One of the drawbacks of typical AFM is that the images obtained are not truely three-dimensional due to the method of data collection. The piezoelectric scanner moves the tip or sample in the x and y directions while the cantilever reflects the sample's topography, or the z meazurements. No matter how sharp the tip, the data collected can never access the underside of the sample (tilting of the sample is far too difficult to do).
As illustrated by the profile of a sphere on a flat surface, the AFM image collected by even a sharp tip displays as a bump (shown in blue). This image can be deconvoluted to slightly improve the image (as shown in red) but the image is still never truely three dimensional.
This limitation to the scanning leads to the "petticoat" effect - all images taken of objects having steep walls or undercut regions appear to have flared sides, as shown for the AFM of a compact disc on the left.
To get a true determination of an object's sides, a special AFM with a unique servo- mechanism and special tip are used [for more details see the paper "Method for imaging sidewalls by atomic force microscopy" by Martin and Wickramasinghe, Appl. Phys. Lett. 64, 1994, pg 2498-2500.]. This AFM not only senses changes in the z direction but also in the x direction - when the tip encounters a wall or sudden drop, the piezoelectric scanner stops moving in the x direction and determines if the sample feature r
The tip used here must be able to reach into the underside of the object. A special boot-shaped tip is designed specifically for this. The tip flares out like a bad pair of bell-bottom pants. The bottom of the tip is flat, which means the lateral resolution is poor, but the flared portions can view the side of an object. This is important in fields such as microprocessor manufacturing, where cross-sections need to be monitored.
Going back to our model of a sphere on a flat plane, the data collected using the special AFM and the fancy tip gives a laterally-distorted image (shown in blue) which requires a special means of deconvolution, but the result resembles the sphere's profile (in red).
So you might be wondering, "Can't I get a similar image using a typical AFM and the boot-shaped tip?" Well, no. In a typical AFM, the tip would move along the substate until it undercuts the sample. The piezoelectric scanner continues, moving the tip past the sample. If the tip doesn't break or scrape the sample off the surface or bend terribly, it will faithfully follow the surface. The resulting image will under-estimate the position of the advancing object, and the size of the tip will cause an
Last revised on 02-06-2001 by Peter Markiewicz