-
1
Effect of Tip Radius on AFM Scans
Guang Li1, K. R. Gadelrab
1, Tewfik Souier
1, Pavel L. Potapov
2, Gang Chen
3 and Matteo Chiesa
1,3*
1 Laboratory for Energy and Nano-Science, Masdar Institute of
Science and Technology, Abu Dhabi, United Arab Emirates 2Center of
Complex Material Analysis, GLOBALFOUNDRIES, Dresden, Germany
3Department of Mechanical Engineering, Masschusetts Institute of
Technology, Cambridge, MA 02139, United States
*Corresponding author: [email protected]
In this study, we try to simulate the effect of Atomic Force
Microscope (AFM) tip radius on the captured details of AFM
scans. While such study does not provide a quantification to the
errors resulted from using a larger AFM tip radius, it
helps to visualize the influence on the captured details of the
scanned surfaces. Furthermore, the extracted results are valid
for the order of magnitude and tip radii commonly utilized in
day to day AFM experiments. The study is performed on
real AFM scans with almost no assumptions related to the tip or
the surface.
To achieve this, an AFM image is extracted in the form of a
matrix that has rows and columns equal to the resolution of
the image. The entries of the matrix are the height values of
the topography recorded by the AFM. A scanning process to
the extracted surface is simulated numerically through a MATLAB
script. The AFM tip is simulated as a virtual sphere
that sweeps the surface by changing the coordinates of its
center while the z value of the region covered by the sphere is
recorded. No mutual penetration condition is fulfilled at every
point in the covered zone. To achieve this, the difference in
height between the sphere surface and the sample topography is
calculated. The recorded z value in that region is the one
that produces no negative z difference (the topography does not
penetrate through the sphere surface). This process is
performed every instance the sphere center is shifted by a step
of one pixel.
Figure 1 shows an AFM scan of indentation using a Berkovich tip
in fused silica. The 3D scan is viewed along the y-z
plane to be able to compare it with generated surfaces from the
simulation. The scan is obtained in tapping mode on MFP
3D Asylum Research with a scan resolution of 512×512.
Figure 1. AFM scan of a fused silica indentation viewed along
the y-z plane to be compared with surfaces generated by
the simulation. The figure shows the remaining indent and some
structure on the surface probably due to some surface
contamination and pile-up.
The simulation is run with different tip radii that have the
values 10nm, 50nm, 100nm and 200nm. In fact, the common
AFM tips commercially available have tip radii in the vicinity
of 10nm. During service, the tip gets slightly blunted till it
stabilizes in the range of 20 to 30nm. Furthermore, the range of
tip radii in this study is extended to 200nm that is almost
mailto:[email protected]
-
2
never encountered in AFM tips; however, this case may occur when
the indenter tip, itself, is employed to get a scan of
the deformed surface after performing an indentation
experiment.
Figure 2 shows the simulation results for different tip radii.
It is seen that the 10nm tip simulation is able to regenerate
the
original surface almost perfectly. Moving to higher values of
the tip radius produces surfaces that are dilated laterally
which affects the lateral dimensions of the surface features. In
addition the expected loss of small details is observed at
larger value of the tip radius generating smoother surfaces
In the axial direction, there is almost no effect of the tip
radius is seen on the height of concave features. On the other
hand, there is an observed difference between the measured
indentation depth with a sharp and a blunted tip. However,
even at a tip radius of 200nm, the error in the measured
indentation depth is about 6.5%. This can be explained by the
wide angle of the remaining indent due to the angle of the
Berkovich tip. Using a smaller angle indenter will result in a
larger error for the same tip radius.
Figure 2. Numerically generated surfaces with different tip
radii. The 10nm tip radius is able to regenerate the surface
accurately. Having a larger tip radius produces a lateral
expansion in the surface features. The depth of the remaining
indent
has about 6.5% error at a tip radius of 200nm.
R=10nm R=50nm
R=100nm R=200nm
-
3
To realize a better understanding of the effect of tip blunting
on the obtained AFM scan, a section is cut across the surface
plot to reveal a 2D view to the indentation depth. Figure 3
demonstrates the section cut for different tip radii. Again,
the
10nm generated plot cannot be distinguished from the original
one. The largest tip radius generates a shallower
indentation depth that is slightly narrower. The error in the
obtained depth is about 8%. From figures 2 and 3, it is seen
that even with a tip radius of 50nm (severely blunted AFM tip),
the error in the scanned featured is low.
Figure 3. A section cut through the numerically simulated
surfaces in comparison to the original AFM scan. The 10nm tip
radius regenerates the surface accurately, while the 200nm tip
radius produces shallower indentation depth and slightly
narrower remaining indent.
The same simulation is performed on a scanned Berkovich tip
utilized to obtain the previous indentation. Figure 4
illustrates the surface plot generated by the AFM scan viewed
along the x-z plane. The indenter tip is quite sharp and the
AFM scan reveals the lapping marks that are clearly observed
along the pyramid edge. A feature is also noticed on the
pyramid surface that is probably a lapping remaining.
Figure 4. An AFM scan of a Berkovich tip utilized to generate
the previously studied indent. The scan accurately captures
the details of the tip and the lapping marks.
-
4
Running the same simulation previously described with the same
range of tip radii, generates the plots in Figure 5. The
most pronounced effect of increasing the tip radius is the clear
rounding of the indenter tip radius and indenter tip edges.
In addition, the fine lapping features on the surface of the
pyramid are smoothened out.
A section cut through the surface plot again shows that 10nm AFM
tip radius is able to regenerate the scanned surface
properly while increasing the AFM tip radius resulted in a clear
blunting in the indenter tip shape. Again, the 50nm tip
AFM tip radius shows minimal effect of the scanned features of
the indenter tip.
The artificial rounding of the indenter tip apex and edges will
result in an error in the calculated tip area function.
However, it is seen that fresh and sharp AFM tips can generates
the indenter tip topography accurately.
Figure 5. Numerically generated surfaces for a Berkovich tip
with different AFM tip radii. The 10nm tip radius accurately
generated the surface while the 200nm tip radius produced a
rounded apex and rounded edges. Such rounding will
introduce errors in the calculated area function of the tip.
R=10nm R=50nm
R=100nm R=200nm
-
5
Figure 6. Comparison between different sections in the Berkovich
tip generated by different AFM tip radii. The section
generated by the 10nm AFM tip can’t be distinguished from the
original scan. The 200nm tip results in a blunted apex that
is significantly different from the true surface.
Conclusion
Tip radius in AFM affects the accuracy of the scanned images.
Sharp tips that have tip radii in the range of 10nm can
accurately generate the topography of the scanned surface.
Increasing the tip radius results in a lateral dilation of
small
features (comparable to the tip radius) and introduces
artificial rounding to sharp regions (Pyramid apex and edges).
Blunt
tips that have radii in the range of 20-30nm introduce minimal
errors that does not significantly influence the topography.
Using blunted tips with tip radii in the range of 200nm will
significantly dilate the scanned topography and result in about
8% error in the estimated depth of deformed surfaces by
Berkovich indenters.