as a function of the focus position. For example, in the top oval in

FIGURE 13.12 Deconvolved 3D STEM images of a conventional thin section. (a) Original (before deconvolution) image at the same focus position as image (c). (b)-(d) Images of the deconvolved data set each differing 50 nm in focus position. The oval dashed lines are added to guide the eye.

Figure 13.12d we can see a continuous line (a membrane structure) that becomes interrupted going through Figure 13.12c and b. In the oval second from the top, a tubular shape is visible in Figure 13.12d, which disappears in Figure 13.12c and a different structure is visible in Figure 13.12b. In the bottom oval an opening between two structures is visible in Figure 13.12c, while it is closed in Figure 13.12d. In the remaining three ovals similar changes can be observed and various changes can be discerned at other positions.

This data thus provides the first proof that 3D STEM can be applied successfully to biological samples with a depth resolution much smaller than the sample thickness. We conclude also that deconvolutiion can be used to enhance the 3D resolution. More accurate deconvolution procedures could be developed to take account of the variation of the PSF with the focus position, which would allow a higher depth resolution in the 3D reconstruction.

0 0

Post a comment