AFM Lab Report
AFM Lab Report
AFM Lab Report
Introduction
Atomic Force Microscopy is a powerful tool used to identify the surface structure of a solid by contouring the top layer with a sharp tipped probe and
amplifying the hills and troughs via laser reflection and detection. Piezoelectric materials are used to finely tune the x and y distance parameters, and a
harmonic oscillator is utilized in close contact mode to move the cantilever
in the z direction. A Scanning Probe Microscopy program processes the data
and calculates the roughness of the sample.
Background
2.1
Probe Mechanics
d1 = KV
(1)
where d1 and d2 are the initial and final lengths respectively, V is the applied
voltage, and K is the piezo constant. [1] K must have a smaller value that
allows for the change in voltage to induce a very small change in distance.
Each dimension (x, y, and z) are controlled by a dierent piezoelectric. Refer
to Figures 3 and 4.
Cantilevers are attached to the electronics and hold the probe in place.
They can be triangular or rectangular in shape and are primarily made of
silicon or silicon nitride. [1] The attached probe can be pyramidal or conical
in shape and a sharp tip is desired for the most accurate results. To obtain the
desired radius of curvature around 2 nanometers, the material is chemically
etched, or ion milled, usually leaving the tip only a couple of molecules thick.
Refer to Figure 5.
To keep the tip close to the sample, a force sensor is used to monitor the
potential energy as a function of distance between the tip and the sample.
This follows the potential energy diagram in Figure 6.
2.2
Contact Mode
In contact mode, the tip is slid across the surface of the sample and a feedback
loop maintains a constant distance between the cantilever and the sample by
vertically moving the scanner to keep the force the same. Refer to Figure 7.
The computer calculates this force following:
F = kz
2
(2)
2.3
In close contact mode, a harmonic oscillator attached to the cantilever oscillates the probe at an amplitude of 20 to 100 nanometers. The tip of the probe
taps the surface at the bottom of the wave and a feedback loop maintains a
constant amplitude. [1] Refer to Figure 8.
This mode is advantageous because the tip gets closer to the sample,
resulting in higher lateral resolution. The tip is harder to break because
the force is smaller and does less damage to soft samples. Lateral forces are
eliminated due to the oscillation of the probe. The only drawback is that
each scan takes longer than contact mode.
Operation
3.1
Probe Exchange
To replace or substitute probes, one can release the scanner and slide it out
for easier access. A magnet connects the cantilever to the scanner and keeps
the probe in place. Using tweezers, carefully take out the cantilever and
monitor which side contains the probe as it cannot be seen with human eyes.
Put the new tip in and check that the magnet has captured the cantilever.
Replace the scanner and begin loading the sample.
3.2
Computer Operation
Load the sample and turn on the computer. The electronics in the AFM are
controlled by the computer, which operates the following components: [1]
Coarse Z Motion Translator: used for moving the probe in the z direction for long distances
Coarse X-Y Translation Stage: used for moving the sample in the x-y
direction for long distances
X and Y Piezoelectric Transducer: used for moving the probe in the
x-y direction for small distances (on the scale of nanometers)
Force Sensor: detects the potential force between the tip and the surface
Z Piezoelectric Ceramic: used for moving the probe in the z direction
over small distances
Feedback Control Unit: connects the force sensor to the computer and
signals the piezoelectric materials to change so as to keep a constant
distance between the probe and the surface so as to not break the tip
X-Y Signal Generator: generates the voltage to be applied to the piezoelectric materials
After the data is obtained and the signal is processed, the computer displays
the surface height as a function of x-y position using the contrast of color.
Refer to Figure 3.
One must align the tip and load the correct software for the current tip.
Approach the tip to the surface of the sample using the coarse and piezoelectric z motion transducers. Set the frequency, amplitude, and position if
using close contact mode. Begin scanning by setting an area to scan and
a time interval for each x and y component. After scanning, the computer
compiles the data and displays a square spectra representing the surface of
the sample, which can be converted to a 3D image.
Data Analysis
Three samples were scanned via Atomic Force Microscopy using contact mode
and close contact mode. Various characteristics of the sample were calculated
4
including the mean of the height, the root mean square of the height, and
the roughness of the sample.
The mean height is defined as the sum of all the measured heights divided
by the total number of measurements (N), and can be written in summation
or integral form following:
N
X
h = 1
hi =
N i=1
(b
1
a)(d
c)
Z bZ
a
h(x, y) dy dx
(3)
hy =
1
(d
c)
h(y) dy
(5)
h,x
1
(b
a)
and,
h,y
1
(d
c)
s
Z
(h(x)
hx )2 dx
(7)
hy )2 dy
(8)
s
Z
(h(y)
These calculations are displayed in the data tables in the Figures of each
spectra.
5
A Silicon coated glass sample was scanned using contact and close contact
tips. Refer to Figures 9 and 10. The dots seen on the sample are the Silicon
bits on the glass. The same area of 55.55 m2 was scanned, using the same
scanning speed. Contact mode calculated a roughness of 38.19 mv while
close contact mode calculated 6.59 nm, showing the higher precision of close
contact mode in comparison to contact mode. The bits of Silicon caused this
sample to have a high standard deviation of height due to the probe being
raised over the lumps. The horizontal dark lines visible across the spectra on
Figure 10 were due to students bumping the table on which the AFM was
placed, causing the probe to jump and miscalculate the height.
A Silicon wafer was scanned using contact and close contact tips. Refer
to Figures 11, 12, and 13. This sample is a lot smoother than the silicon
coated glass sample, as the roughness calculated using contact mode and close
contact mode were 13.66 mv and 71.47 mv respectively. As seen from the
spectrum of Figure 11, the sample is very smooth with only a few aberrations.
In Figure 12, a smaller portion of the total area was selected (47.72 m2 of
the total 888.79 m2 ), and the roughness increased to 128.09 mv due to
the smaller area having greater variations in height comparably, whereas the
larger area contained a great portion of it that was relatively flat. The flat
area can be seen on the left hand side of the 3D spectra of Figure 13, where
on the right side it dips and creates a larger roughness.
A thermally evaporated Aluminum sample was scanned using a close
contact tip and the height of the sample was analyzed using the line analysis
tool of the program. Refer to Figures 14 and 15. An area of 2.22 m2 was
scanned and found to have a roughness of 20.67 nanometers. This very high
roughness is due to the very tall lump in the top right corner of the spectrum,
causing the standard deviation to increase greatly. Using line analysis, one
may compare a roughness that doesnt include the bump to one that does.
Figure 15 shows Line 1 having a roughness of 10.80 nm (not including the
bump) and Line 2 having a roughness of 60.11 nm (including the bump).The
bump on Line 2 is shown on the right side of the 2D height graph in Figure 15.
The data from Line 1 and Line 2 were fit to a summation of functions using
Mathematica and the resulting equations for height were used to calculate
the mean height and roughness of the samples using equations 4,5 and 7,8
displayed on the last page of the report. The program calculated a roughness
average of 10.80 nm for Line 1, and using Mathematica, a value of 14.789 nm
was calculated, corresponding to the red highlighted result on the page after
the figures. This approximation is higher than the actual due to the fitted
6
line not hitting every point on the spectra, leading the area under the curve
to be greater. The program calculated a roughness average of 60.11 nm for
Line 2, and using Mathematica, a value of 49.366 nanometers was calculated.
This result is smaller than the actual due to my fitting equation not being
able to model the quick bump around y = 1.25m, leading to a smaller area
under the curve, and thus a lower roughness.
Conclusion
References
[1] Shi, Hongtao. PHYS 366:
Intermediate Experimental Physics,
Atomic
Force
Microscopy
2014.
url:
http://www.physastro.sonoma.edu/people/faculty/shi/p366/lectures/afm.pdf
10
11
12
13
14