Ultrasonics in an Atomic Force Microscope
Mark S. Skilbeck, Rachel S. Edwards, Neil R. Wilson
Email: m.skilbeck@warwick.ac.uk
Department of Physics, University of Warwick, Coventry, CV4 7AL, UK
Atomic Force Microscopy (AFM)
Four-quadrant
photodiode
Laser
Cantilever
Transducer (UFM)
Ultrasonic Force Microscopy (UFM)
Lateral
Deflection
Drive
• Sample is mounted on a piezoelectric transducer.
• Sample is oscillated at high frequencies (> 1 MHz).
• Cantilever cannot respond to oscillations at these
frequencies, instead the tip indents and retracts from the
sample.
• Due to nonlinearity of force vs indentation relation,
average force on the cantilever increases with increasing
oscillation amplitude, causing the cantilever to deflect.
• Increase also depends on surface stiffness.
• Oscillation is amplitude modulated at lower (kHz)
frequencies, to distinguish between topography and
surface stiffness information.
• Oscillation also induces superlubricity between tip and
sample, allowing scanning of delicate samples.
4 μm
10-5
10-6
Above: Raw signals in UFM experiments.
Below: Gold on silicon sample, topography (left, 300nm full scale)
and UFM response (right), showing a lower response (darker) for
gold, meaning it is softer.
10-7
10-8
8 μm
10-9
Current (A)
Sample
Real
Current
• Nanoscale tip (typical radius < 10 nm) on the end of a
flexible cantilever.
• Contact with sample surface, measure deflection of
cantilever (typically using a reflected laser and four
quadrant photo-diode).
• Sample can be scanned and changes in deflection
correspond to surface topography.
• Very high resolution in-plane (typically 10s nm) and
out-of-plane (typically 100 pm).
• Physical contact allows for investigation of other
properties, such as mechanical properties, and
conductivity.
• Cantilever can also twist, allowing detection of lateral
motion and forces.
Idealised
10-10
Sample stage
(X-Y Motion)
10-11
Conductivity image of carbon nanotubes, using superlubricity
to prevent sample damage. Sample is biased to 2V using a gold
contact off the right side of the image.
Non-Destructive Testing (NDT)
Detection of Ultrasound
• At lower (< MHz) frequencies, cantilever will follow
motion of the surface.
• Can be used as a high sensitivity, high resolution
pickup for ultrasound experiments, using waves with
out of plane motion.
• In plane motion also detectable using the lateral
channel, though only in one direction.
• Measure the deflection of the cantilever when in
contact with the sample, with ultrasound generation
by an electro-magnetic acoustic transducer (EMAT)
driven with a high voltage pulse.
• On an aluminium plate, fast S0 mode and slower,
dispersed A0 mode both visible, with minimal
generation noise.
EMAT
0.5 mm aluminium sheet
Wave Direction
~150 mm EMAT to crack
Through crack ~20 μm wide
Scan line
10 mm
150 mm
20 mm
80 mm
150 mm
Above: Layout of sample for testing ultrasonic pickup in an AFM. The
EMAT generates out of plane motion, resulting in Lamb waves on
plate samples. The tip is in contact on the scan line, near the crack.
Below: Typical A-Scan (cantilever deflection) for tip contact on the
side of the crack closest to the EMAT.
We acknowledge support from the University of
Warwick through a Chancellor’s scholarship to MSS and
support from the EPSRC through grant EP/J015202/1.
We also thank Oxford Instruments Asylum Research,
Inc. for loan of a dual gain ORCA.
Further Reading
1.
2.
3.
4.
Skilbeck, M. S. et al., Nanotechnology, 25, 335708 (2014).
Yamanaka, K., Ogiso, H., & Kolosov, O., Applied Physics Letters, 64, 178 (1994).
Dinelli, F., Castell, M., & Ritchie, D., Philosophical Magazine A, 80, 2299 (2000).
Clough, A. R., & Edwards, R. S., Journal of Applied Physics, 111, 104906 (2012).
• Cracks and other defects cause changes in the ultrasound
signal; looking at these changes allows the cracks to be
detected.
• Here we measured multiple A-Scans across a crack, as
pictured in “Detection of Ultrasound,” with 20 μm steps
between points, compiled into a B-Scan (left).
Above: B-Scan across a crack using an AFM as a pickup. The crack is
located at ~3 mm and < 3 mm positions are on the EMAT side of the
crack. 50nm full data scale.
Below: Sonograms for tip position at 0 mm (top) and 2.6 mm (just
before crack, bottom). Dispersion curves for S0 and A0 Lamb waves with
a 150 mm travel distance are overlaid.
»» Crack is non parallel to sides of the sample to reduce
effects from edge reflections.
»» Crack goes all the way through the sample to maximise
the effects.
»» Clear change is visible when the crack is crossed,
including a reduction in amplitude and a time delay.
• Sonograms are time windowed Fourier transforms of the
signal, showing frequency content at each time.
»» Dispersion curves can be overlaid to see the mode
content.
»» Shown here (left) is that the Lamb wave A0 mode is
strong in this experiment, and a weaker S0 mode is also
visible (which is expected as the A0 mode has more outof-plane content than the S0 mode).
• Regions on a sonogram can be averaged to see how
the amplitude within that frequency and time window
changes with position.
»» As seen below, there is enhancement of the signals
before the crack (due to constructive interference with
reflections), followed by a near complete drop off after.
Mean values for regions indicated by boxes on the sonograms. Black is the
S0 mode, red is the A0 mode and yellow is not a direct Lamb wave mode.
ULTRASOUND GROUP
D E P A R T M E N T
O F
P H Y S I C S
warwick
microscopy