Retrofitting an atomic force microscope with photothermal excitation for a clean cantilever response in low Q environments

Abstract

Dual resonance excitation system for the contact mode of atomic force microscopy Rev. Sci. Instrum. 83, 043703 (2012) Dual frequency open-loop electric potential microscopy for local potential measurements in electrolyte solution with high ionic strength Rev. Sci. Instrum. 83, 033709 (2012) Note: A scanning electron microscope sample holder for bidirectional characterization of atomic force microscope probe tips Rev. Sci. Instrum. 83, 036108 (2012) Analysis of the contrast mechanism in bimodal atomic force microscopy combining amplitude modulation and band excitation J. Appl It is well known that the low-Q regime in dynamic atomic force microscopy is afflicted by instru-mental artifacts (known as "the forest of peaks") caused by piezoacoustic excitation of the cantilever. In this article, we unveil additional issues associated with piezoacoustic excitation that become ap-parent and problematic at low Q values. We present the design of a photothermal excitation system that resolves these issues, and demonstrate its performance on force spectroscopy at the interface of gold and an ionic liquid with an overdamped cantilever (Q < 0.5). Finally, challenges in the interpre-tation of low-Q dynamic AFM measurements are discussed. © 2012 American Institute of Physics. [http://dx.doi.org/10.1063/1.4712286]

Full text

Page 1

Retrofitting an atomic force microscope with photothermal excitation for aRetrofitting an atomic force microscope with photothermal excitation for a
clean cantilever response in low Q environmentsclean cantilever response in low Q environments
Aleksander Labuda, Kei Kobayashi, Yoichi Miyahara, and Peter Grütter

Citation: Rev. Sci. Instrum. 8383, 053703 (2012); doi: 10.1063/1.4712286
View online: http://dx.doi.org/10.1063/1.4712286
View Table of Contents: http://rsi.aip.org/resource/1/RSINAK/v83/i5
Published by the American Institute of Physics.

Related ArticlesRelated Articles
Imaging of subatomic electron cloud interactions: Effect of higher harmonics processing in noncontact atomic
force microscopy
Appl. Phys. Lett. 100, 163104 (2012)
Dual resonance excitation system for the contact mode of atomic force microscopy
Rev. Sci. Instrum. 83, 043703 (2012)
Dual frequency open-loop electric potential microscopy for local potential measurements in electrolyte solution
with high ionic strength
Rev. Sci. Instrum. 83, 033709 (2012)
Note: A scanning electron microscope sample holder for bidirectional characterization of atomic force microscope
probe tips
Rev. Sci. Instrum. 83, 036108 (2012)
Analysis of the contrast mechanism in bimodal atomic force microscopy combining amplitude modulation and
band excitation
J. Appl. Phys. 111, 054909 (2012)

Additional information on Rev. Sci. Instrum.Additional information on Rev. Sci. Instrum.
Journal Homepage: http://rsi.aip.org
Journal Information: http://rsi.aip.org/about/about_the_journal
Top downloads: http://rsi.aip.org/features/most_downloaded
Information for Authors: http://rsi.aip.org/authors
Downloaded 10 May 2012 to 132.206.186.16. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

Page 2

REVIEW OF SCIENTIFIC INSTRUMENTS 83, 053703 (2012)
Retrofitting an atomic force microscope with photothermal excitation
for a clean cantilever response in low Q environments
Aleksander Labuda,1Kei Kobayashi,2Yoichi Miyahara,1and Peter Grütter1
1Department of Physics, McGill University, Montreal, Quebec H3A 2T8, Canada
2Office of Society-Academia Collaboration for Innovation, Kyoto University, Katsura, Nishikyo,
Kyoto 615-8520, Japan
(Received 27 February 2012; accepted 22 April 2012; published online 10 May 2012)
It is well known that the low-Q regime in dynamic atomic force microscopy is afflicted by instru-
mental artifacts (known as “the forest of peaks”) caused by piezoacoustic excitation of the cantilever.
In this article, we unveil additional issues associated with piezoacoustic excitation that become ap-
parent and problematic at low Q values. We present the design of a photothermal excitation system
that resolves these issues, and demonstrate its performance on force spectroscopy at the interface of
gold and an ionic liquid with an overdamped cantilever (Q < 0.5). Finally, challenges in the interpre-
tation of low-Q dynamic AFM measurements are discussed. © 2012 American Institute of Physics.
[http://dx.doi.org/10.1063/1.4712286]
I. INTRODUCTION
Dynamic atomic force microscopy1(dAFM) has become
a ubiquitous tool in surface science in the last two decades,
and its versatility is made evident by the wide span of stud-
ies in environments ranging from ultrahigh vacuum, to cryo-
genic temperatures, and to liquids. It has recently been iden-
tified that piezoacoustic excitation of cantilevers can prevent
accurate interpretation of data in all of these environments,2–4
when using frequency modulation (FM) AFM.5This article
extends the discussion to low-Q environments (near and be-
low Q values of 1), where amplitude modulation (AM) AFM
is the preferred dAFM method.
Section II discusses problems with piezoacoustic exci-
tation that are unique to low-Q environments – in contrast
to the well-behaved photothermal excitation.6Afterwards,
the design of our photothermal excitation unit – dubbed
“the photothermal panther” – is presented and its perfor-
mance is benchmarked in water. Then, photothermal ex-
citation is employed for imaging and force spectroscopy
on Au(111) in 1-butyl-3-methylimidazolium hexafluorophos-
phate [BMIM][PF6] – a highly viscous ionic liquid. The pho-
tothermal dAFM measurement is compared to simultaneously
acquired static AFM (sAFM) data, as both provide indepen-
dent measurements of the solvation stiffness profile. Finally,
limitations and complications that pertain to dAFM in low-Q
environments are discussed.
II. PROBLEMS WITH PIEZOACOUSTIC EXCITATION
The benefits of photothermal excitation over piezoacous-
tic excitation in FM-AFM have recently been outlined in great
detail.4The “forest of peaks”7observed using piezoacoustic
excitation severely complicates the interpretation of the mea-
sured signals. In fact, accurate recovery of the conservative
and dissipative forces may become impossible if the tempera-
tureoftheinstrumentchangesbyaslittleasafewmKbecause
the “forest of peaks” is temperature-dependent4– it can drift
and distort throughout the experiment.
Driving the cantilever at a fixed frequency greatly sim-
plifies the interpretation of data as the frequency-dependence
of the piezoacoustic excitation can be disregarded. This fact
favors the use of AM-AFM for piezoacoustic excitation in
liquids. However, the ratio of the cantilever excitation that
arises from fluid vibrations (fluid-borne excitation), as op-
posed to movement of the cantilever base (structure-borne
excitation),8is unknown a priori; knowledge of this ratio
is necessary for extracting quantitative force and dissipation
measurements.9
Nevertheless, fundamental problems with piezoacoustic
excitation persist in low-Q environments when used in con-
junction with the optical beam deflection method:10the mea-
sured oscillation amplitude of the cantilever and the true os-
cillation of the tip do not relate in any simple manner, as is
graphically represented in Figure 1.
To illustrate this problem, we have used recent modelling
of cantilever dynamics in liquids9to plot the transfer func-
tion of the cantilever in the limiting case of structure-borne
FIG. 1. (a) Piezoacoustic excitation drives the cantilever by moving the base
of the cantilever, leading to complex cantilever dynamics that are frequency
dependent: the measured angle of the cantilever tip does not relate to os-
cillation amplitude in any simple way for Q factors around and below 1.
(b) Photothermal excitation drives the cantilever via stress-induced cantilever
bending: the shape of the cantilever is well defined. (Artistic rendition of
cantilevers and gold background courtesy of Magdalena Wielopolski).
0034-6748/2012/83(5)/053703/8/$30.00© 2012 American Institute of Physics
83, 053703-1
Downloaded 10 May 2012 to 132.206.186.16. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

Page 3

053703-2Labuda et al.Rev. Sci. Instrum. 83, 053703 (2012)
loading only. We restrict the following discussion to the first
eigenmode, as it suffices to illustrate the primary problem at
hand. Figure 2 demonstrates that the measured oscillation am-
plitude strongly deviates from the true amplitude if it is driven
off-resonance. In other words, the deflection sensitivity (in
unitsofnm/V)isonlyvalidaroundtheresonance.Conversely,
photothermal excitation has a constant sensitivity throughout
the usable spectrum because the bending shape is frequency-
independent.11Figure 2(b) suggests that a measurement of os-
cillation amplitude is reliable throughout the spectrum when
using photothermal excitation, as opposed to piezoacoustic
excitation.
This calibration predicament becomes especially prob-
lematic in overdamped environments (Q < 0.5), where the
cantilever is excited below resonance. In fact, it would per-
haps be foolish to excite the cantilever at its natural frequency,
because the second eigenmode may have a larger response at
that frequency. In any case, for piezoacoustic excitation, the
dAFM sensitivity becomes a function of the drive frequency,
the natural frequency, the Q factor, and the ratio of fluid-borne
to structure-borne excitation. Calibrating the AFM sensitivity
becomes difficult, if not impossible.
The above-mentioned problems do not only affect
interpretation of acquired data, they can also prevent regular
topographical imaging. Depending on the ratio between fluid-
borne and structure-borne excitation, the measured amplitude
may increase or decrease as the tip approaches the surface,
potentially prohibiting stable feedback required for imaging.
For example, if structure-borne excitation dominates, the
base of the cantilever moves with an amplitude larger than
the cantilever tip, as illustrated in Figure 1(a). In that case, the
measured angular deflection of the cantilever becomes larger
(a)
(b)
FIG. 2. (a) The first eigenmode of a driven cantilever with Q = 2 is simu-
lated for photothermal excitation (with fixed force amplitude) and for piezoa-
coustic excitation (with fixed base amplitude) around the natural frequency
(10 kHz). The photothermal response is simply that of a harmonic oscillator,
i.e., the transfer function of the cantilever. As illustrated in Figure 1(b), the
amplitude measured by piezoacoustic excitation does not correspond to the
actual oscillation amplitude, except near the resonance frequency. (b) In other
words, there is an amplitude calibration error when driving the cantilever off-
resonance, which becomes inevitable for very low Q environments.
as the cantilever approaches the surface, rendering AM-AFM
imaging difficult, if not impossible.
All the above-mentioned problems are resolved by em-
ploying photothermal excitation instead of piezoacoustic
excitation.
III. THE PHOTOTHERMAL PANTHER
The photothermal panther is a retrofit to our exist-
ing home-built electrochemical (EC) AFM12designed for
atomic-scale friction force microscopy in electrochemical
environments.13The only hardware modification to the pre-
vious design was the addition of a magnetic docking stage
that allows the panther to be mounted effortlessly and rigidly
with ∼μm repeatability.
A. Optomechanical design
The design of the photothermal panther and its integra-
tiontotheexistingECAFMisbestexplainedgraphically.This
section describes some additional details, with reference to
Figure 3.
A 20 mW blue laser-diode (DL-LS5042, Sanyo) light
beam is collimated to ∼1 mm by an aspheric collimation lens
(352610-A, Thorlabs) with 4 mm effective focal length. The
collimated light beam is reflected off a polarizing beamsplit-
ter towards a 567 nm longpass dichroic mirror (DMLP567R,
Thorlabs) that combines the red detection light beam and the
blue excitation light beam. The existing optics of the ECAFM
focus both light beams onto the cantilever, as well as rotate
their polarization directions by 90◦. The rotation in polariza-
tionguidestheredlighttowardsthephotodetectorafterreflec-
tion, as well as prevents blue light from returning into the blue
laser-diode which could cause lasing instabilities. Instead, the
blue light is mostly absorbed by a yellow filter, which also
acts as a sight glass for aiding the alignment of both light
beams onto the cantilever. Two micrometer thumbscrews con-
trol the rotation angle of the laser-diode and the beamsplitter,
thereby enabling 2-axis positioning of the blue laser onto the
cantilever.
Additionally, since the previous description of our home-
built ECAFM,12the collimated diameter of the detection light
beam was reduced to 0.65 mm, and the effective focal length
was increased to 25 mm. These changes reduced the diver-
gence of the light beam by 4.5×, and therefore reduced the
detection noise by the same ratio.14
The fully assembled ECAFM 3D technical drawing
and the internal components can be interactively viewed in
Figure 4.
B. Electronics
The blue laser-diode is driven in constant-power mode
(WLD3343, Wavelength Technology) by using the in-
tegrated photodetector of the laser-diode for feedback.
The drive current is modulated at 300 MHz using a
voltage-controlled oscillator (POS-535+, Mini-Circuits) to
Downloaded 10 May 2012 to 132.206.186.16. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

Page 4

053703-3Labuda et al.Rev. Sci. Instrum. 83, 053703 (2012)
FIG. 3. The photothermal panther integrates a blue light beam into the existing AFM optical system to enable photothermal excitation of the cantilever.
Downloaded 10 May 2012 to 132.206.186.16. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

Page 5

053703-4Labuda et al.Rev. Sci. Instrum. 83, 053703 (2012)
FIG. 4. Interactive 3D technical drawings of (a) the assembled ECAFM and (b) the internal optics and mechanics. Click image to activate. Use ctrl, shift, to
zoom and pan.
reduce optical feedback and interference as described in
Ref. 15.
Most importantly, the blue light power is modulated for
photothermal excitation of the cantilever using a power split-
ter/combiner (PRSC-2050, Mini-Circuits). An input AC drive
signal modulates the laser power output with a modulation
depth determined by the drive amplitude. The circuit allows
modulation of the laser power for frequencies between 1 kHz
and 3 MHz, which covers the range of typical cantilever drive
frequencies.
Also, our previous controller was replaced by the Nano-
nis SPM controller, allowing reliable imaging, the acquisition
of transfer functions, and enabling fast force spectroscopy.
IV. PROOF-OF-PRINCIPLE IN MODERATE Q
We first demonstrate the capabilities of the photothermal
panther in a moderate Q environment – water – where a clear
cantilever resonance peak is observable. A gold-coated can-
tilever was used because it has higher photothermal efficiency
than an uncoated one.16
A photothermally driven transfer function of the can-
tilever is shown in Figure 5(a). Overlaid is a fit to a har-
monic oscillator model with a multiplicative 1/f background;
the model fits the data impeccably. The 1/f background (with
anexponent of0.36setasafreefittingparameter)isattributed
to the frequency-dependence of the photothermal driving ef-
ficiency, i.e., the photothermal excitation transfer function.4
The close agreement to the harmonic oscillator model is
reassuring, as any AM-AFM theory1,17,18used to interpret the
magnitude and phase response throughout the experiment is
based on the assumption that the cantilever can be described
as a harmonic oscillator.
In Figure 5(b), the linearity of the system is demonstrated
by driving the cantilever on-resonance at high power (15 mW
at roughly 50% modulation depth) and measuring its deflec-
tion. A sinusoidal fit suggests highly linear behaviour; the
residual plot deviates by no more than 0.3% of the oscillation
amplitude (not shown).
Finally, the imaging capability of the system is illustrated
in Figure 5(c) by the constant-amplitude topography image of
Au(111) steps acquired in water using AM-AFM.
V. IMAGING IN OVERDAMPED ENVIRONMENTS
A gold sample was imaged in [BMIM][PF6] using AM-
AFM. Figure 6 presents a constant-amplitude topography im-
age of a Au(111)-oriented grain, along with the phase signal
Downloaded 10 May 2012 to 132.206.186.16. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

Page 6

053703-5Labuda et al. Rev. Sci. Instrum. 83, 053703 (2012)
FIG. 5. (a) The photothermally driven cantilever amplitude response in wa-
ter far from the surface. Cantilever stiffness: 1.34 N/m and Q: 3.6. A har-
monic oscillator function with 1/f background is fit and overlaid. (b) Driving
the cantilever on-resonance with a 8mW modulation amplitude demonstrates
the linearity of the system up to large oscillation amplitudes. A sinusoidal
function was fit and overlaid. (c) Constant-amplitude topography image of
Au(111) in water using photothermal AM-AFM. (Scan width: 500 nm; scan
speed: 3 lines/s; pixel acquisition rate: 1 kHz).
and amplitude error signal. The subtle variations in topog-
raphy loosely correlate with the unidentified objects clearly
seen in the phase image. It was impossible to image the under-
lying Au(111) steps using dAFM, despite the high cantilever
stiffness (28.5 N/m). In fact, this cantilever was too soft to
penetrate the last layer of ionic liquid, necessary to observe
the gold surface. The root of this problem will be discussed
more thoroughly in Sec. VII.
Due to the low stiffness of the cantilever, it was neces-
sary to switch imaging methods to sAFM in order to pene-
trate last layer and reveal the underlying Au(111) steps, as
shown in Figure 6. Interestingly, there is a fine line between
imaging and damaging the underlying gold substrate: a 40 nN
set point was necessary to image, while 150 nN caused sig-
nificant damage to the Au(111) surface, as observed by the
damaged square area in the sAFM image in Figure 6.
Interestingly, switching between dAFM and sAFM with
a single cantilever extends the dynamic range of usable force
set points, allowing the imaging of both the solid surface and
its solvation layers.
VI. FORCE SPECTROSCOPY IN OVERDAMPED
ENVIRONMENTS
After imaging the surface, force spectroscopy was per-
formed in the center of the Au(111) grain seen in Figure 6.
A. Calibration
The dynamic cantilever stiffness had been determined
prior to imaging from a power spectral density (PSD) in air,
using the Sader method19,20: kc= 28.5N/m. Immediately be-
fore the force spectroscopy experiment, the deflection sensi-
tivity of the AFM was calibrated ∼1μm from the surface by
measuring a PSD of the cantilever deflection in the ionic liq-
uid, as seen in Figure 7.
A PSD was measured with the blue laser-diode both
turned on and off, in order to verify that the blue light was
not affecting the cantilever deflection signal in the frequency
range of interest.21The blue light causes additional 1/f can-
tilever bending below 2 kHz because of the coupling to light
power fluctuations caused by the gold coating.22
These verifications confirm that the measured PSD is in
fact a thermal spectrum of the cantilever above 2 kHz. Be-
cause the cantilever is overdamped, its first eigenmode can
be modelled as a massless harmonic oscillator, using the for-
mulism proposed in a recent communication.23In this case,
the cantilever spectrum is accurately modelled as a first-order
low-pass filter driven by a white thermal force spectrum that
FIG. 6. Dynamic (AM) AFM was used to image Au(111) in [BMIM][PF6]. Gold steps could not be observed as the range of imaging setpoints only allows the
imaging of solvation layer above the surface. Switching to static AFM enables the imaging of underlying Au(111) with a 40 nN setpoint; however, a 150 nN
setpoint damages the surface which can observed as a square in the image from the previous scan. (Image size: 300 nm × 300 nm; scan speed: 5lines/s; pixel
acquisition rate: 5 kHz).
Downloaded 10 May 2012 to 132.206.186.16. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

Page 7

053703-6Labuda et al.Rev. Sci. Instrum. 83, 053703 (2012)
FIG. 7. Thermal spectrum of the cantilever in air, used to calibrate the stiff-
ness, and the thermal spectrum in [BMIM][PF6] near the gold surface, used
to calibrate the damping. Turning on the blue light confirms that the measured
spectrum around 10 kHz is in fact a thermal spectrum, which therefore can
be used to calibrate the AFM using the fluctuation-dissipation theorem and
the measured damping.
can bedetermined bythefluctuation-dissipation theorem.24In
summary, the thermal PSD for overdamped cantilevers equals
4kBT
kc(2πfro)
1
1 + ( f/fro)2,
(1)
where frois the roll-off frequency of the low-pass filter, and
kBT is the thermal energy of the harmonic oscillator. Fitting
the measured PSD in Figure 7 with this equation determines
froand calibrates the sensitivity of the AFM.
Finally, the transfer function of the cantilever can be de-
fined at any drive frequency f by its magnitude
|C| =
1
kc
1
?
?
1 +
f
fro
?2
(2)
and phase
θC= −atan
?
f
fro
?
.
(3)
B. Results
Two hundred approach curves were performed in the
span of 100 s. Both dAFM and sAFM signals were mea-
sured simultaneously. The dAFM was performed by oscillat-
ing the cantilever with a constant photothermal driving force
of 7 nN amplitude at 20 kHz, and measuring its magnitude
response |C| and phase response θC. All 200 approach curves
were highly reproducible, as demonstrated by the movie in
Figure 8. The high level of reproducibility (within noise) sug-
gests that the tip remained stable throughout the measurement
and warrants averaging the 200 approach curves.
Figure 9(a) presents the averaged dAFM data, along with
the averaged sAFM deflection. Note that the dAFM oscilla-
tionamplitudewasbelow200pm,whichismuchsmallerthan
the [BMIM][PF6] ion-pair diameter, and the sAFM deflection
remained below 60 pm across the solvation profile. Therefore,
the dAFM and sAFM are independent measurements25of the
solvation profile.
FIG. 8. Movie of 200 approach curves towards Au(111) in [BMIM][PF6],
acquired after the calibration in Figure 7. The magnitude response |C| and
phase response θCof the cantilever is shown. The thick lines represent the
averaged profiles, while the thin lines represent single approach curves. Two
approach curves were acquired per second, with an approach speed of 70
nm/s. The sample position was corrected for drift; the drift was mostly linear
with a rate of 2 pm/s. No isolation hood was used. (enhanced online) [URL:
http://dx.doi.org/10.1063/1.4712286.1].
InFigure9(b),theinteractionstiffnessanddamping mea-
sured by dAFM were obtained by23
?sinθC
and
ki=cosθC
|C|
where |Cs| and θCsare the magnitude and phase response mea-
sured at the start of the experiment during the calibration pro-
cedure. Although the calibration of the cantilever assumed a
massless harmonic oscillator model, Eqs. (4)and(5) apply for
AM-AFM at all Q values, and are mathematically identical to
standard AM-AFM theory.18
The interaction force measured by sAFM is proportional
to the tip position, related by kc, and the corresponding stiff-
ness, plotted in Figure 9(c), is calculated as the derivative of
interaction force with respect to tip-sample distance.
γi= −1
ω
|C|
−sinθCs
|Cs|
?
(4)
−cosθCs
|Cs|
,
(5)
VII. DISCUSSION
Despite resolving all the complications that arise due to
piezoacoustic excitation (discussed in Sec. II) by the use of
clean photothermal excitation, the interpretation of measured
signals remains non-trivial. This section discusses the results
in Figure 9.
A. Offset in stiffness and damping
Note that there are offsets in stiffness and damping at
4.5 nm tip-sample distance for the dAFM data in Figure 9(b).
These offsets represent changes in stiffness and damping be-
tween the 4.5 nm and ∼1μm tip-sample distances; the latter
is the distance at which the cantilever was calibrated.
The offsetin damping can be mostly attributed to squeeze
film damping of the cantilever. The offset in stiffness is due
Downloaded 10 May 2012 to 132.206.186.16. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

Page 8

053703-7Labuda et al.Rev. Sci. Instrum. 83, 053703 (2012)
FIG. 9. (a) Averaged data from Figure 8 and the simultaneously acquired
(averaged)deflectionsignal.(b)Themeasureddatawereconvertedintophys-
ical signals. (c) The independent measurements of stiffness (dynamic and
static) are plotted together. The two correspond very well (up to an offset)
far from the surface, however dynamic AFM breaks down near the surface
because the interaction stiffness becomes a large fraction of the cantilever
stiffness (28.5 N/m).
to either a calibration error or a true change in effective can-
tilever stiffness. These offsets are instrumental artifacts that
should be disregarded during the interpretation of tip-sample
physics, and could be minimized by calibrating the cantilever
much closer to the sample surface, say 15 nm instead of 1μm
away. In other words, these offsets relate to cantilever-sample
interactions, rather than tip-sample interactions.
Nevertheless, the offset in stiffness is worrisome, as it
suggests that the assumption of a constant cantilever stiff-
ness (in air versus liquid) may be wrong. In fact, because the
damping and mass loading of the cantilever are not uniform
across the full length of the cantilever (due to the proximity
to the surface at an angle), the modal shape of the cantilever
may be skewed and cause a change in effective stiffness26
that depends on the cantilever-sample distance. For the data in
Figure 9, this effect is small.
B. Dynamic vs static stiffness
Figure 9(c) overlays the stiffness profile as measured by
the dynamic and static AFM methods. Disregarding the offset
in dynamic stiffness, explained in Sec. VII A, the two datasets
agree very well beyond 2 nm from the surface. But below 2
nm, for the two or three last hydration layers, both methods
deviate significantly.
Hooke’s law, used to interpret the sAFM measurement,
remains accurate even for situations where the interaction
stiffness greatly exceeds the cantilever stiffness – as long as
the quasistatic condition is respected and no jump instabilities
occur. We believe the sAFM stiffness profile in Figure 9(c) to
be very accurate.
On the other hand, the stiffness profile extracted from
dynamic AFM becomes questionable close to the surface
because the solvation layer stiffness rises to a significant
percentage of the cantilever stiffness (>10%). Therefore, the
point-mass model approximation (Eq. (1)) begins to break
down in proximity to the sample surface as the shape of
the pinned cantilever takes precedence.27Stiffer cantilevers
are required to obtain accurate data for the solvation layers
closest to the surface. Note that this shortcoming is universal
to all dAFM techniques, irrespective of the Q factor of the
cantilever, and therefore does not affect the validity of the
massless model used to calibrate the cantilever. Even if the
calibration (Eqs. (1)-(3)) is perfectly accurate, the assumption
of a point-mass that led to the derivation of AM-AFM theory
(Eqs. (4)and(5)) is violated, and the dynamic stiffness
measurement is expected to be incorrect.
C. Oscillations in damping
Damping could not be measured with sAFM as it is an in-
trinsicallyvelocity-dependent quantity. Itishowever expected
that the damping profile measured by dAFM loses accuracy in
proximity to the surface for the same reasons as the stiffness
profile, described in Sec. VII B.
The monotonically increasing damping profile as the tip
approaches to 2 nm from the surface can therefore be trusted.
However, the oscillations in damping between 1–2 nm are
highly questionable, and the decrease in damping below 1 nm
is definitely meaningless.
D. Tip-sample distance
A commonly arising issue when imaging in liquids is the
determination of the tip-sample distance. The plot thickens
in viscous ionic liquids at the gold electrode as the last liq-
uid layer may require very large pressure to be displaced by
the tip. Therefore, there is no clear force regime between the
liquid removal, and the compression and damage of the gold
Downloaded 10 May 2012 to 132.206.186.16. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

Page 9

053703-8Labuda et al.Rev. Sci. Instrum. 83, 053703 (2012)
electrode; i.e., there is no hard-contact regime which can be
used to zero the tip-sample distance. In fact, it is possible
that the gold surface and the tip undergo significant defor-
mation before the last liquid layer is displaced. This observa-
tion is consistent with the low damage threshold (somewhere
between 40 nN and 150 nN) of the gold sample observed in
Figure 6. Given these complications, the tip-sample distance
in Figure 9 is poorly defined and somewhat arbitrary.
VIII. CONCLUSIONS
Our home-built AFM12was retrofit with a photothermal
excitation unit with minimal modification to the original de-
sign. Photothermal excitation provides a reliable method for
exciting cantilevers in liquid environments and allows much
easier interpretation of tip-sample physics than in the case of
piezoacoustic excitation.
Static and dynamic AFM can serve as complimentary
techniques for imaging the solid-liquid interface, allowing
imaging of both the solid surface and liquid layers, respec-
tively. In essence, both imaging methods can be combined
to span a higher dynamic range of force measurements and
imaging setpoints.
A comparison between dynamic and static force spec-
troscopy performed simultaneously suggests that dynamic
AFM reliability decreases as the interaction stiffness in-
creases up to and above the cantilever stiffness. This is ex-
plained by the fact that the cantilever changes modal shape
to a pinned cantilever, making the original point-mass ap-
proximation fail. Cantilevers much stiffer than the interaction
should be used to maintain accurate recovery of stiffness and
damping profiles.
The complications that arise when imaging in viscous
liquids are plentiful; they relate to cantilever dynamics and
strong tip-sample interactions (relative to the cantilever stiff-
ness). Photothermal excitation allowed us to identify and ad-
dress these difficulties confidently, because all the instrumen-
tal complications caused by piezoacoustic excitation were
abolished.
ACKNOWLEDGMENTS
We acknowledge valuable discussions with Hirofumi Ya-
mada, Daniel Kiracofe, and William Paul, as well as the
generosity of SPECS Surface Nano Analysis Inc., NSERC,
FQRNT, and CIfAR.
1R. García and R. Perez, Surf. Sci. Rep. 47, 197–301 (2002).
2R. Proksch and S. V. Kalinin, Nanotechnology 21, 455705 (2010).
3A. Labuda, Y. Miyahara, L. Cockins, and P. Grütter, Phys. Rev. B 84,
125433 (2011).
4A. Labuda, K. Kobayashi, D. Kiracofe, K. Suzuki, P. H. Grütter, and H.
Yamada, AIP Adv. 1, 022136 (2011).
5T. R. Albrecht, P. Grutter, D. Horne, and D. Rugar, J. Appl. Phys. 69, 668–
673 (1991).
6G. C. Ratcliff, D. A. Erie, and R. Superfine, Appl. Phys. Lett. 72, 1911
(1998).
7T. E. Schäffer, J. P. Cleveland, F. Ohnesorge, D. a. Walters, and P. K.
Hansma, J. Appl. Phys. 80, 3622 (1996).
8X. Xu and A. Raman, J. Appl. Phys. 102, 034303 (2007).
9D. Kiracofe and A. Raman, Nanotechnology 22, 485502 (2011).
10G. Meyer and N. M. Amer, Appl. Phys. Lett. 53, 1045–1047
(1988).
11To be exact, the bending shape converges to perfectly circular as
f → 0 because the cantilever temperature profile becomes uniform at long
timescales. This suggests a frequency-dependence and a deviation from the
first-eigenmode approximation used in this article. However, the deviations
are small (a few percent) in the usable frequency range for dAFM using
photothermal excitation, and are negligible in our comparison with piezoa-
coustic excitation where deviations can exceed an order of magnitude.
12A. Labuda, W. Paul, B. Pietrobon, R. B. Lennox, P. H. Grütter, and R.
Bennewitz, Rev. Sci. Instrum. 81, 083701 (2010).
13A. Labuda, F. Hausen, N. N. Gosvami, P. H. Grütter, R. B. Lennox, and R.
Bennewitz, Langmuir 27, 2561–2566 (2011).
14A. Labuda and P. H. Grütter, Rev. Sci. Instrum. 82, 013704 (2011).
15T. Fukuma, M. Kimura, K. Kobayashi, K. Matsushige, and H. Yamada,
Rev. Sci. Instrum. 76, 053704 (2005).
16D. Kiracofe, K. Kobayashi, A. Labuda, A. Raman, and H. Yamada, Rev.
Sci. Instrum. 82, 013702 (2011).
17S. N. Magonov, V. Elings, and M.-H. Whangbo, Surf. Sci. 375, L385–L391
(1997).
18S. J. O’Shea and M. E. Welland, Langmuir 14, 4186–4197 (1998).
19J. E. Sader, J. W. M. Chon, and P. Mulvaney, Rev. Sci. Instrum. 70, 3967
(1999).
20M. J. Higgins, R. Proksch, J. E. Sader, M. Polcik, S. Mc Endoo, J. P. Cleve-
land, and S. P. Jarvis, Rev. Sci. Instrum. 77, 013701 (2006).
21O. Marti, A. Ruf, M. Hipp, H. Bielefeldt, J. Colchero, and J. Mlynek,
Ultramicroscopy 42–44, 345–350 (1992).
22A. Labuda, J. R. Bates, and P. H. Grütter, Nanotechnology 23, 025503
(2012).
23A. Labuda and P. Grütter, Langmuir 28, 5319–5322 (2012).
24M. T. Clark, J. E. Sader, J. P. Cleveland, and M. R. Paul, Phys. Rev. E 81,
046306 (2010).
25S. Kawai, T. Glatzel, S. Koch, B. Such, A. Baratoff, and E. Meyer, Phys.
Rev. B 80, 085422 (2009).
26D. Kiracofe and A. Raman, J. Appl. Phys. 107, 033506 (2010).
27U. Rabe, K. Janser, and W. Arnold, Rev. Sci. Instrum. 67, 3281
(1996).
Downloaded 10 May 2012 to 132.206.186.16. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions