Biomedical Engineering Reference
In-Depth Information
Chapter 5
AFM data needs to be processed, displayed and analysed to get the most out of it. AFM
data has several advantages over data from other microscopes in terms of data analysis.
Firstly, the data is in digital form, therefore there is no digitization step required. As AFM
data is already calibrated, there is no need for a scale bar overlaid over the images because
the scale is built into the data file. AFM data is by its nature three dimensional, so there are
no calculations or guesswork involved in getting three-dimensional data. The height data
is in the files and is easy to measure. Finally, AFM data often has different channels of
information that are all recorded at exactly the same time, and in exactly the same position,
removing the need for any 'alignment' steps. The data is saved by the AFM software as a
file containing all the height and other data recorded by the instrument. Typically, the
AFM software provided by the instrument's manufacturer has two main functions, image
acquisition, and image processing. These could be separate programs, or may operate
together. The latter part allows the user to open the file, and carry out the operations that
will be described in this chapter. AFM data files are always proprietary formats; the data is
not stored as simple image formats such as .tiff, .jpeg or .bmp which can be used by other
microscope types. Such formats are not suitable for AFM images, because they cannot
record the three-dimensional height data of AFM images, nor the additional data that can
be obtained along with the height data (phase shift, cantilever deflection, friction, etc.).
In addition to the actual data points, AFM files usually contain a lot of other information,
such as scanning parameters, etc. The ADCs in AFMs usually record the data as 16 bit
numbers (see Chapter 2), so the data must be stored also as 16 bit numbers to maintain the
best resolution in the data. Optical microscope images and other photographs are usually
stored as 8 bit numbers (256 shades per colour channel), because this is as many shades as
the human eye can discriminate. Thus, these image formats are not adequate for AFM
images, which usually contain far more information and resolution than can be simply
visualized by eye, at least along the
z
(height) axis.
AFM file formats
AFM files vary widely in format from one instrument to another, but most formats are
along the following lines. The start of the file will contain a large 'header' chunk, which
may or may not be human-readable, but is easily read by AFM software. This is followed
by a data section, containing the actual images. The header will normally contain some or
all of the scanning parameters used to obtain the image (scan speed, gain settings, modes
used, etc.). The most important function of the header is that it contains information
allowing the second part of the file - the actual data - to be decoded. That is, it declares the
scan size, and the 'conversion factor' which allows decoding from the 16 bit numbers in
the data section into real data (distances, phase shift values, deflection values, etc.). It does
not make sense to record all the data as real numbers, as this would reduce the precision
available in the data. However, this storage method, combined with the large number of
