High-speed atomic force microscopy in liquid T. Sulchek,a) R. Hsieh

advertisement
REVIEW OF SCIENTIFIC INSTRUMENTS
VOLUME 71, NUMBER 5
MAY 2000
High-speed atomic force microscopy in liquid
T. Sulchek,a) R. Hsieh, J. D. Adams, S. C. Minne, and C. F. Quate
E. L. Ginzton Laboratory, Stanford University, California 94305-4085
D. M. Adderton
NanoDevices Incorporated, Santa Barbara, California 93105
~Received
23 November 1999; accepted for publication 17 January 2000!
High-speed constant force imaging with the atomic force microscope ~AFM! has been achieved in
liquid. By using a standard optical lever AFM, and a cantilever with an integrated zinc oxide ~ZnO!
piezoelectric actuator, an imaging bandwidth of 38 kHz has been achieved; nearly 100 times faster
than conventional AFMs. For typical samples, this bandwidth corresponds to tip velocities in excess of
3 mm/s. High-speed AFM imaging in liquid will ~1! permit chemical and biological AFM observations
to occur at speeds previously inaccessible, and ~2! significantly decrease measurement times in
standard AFM liquid operation. © 2000 American Institute of Physics.
@S0034-6748~00!01705-6#
I. INTRODUCTION
One of the advantages of the atomic force microscope
is its ability to image a wide range of samples in a wide
range of environments. Imaging in liquids offers the potential
to investigate biological samples in their physiological
environment as well as eliminate capillary forces between the
tip and the sample.1,2 Lindsay et al.,3 Weisenhorn et al.,4 Butt
et al.,5 and many more,6 have shown that nanometer scale
resolution on biological samples can be achieved in
physiological conditions.
Contact mode AFM operation involves scanning a tip over
a surface such that the tip and sample are continuously in
contact. The contact mode of imaging is subdivided into two
modes: Constant height imaging and constant force imaging. In
constant height imaging, the tip remains at afixed height
relative to the sample, but the force between the tip and the
sample vary with topography. Variation of the tip/ sample
force can cause unwanted imaging artifacts as well as
premature tip wear. Imaging in the constant height mode is a
special problem with soft samples, where vertical forces can
cause deformation, and shear forces can sweep delicate
samples completely off a surface.7 These problems can be
minimized by maintaining a constant force on the sample to
insure reliable contact mode AFM operation.
In the constant force mode of imaging, the deflection of
the cantilever, and hence the force applied to the surface, is
kept fixed. This is accomplished by monitoring the deflection
of the AFM cantilever, and using that signal in a feedback loop
to control the vertical position of the sample. In this
configuration, the tip, and the force applied by the tip, remain
fixed while the sample scans in three dimensions beneath the
tip. In most AFMs, the three dimensional scanner is a conventional piezotube. The tube bends laterally to perform the x–
y scan, and extends vertically in the z direction to maintain the
constant force on the tip. The conventional constant force
mode AFM operates as follows: As the cantilever scans
~AFM!
a!
Electronic mail: sulchek@stanford.eduover
0034-6748/2000/71(5)/2097/3/$17.00
produces a signal representing a deviation from constant force
~an error signal!. This signal is directed through a
proportional-integral-differential ~PID! feedback circuit,
amplified, and then sent to the piezotube z-axis actuator. The
piezotube will move so as to follow the sample topography in
order to restore the constant force, or return the error signal to
zero.
One of the primary disadvantages of the AFM is the slow
imaging speed. The speed of the AFM is limited by the speed
of the PID control loop, which is in turn limited by the
resonant frequency of the piezotube. Piezotubes commonly
have resonant frequencies of a few hundred hertz. For most
samples, a bandwidth of a few hundred hertz corresponds to
scan speeds of a few tens of microns per second. At this rate, a
single moderately sized 5123512 pixel image can take several
minutes to acquire. Observing biological and chemical
processes that occur on time scales faster than this are beyond
the reach of AFM systems based on piezotubes.
High speed imaging has become a hallmark advance in
constant force AFM operation. One of the main techniques to
increase the speed of the AFM is to replace the primary
feedback actuator ~the slow piezotube! with a smaller, faster
actuator. MEMS integration of a piezoelectric actuator onto the
cantilever has been an approach that several groups have
used.8–11 Manalis et al.12 have achieved tip speeds of up to 1
cm/s by integrating a thin layer of zinc oxide ~ZnO! on the
base of a cantilever while measuring deflection with a standard
optical lever system.
However, up to now, high speed operation in contact mode
has been limited to operation in air. Viani et al.13 have used
small cantilevers with high resonant frequencies to obtain fast
images of biopolymers in liquid using the tapping mode. Lee et
al.14 used an integrated piezoelectric actuator in fluid to drive a
cantilever into resonance and sensed deflection by measuring
current to the actuator. This method proved useful for
intermittent contact mode imaging, although the bandwidth
was unsuitable for high speed operation.
a surface, its deflection
2097
© 2000 American Institute of Physics
Downloaded 08 Apr 2004 to 133.28.19.12. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
2098
Rev. Sci. Instrum., Vol. 71, No. 5, May 2000
Sulchek e t al .
FIG. 3. The closed loop transfer function of a ZnO cantilever, showing both
amplitude and phase. With a margin in phase of 45 degrees, the imaging
bandwidth is 38 kHz.
FIG. 1. A schematic diagram for an AFM with an integrated piezoelectric
actuator operating in liquid. The laser beam travels through the Plexiglas fluid
cell, through the drop of liquid, and is reflected off the back of the cantilever
into a split photodiode. A meniscus holds the liquid in place between the fluid
cell and the sample.
circuiting through the fluid. Several materials for passivation
were used including spray coated Teflon, poly~methylmethacrylate! electron resist, and poly-dimethylsiloxane
~PDMS! elastomer.16 We found the best results to be ob-
II. EXPERIMENT
A schematic of the system used for this work is shown in
Fig. 1. The silicon cantilever consists of a piezoelectric ZnO/
silicon bimorph. This bimorph forms the integrated actuator.
The actuation occurs when a voltage is applied to the ZnO
electrodes. The electric field expands or contracts the ZnO in
proportion to the field; this creates stress between the piezoelectric and the silicon which is relieved by bending. The
device used here has been described previously.15
During imaging, the cantilever is bent to follow the
sample topography by applying a voltage to the ZnO electrodes that is proportional to the tip–sample force signal, as
detected with an optical lever. The voltage signals are sent
from the PID control amplifier to the cantilever holder,
through a coaxial cable. The holder connects to the cantilever’s die through wire bonds. On the cantilever die, the wire
bonds connect to thin film metal traces that extend to the ZnO
actuator.
A drop of liquid is inserted between the sample and the
fluid cell so that the cantilever is completely immersed in the
liquid. The ZnO actuator, the traces along the cantilever die,
and the wire bonds from the die to the cantilever holder all
must be passivated to protect the electrical signals from short
FIG. 2. The transfer function of a freely vibrating ZnO cantilever in both air
and water. The resonance is both damped and shifted to a lower frequency by
the more viscous liquid environment. The Q changes from 110 to 3 and the
frequency shifts from 56 to 35 kHz. The drive amplitude was 100 mV peak to
peak, and the photodiode was filtered to a 1 MHz bandwidth.
FIG. 4. Two high speed images in liquid. ~a! is a one micron scan of the
surface of gold taken at a tip speed of 0.8 mm/s ~scan rate of 400 Hz!. The
grains of gold have a height of approximately 5 nm and a lateral dimension of
approximately 50 nm. ~b! shows an image of a grating with a 10 micron pitch,
taken at a tip speed of over 3 mm/s ~scan rate of 55 Hz!. Imaging times for
these 5123512 pixel images were approximately 1 and 10 s, respectively.
Downloaded 08 Apr 2004 to 133.28.19.12. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
Rev. Sci. Instrum., Vol. 71, No. 5, May 2000
tained by passivation with the PDMS elastomer. The elastomer
was applied by dipping the holder, die, and the cantilever’s
actuator directly into the elastomer. Micromanipulators were
used to ensure that the entire actuator region was covered,
while leaving the tip region uncovered. We found the
passivation to hold up well in both pure water and ionic
solutions, as long as the elastomer completely covered the
actuator region.
I I I . R E S U L T S AN D D I S C U S S I O N
Figure 2 shows the transfer function of the freely vibrating
ZnO cantilever in air and in water. In air, the resonant
frequency is 56 kHz with a Q of 110. In liquid, we expect a
lower resonant frequency due to the increased damping effect
of water.17 In liquid the resonance drops to 35 kHz and the Q to
3. A particularly advantageous result from using an integrated
actuator in liquid is that the system’s response shows none of
the spurious peaks associated with external drives.18
The bandwidth of the system is determined by the actuator’s ability to follow a modulation of the setpoint while the tip
is in contact with a surface. In Fig. 3 we plot the closed loop
frequency response of the ZnO cantilever in feedback. When
we set the phase margin at 45 degrees, the bandwidth of the
system is 38 kHz, almost 100 times that of a piezotube
~typically 600 Hz!. The minimum detectable distance is 0.3 Å
in a 12 kHz bandwidth; showing no loss of resolution for the
high speed system.
Two high speed images taken in water are shown in Figs.
4~a! and 4~b!. The images were taken by raster scanning the
sample using a 1 in. long piezotube. Figure 4 ~a! is a one
micron scan of the surface of gold. The scan rate was 400 Hz
corresponding to a tip speed of 0.8 mm/s. The grains of gold,
with a height of under 5 nm, are approximately 50 nm across.
This high resolution ~5123512 pixels! image was acquired in
just over one second. Figure 4 ~b! is a 200 nm grating with a 10
mm pitch taken with a scan size of 30 microns at a scan rate of
55 Hz ~3 mm/s tip speed!.
For images containing 200 lines, a scan rate of 400 Hz
allows two frames to be acquired each second. At this refresh
rate, we were able to center and zoom in on a single feature
with real time feedback. Although our servo loop will oper-ate
with faster tip speeds, a mechanical resonance in our x – y
scanner limits the scan rate. The ability to acquire several
images every second allows the possibility of real-time nanometer scale biological and chemical imaging.
AFM in liquid
2099
ACKNOW LEDGM ENTS
The authors would like to thank Jim Zesch for all of his
help. One of the authors ~J.D.A.! acknowledges the support of
a NSF graduate fellowship, another author ~D.M.A.! acknowledges support from NSF DMI-9903522. This work was
supported by the Defense Advanced Research Projects Agency
and the Semiconductor Research Corporation.
1
B. Drake, C. B. Prater, A. L. Weisenhorn, S. A. Gould, T. R. Albrecht, C.
F. Quate, D. S. Cannell, H. G. Hansma, and P. K. Hansma, Science 243,
1586 ~1989!.
2
A. L. Weisenhorn, P. K. Hansma, T. R. Albrecht, and C. F. Quate, Appl.
Phys. Lett. 54, 2651 ~1989!.
3 S. M. Lindsay, L. A. Nagahara, T. Thundat, U. Knipping, and R. L. Rill,
J. Biomol. Struct. Dyn. 7, 279 ~1998!.
4
A. L. Weisenhorn, M. Egger, F. Ohnesorge, S. A. C. Gould, and S-P. Heyn,
Langmuir 7, 8 ~1991!.
5
H. J. Butt, E. K. Wolff, S. A. Gould, N. B. Dixon, C. M. Peterson, and P.
K.
Hansma, J. Struct. Biol. 105, 54 ~1990!.
6
Z. Shao and J. Yang, Quarterly Rev. Biophys. 28, 195 ~1995!.
7
T. Shibata-Saka, J. Masai, T. Tagawa, T. Sorin, and S. Kondo, Thin Solid
Films 273, 297 ~1996!.
8 S. Akamine, T. R. Albrecht, M. J. Zdeblick, and C. F. Quate, IEEE Electron Device Lett. 10, 490 ~1989!.
9 T. Itoh and T. Suga, in The Proceedings of Transducers ’95 ~Royal Swedish Academy of Engineering Science, Stockholm, 1995!, Vol. IV A, p. 632.
10
T. Fujii, S. Watanabe, M. Suzuki, and T. Fujiu, J. Vac. Sci. Technol. B 13,
1119 ~1995!.
11
S. R. Manalis, S. C. Minne, and C. F. Quate, Appl. Phys. Lett. 68, 871
~1996!.
12 S. R. Manalis, S. C. Minne, A. Atalar, and C. F. Quate, Rev. Sci. Instrum.
67, 3294 ~1996!.
13
M. B. Viani et al., Rev. Sci. Instrum. 70, 4300 ~1999!.
14
C. Lee, T. Itoh, T. Ohashi, R. Maeda, and T. Suga, J. Vac. Sci. Technol. B
15, 1559 ~1997!.
15
S. C. Minne, S. R. Manalis, and C. F. Quate, Appl. Phys. Lett. 67, 3918
~1995!.
16
Poly-dimethylsiloxane obtained from Dow Corning.
17
C. Y. Chen, R. J. Warmack, T. Thundat, D. P. Allison, and A. Huang, Rev.
Sci. Instrum. 65, 2532 ~1994!.
18
C. A. J. Putman, K. O. Van der Werf, B. G. De Grooth, N. F. Van Hulst,
and J. Greve, Appl. Phys. Lett. 64, 2454 ~1994!.
Downloaded 08 Apr 2004 to 133.28.19.12. Redistribution subject to AIP license or copyright, see http://rsi.aip.org/rsi/copyright.jsp
Download