INSTRUMENTATION AND EXPERIMENTATION - Nanoindentation with Biological Applications - page 71

stiffness or adhesion of each pixel. If that image were of a 1 μm × 1 μm
region of a biomaterial, each pixel would then represent a 2 nm × 2 nm
sample region. One might then imagine that it is possible to measure the
elastic modulus E of a complex surface at every pixel, and create a
highly spatially resolved map of this important mechanical property to
render its quantitative variation along a sample such as a hydrated tissue
or an individual cell. Certainly, it is an attractive goal to correlate the
spatial features of a structurally heterogeneous biomaterial with the
mechanical properties of individual regions. However, although such
maps of mechanical properties (sometimes called force-volume maps to
indicate that they register a range of force-calculated property values
along an x-y image plane) are straightforward to obtain via automated
software, it is difficult for such data collection approaches to accurately
reflect the mechanical properties of the biological material - uncoupled
from mechanical history of adjacent pixels, sample topography, and
ignorance of the indenter probe geometry at the indentation depths
employed.
The data acquisition in such mapping is in contrast to instrumented
indentation of heterogeneous samples, where indentations are small
enough that the (plastically) strained indentation volume is less than the
phase volume, and spaced widely enough that these strained volumes do
not overlap but sample distinct phase regions. If the material were
perfectly flat and the indentation responses perfectly elastic, such that
no prior deformation was stored within the material adjacent to the
current indentation volume, this pixellated mapping could render an
approximately accurate estimate of elastic properties. If only the latter
condition is met but the sample has surface roughness on the order of the
nm-scale indentation depths, the assumptions of normal contact on the
surface may not be met and thus the estimated elastic properties will be
inaccurate due to errors in the probe-surface contact area. Additionally,
the analysis of such indentations is never well defined at a phase
interface, at which the strains and stresses are distributed nonuniformly,
so image areas comprising many interfacial boundaries will be difficult
to quantify accurately. Finally, viscoelasticity is typically neglected in
such mapping, as the loading times sufficient to suppress viscous
responses should differ from pixel to pixel but are maintained constant