Constant tip-surface distance with atomic force microscopy via quality factor feedback

Lin Fan, Daniel Potter, and Todd Sulchek Citation: Rev. Sci. Instrum. 83, 023706 (2012); doi: 10.1063/1.3683236 View online: http://dx.doi.org/10.1063/1.3683236 View Table of Contents: http://rsi.aip.org/resource/1/RSINAK/v83/i2 Published by the American Institute of Physics.

Related Articles
Contrast distortion induced by modulation voltage in scanning capacitance microscopy Appl. Phys. Lett. 101, 083503 (2012) Internal resonance based sensing in non-contact atomic force microscopy Appl. Phys. Lett. 101, 053106 (2012) Joint strength measurements of individual fiber-fiber bonds: An atomic force microscopy based method Rev. Sci. Instrum. 83, 073902 (2012) Design and performance of a combined secondary ion mass spectrometry-scanning probe microscopy instrument for high sensitivity and high-resolution elemental three-dimensional analysis Rev. Sci. Instrum. 83, 063702 (2012) High-speed Lissajous-scan atomic force microscopy: Scan pattern planning and control design issues Rev. Sci. Instrum. 83, 063701 (2012)

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 24 Aug 2012 to 130.207.50.37. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

REVIEW OF SCIENTIFIC INSTRUMENTS 83, 023706 (2012)

Constant tip-surface distance with atomic force microscopy via quality factor feedback
Lin Fan, Daniel Potter, and Todd Sulchek
George W. Woodruff School of Mechanical Engineering, Parker H. Petit Institute for Bioengineering and Bioscience, Georgia Institute of Technology, Atlanta, Georgia 30332, USA

(Received 18 October 2011; accepted 21 January 2012; published online 10 February 2012) The atomic force microscope (AFM) is a powerful and widely used instrument to image topography and measure forces at the micrometer and nanometer length scale. Because of the high degree of operating accuracy required of the instrument, small thermal and mechanical drifts of the cantilever and piezoactuator systems hamper measurements as the AFM tip drifts spatially relative to the sample surface. To compensate for the drift, we control the tip-surface distance by monitoring the cantilever quality factor (Q) in a closed loop. Brownian thermal uctuations provide sufcient actuation to accurately determine cantilever Q by tting the thermal noise spectrum to a Lorentzian function. We show that the cantilever damping is sufciently affected by the tip-surface distance so that the tip position of soft cantilevers can be maintained within 40 nm of a setpoint in air and within 3 nm in water with 95% reliability. Utilizing this method to hover the tip above a sample surface, we have the capability to study sensitive interactions at the nanometer length scale over long periods of time. 2012 American Institute of Physics. [doi:10.1063/1.3683236]
I. INTRODUCTION

Atomic force microscopy (AFM) is a prominent tool for nanoscale measurements1 including the study of surface topography and molecular interactions.2 Thermal bending of the cantilever3 and thermal drift in the AFM components4, 5 both result in a time-varying tip-surface distance. Accurate tip-surface distance control is critical to several imaging6 and force measurement modes. For example, measurement of interaction potentials with Brownian force reconstruction is extremely sensitive to tip-sample drift.79 Moreover, the ability to stably hover the tip over the surface for long periods of time would allow sampling of rare attractive events which could then be used to probe deeper into the attractive energy potential. Several methods have been developed in an attempt to reduce the impact of drift on AFM measurements, including optical tracking of the probe,10 drift prediction to compensate for motion, attempts to control thermal conditions to which the AFM is exposed, or simply completing measurements quickly to avoid drift altogether.4, 11 Thermal uctuations of the cantilever were also monitored to feedback to the change in resonance frequency for non-contact AFM.12 The separation distance between the cantilever tip and the surface strongly affects the cantilever dynamics through changes in the cantilever damping and mass loading. While hydrodynamic damping, due to the inertial and viscous components of uid resistance, is largely independent of the tipsurface distance, squeeze damping, an additional drag due to the forcing of uid out from the gap between the tip and the surface, is inversely related to the time-averaged tipsurface distance.1318 The quality factor (Q), inversely related to the damping of an oscillator, is characterized by the sharpness of a resonance response. Because Q decreases with increased damping, Q correlates positively with tip-surface distance. The effect of decreasing tip-surface distance therefore decreases the cantilever Q until the cantilever approaches
0034-6748/2012/83(2)/023706/4/$30.00

within a few nanometers of the surface, at which point the snap-to-contact point occurs.19 In viscous uids such as water, Q has also been observed to decrease with distance, but to a lesser degree due to more dominant hydrodynamic damping.13, 15, 17, 18 This study explores a new method to minimize tipsurface drift by feeding back to the cantilever quality factor (Q). Because the cantilever Q is a monotonic function of the tip-surface distance, maintaining a constant Q also maintains a constant distance. We rst determine the relationship between Q and tip-surface distance. We then implement a proportional feedback controller to translate the AFM vertical piezoactuator as necessary to maintain a constant Q and hence, a constant tip-surface distance. We demonstrate the simple and fast method in air and the applicability of the method in water environments.

II. MATERIALS AND METHODS

We used an atomic force microscope (MFP-3D from Asylum Research) integrated with an inverted optical microscope (Nikon TiE) for all experiments. Real time data acquisition, analysis, and control routines were written in IGOR Pro 6.0 (WaveMetrics). For experiments in air, we used soft, triangular silicon nitride cantilevers (Bruker MLCT, cantilevers D, E, and F), with nominal spring constants of 0.03 N/m, 0.1 N/m, and 0.6 N/m and natural resonance frequencies in air of 15 kHz, 38 kHz, and 125 kHz, respectively. For experiments in water, we examined the second resonance frequency of cantilever F, which has a natural second resonance frequency of 230 kHz in water. All experiments were conducted in ambient air or distilled water and cleaned glass microscope slides were used as surfaces. We determined the quality factors in real time by tting a Lorentzian function to the averaged thermal noise spectra. Thermal noise spectra were collected with
2012 American Institute of Physics

83, 023706-1

Downloaded 24 Aug 2012 to 130.207.50.37. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

023706-2

Fan, Potter, and Sulchek

Rev. Sci. Instrum. 83, 023706 (2012)

FIG. 1. Schematic of feedback control loop and AFM system. The tipsurface distance changes with time due to drift of the AFM components.

a bandwidth from 1 to 500 kHz and averaged from 10 to 1000 times. We used cantilever force-distance measurements to accurately determine tip-surface distance.
III. RESULTS AND DISCUSSION A. Feedback control

We used the transfer function describing the relationship between Q and tip-surface distance in our feedback control. To maintain a constant tip-surface distance using feedback, Q is measured and compared to the setpoint to obtain the Q error, as shown in Figure 1. The Q error is input to a simple proportional controller with a gain of unity, which uses the transfer function to determine the tip-surface drift distance. The controller converts the distance to the corresponding voltage signal necessary to negate the cantilever drift through movement of the piezoactuator. With this control, we can maintain a constant Q and equivalently, a constant tip-surface distance.
B. Cantilever Q versus tip-surface distance in air

FIG. 2. Quality factor dependence on distance above snap-to-contact with linear relationship (Q = 2.62Z + 13.17) shown for distances greater than 0.03 m. Cantilever Q is linearly dependent on tip-surface distance far from the surface (Z > 0.1 m), while closer to the surface (Z < 0.1 m), the sensitivity drastically increases. The snap-to-contact of the tip to the surface occurs because of capillary wetting attraction at the probe tip. The inset plot shows thermal resonance curves at corresponding points identied in the main plot.

We determined the relationship between the cantilever Q and tip-surface distance by acquiring thermal noise spectra as we translated the tip towards the surface in small steps for distances from 3 m to contact with the surface (Figure 2). Thermal spectra were collected through a fast Fourier transform of the cantilever deection data (Figure 2 inset) and Q values were determined from real time Lorentzian curve ts. The probe was translated progressively closer to the surface with decreasing step size to accurately determine the changes in Q due to more signicant squeeze damping near the surface. These data were collected in air using a cantilever (E) with a spring constant of 0.1 N/m and a natural resonance frequency of 38 kHz. The data in Figure 2 show that Q scales linearly with tip-surface distance (R2 = 0.997) down to 0.03 m above the snap-to-contact point, before decreasing dramatically over small step sizes. As a result, in the linear region we can accurately determine changes in tip-surface distance by measuring changes in cantilever Q.
C. Tip-surface distance control in air

cantilever (E) over 24 h with and without feedback. Without feedback, the tip drifted almost a micrometer during the rst 5 h of the experiment, even after signicant equilibration. With feedback, the tip-surface distance setpoint (1.67 m) was continuously maintained with no steady-state error. It is important to note that drift changes nonlinearly with time and uctuates unpredictably. Therefore, drift compensation methods based upon linearly increasing drift would not be effective over long time periods. From the uctuation of the Q measurement in the control data in Figure 3, the standard deviation of tip position relative to the setpoint was 35 nm. Therefore, based on a normal distribution of tip position deviation, we expect this method to maintain a tip within 70 nm of a setpoint in air with 95% reliability using cantilever E. It is important to note that the method can be improved by increasing the acquisition rate and precision of Q measurements. Also, in air, the soft lever exhibited signicant snap-to-contact behavior near the

By maintaining a constant cantilever Q through the proportional control loop we can hold a desired tip-surface distance and eliminate experimental drift. In Figure 3, we show the stark difference in drift of the tip-surface distance of a

FIG. 3. Distance above snap-to-contact with and without proportional feedback control with time. The tip position was determined from the cantilever Q.

Downloaded 24 Aug 2012 to 130.207.50.37. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

023706-3

Fan, Potter, and Sulchek

Rev. Sci. Instrum. 83, 023706 (2012)

TABLE I. Effect of cantilever geometry on the precision of tip-surface distance control in air. The setpoint distance is 1.6 m. Cantilever (Bruker MLCT) D E F Length, width (m) 225, 20 140, 18 85, 18 Spring constant (N/m) 0.03 0.1 0.6 Hovering error (nm) 160 100 50

surface. This increased the difculty of keeping the tip steady below tip-surface distances of 30 nm.
D. Effect of cantilever properties and setpoint distance on tip-surface distance control in air

We explored the effects of cantilever properties such as size and spring constant on the precision of tip-surface distance control in air. We used three cantilevers (D, E, and F) with different sizes and spring constants in feedback control at a setpoint distance of 1.6 m. The data in Table I show that the setpoint can be maintained with less error using smaller, stiffer cantilevers. The hovering error refers to the distance from a setpoint within which we are able to maintain the tip with 95% reliability based on a normal distribution of tip position deviation, equal to two standard deviations. At least 8 h of data were collected in each instance. We also investigated the effect of setpoint distance on the precision of tip-surface distance control in air using cantilever F at three setpoints. As shown in Table II, setpoints closer to the surface yield a smaller hovering error. The same technique was used to measure cantilever Q in all cases and no steady-state error was observed. The data show that we are capable of maintaining a tip within 40 nm of a setpoint based on a normal distribution of tip position deviation, and even greater precision may be achieved by hovering closer to the surface using smaller, stiffer cantilevers.
E. Cantilever Q measurement precision

FIG. 4. Relationship between the number of samples collected per measurement and the standard deviation of Q measurements using cantilever E in air. The inset plot shows the same relationship for a smaller, stiffer cantilever (F) in water.

ples and hence, the measurement time, while still maintaining precision is important for minimizing tip-surface drift during data acquisition. In mapping the relationship between Q and tip-surface distance, 250 samples of the thermal spectra were averaged for each Q measurement in air, which required 2.5 min. In water, between 20 and 40 samples were averaged per measurement, which required between 20 and 40 s. The time taken for Q measurements can be optimized by considering sources of error resulting from both the Q measurement variability and the drift rate.

F. Cantilever Q versus tip-surface distance in water

The current methods we use to measure the cantilever Q result in a tradeoff between the precision of the measurement and the data collection time. More precise measurements can be achieved by simply increasing the number of samples and averaging the thermal spectra used in the Lorentzian curve tting. In Figure 4, we show the standard deviation of Q measurement versus the number of samples per measurement using cantilever E for experiments in air. From this plot we see that there is a signicant increase in precision achieved by collecting over 200 samples. Optimizing the number of samTABLE II. Cantilever F: Effect of setpoint distance on hovering error in air. Setpoint distance (m) 2.6 1.6 0.6 Hovering error (nm) 70 50 40

Performing AFM measurements in water environments is necessary to probe biological interactions. One advantage of AFM measurements in liquid is that the cantilever tip can be brought closer to the surface without capillary wetting attraction. However, damping in water is signicantly higher than in air due to increased viscous damping.20 To explore the possibility of Q feedback in a liquid, we repeated the Q versus tip-surface distance measurements in water to determine if a monotonic decrease in Q exists near the surface. We did not observe a sufcient change in Q with distance in the rst resonance mode. However, we found that damping in liquids is less signicant for higher order resonance modes and for stiffer cantilevers. Therefore, we utilized the second resonance peak of a small, stiff cantilever (F) with a spring constant of 0.6 N/m and a second resonance frequency of 230 kHz in water. Thermal spectra were collected and Q was measured for tip-surface distances less than one micrometer. Cantilever force-distance measurements were taken before and after each Q measurement to accurately quantify tip-surface distance, and the time was recorded to monitor the drift rate. The tip-surface distance was determined as the average of the distance before and after each Q measurement to account for drift during the measurement process. The average drift rate observed in water experiments was 0.19 0.06 nm/s. Several Q measurements were gathered at specic tipsurface distances and averaged. Cantilever Q showed minimal change with distance until the tip was close to the surface.

Downloaded 24 Aug 2012 to 130.207.50.37. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions

023706-4

Fan, Potter, and Sulchek

Rev. Sci. Instrum. 83, 023706 (2012)

FIG. 5. Quality factor dependence on tip-surface distance in water. Cantilever Q remains constant at distances of more than 10 nm, but appears to decrease suddenly at closer distances (dZ/dQ = 8 nm). The error bars represent one standard deviation from the mean. At least 15 Q measurements were taken around each point. The inset plot shows the second resonance peak and Lorentzian curve t at a tip-surface distance of 10 nm.

We have demonstrated that the tip position of soft levers can be maintained within 40 nm of a setpoint (2 standard deviations) in air over the course of many hours. The sensitivity of the method is sufcient to achieve 3 nm hovering accuracy in water. Utilizing smaller, stiffer cantilevers and hovering closer to the surface yielded greater precision in air. In addition to increasing the speed and precision of Q measurements, we plan to investigate different cantilever geometries, oscillation amplitudes, and media to improve the method in the future. This capability may be utilized to study sensitive interactions between the tip and a sample surface over long periods of time. Because the sizes of biological molecules such as proteins are commensurate to the scale of the hovering accuracy, this technique could potentially prove useful in studying biomolecular interactions.21
ACKNOWLEDGMENTS

In Figure 5, we plot the relationship close to the surface and observe a decrease in Q from less than 10 nm to 2 nm. The decrease in Q is monotonic close to the surface. Therefore, the same principle of Q feedback can be used to maintain constant tip-surface distance in liquid environments. We can estimate the hovering error in water by calculating the variability in tip-surface distance during feedback, Z = Q(dZ/dQ), where Z is the tip-surface distance, and Q and Z are the variability of Q and Z measurements, respectively. Assuming a linear correspondence between Q and Z at less than 5 nm above the surface, a normal distribution for Q measurement deviation, and a Q value of 0.2 (the standard deviation of the Q measurement), we estimate that tip position can be maintained within 3 nm of a setpoint in water with 95% reliability using cantilever F. As with hovering in air, it is critical to increase the speed and precision with which the cantilever Q can be measured to improve the method in water. Optimizing cantilever properties such as size and stiffness may allow for more precise hovering as well. These results for experiments in water agree with past studies of thermally driven and low-amplitude, acoustically driven AFM cantilever oscillations in liquid where cantilever Q remains relatively constant until reaching a tip-surface distance of less than 10 nm, at which point Q decreases. The observed decrease in Q could be a combination of several factors such as squeeze lm effects and increased viscous drag near a solid surface.13, 15, 17, 18 However, the decrease cannot be attributed to intermittent contact of the cantilever tip with the surface as the thermally excited oscillation amplitudes for the experiments were over an order of magnitude less than the measured tip-surface distances.
IV. CONCLUSIONS

The authors thank the National Science Foundation (NSF) for support of this research (CBET-CAREER1055437). The authors also thank the Presidents Undergraduate Research Award (PURA) program at Georgia Tech for providing funding to engage undergraduates in research.

We have shown that the cantilever quality factor can be used to accurately maintain a constant tip-surface distance.

Binnig, C. F. Quate, and C. Gerber, Phys. Rev. Lett. 56(9), 930 (1986). Rief, F. Oesterhelt, B. Heymann, and H. E. Gaub, Science 275(5304), 1295. 3 M. Radmacher, J. P. Cleveland, and P. K. Hansma, Scanning 17(2), 117 (1995). 4 E. Tranvouez, E. Boer-Duchemin, G. Comtet, and G. Dujardin, Rev. Sci. Instrum. 78(11), 5 (2007). 5 D. Rugar, H. J. Mamin, R. Erlandsson, J. E. Stern, and B. D. Terris, Rev. Sci. Instrum. 59(11), 2337 (1988). 6 Y. Martin, C. C. Williams, and H. K. Wickramasinghe, J. Appl. Phys. 61(10), 4723 (1987). 7 J. P. Cleveland, T. E. Schaffer, and P. K. Hansma, Phys. Rev. B 52(12), R8692 (1995). 8 O. H. Willemsen, L. Kuipers, K. O. van der Werf, B. G. de Grooth, and J. Greve, Langmuir 16(9), 4339 (2000). 9 P. D. Ashby and C. M. Lieber, J. Am. Chem. Soc. 126(51), 16973 (2004). 10 G. M. King, A. R. Carter, A. B. Churnside, L. S. Eberle, and T. T. Perkins, Nano Lett. 9(4), 1451 (2009). 11 M. Abe, Y. Sugimoto, T. Namikawa, K. Morita, N. Oyabu, and S. Morita, Appl. Phys. Lett. 90(20), 203103 (2007). 12 A. Gannepalli, A. Sebastian, J. Cleveland, and M. Salapaka, Appl. Phys. Lett. 87(11), 111901 (2005). 13 C. Harrison, E. Tavernier, O. Vancauwenberghe, E. Donzier, K. Hsu, A. R. H. Goodwin, F. Marty, and B. Mercier, Sens. Actuators, A 134(2), 414 (2007). 14 G. Meyer and N. M. Amer, Appl. Phys. Lett. 53(12), 1045 (1988). 15 I. Nnebe, and J. W. Schneider, Langmuir 20(8), 3195 (2004). 16 G. Leveque, P. Girard, S. Belaidi, and G. C. Solal, Rev. Sci. Instrum. 68(11), 4137 (1997). 17 S. Basak, A. Raman, and S. V. Garimella, J. Appl. Phys. 99(11), 114906 (2006). 18 T. Naik, E. K. Longmire, and S. C. Mantell, Sens. Actuators, A 102(3), 240 (2003). 19 M. Luna, J. Colchero, J. Gomez-Herrero, and A. M. Baro, Appl. Surf. Sci. 157(4), 285 (2000). 20 T. Sulchek, R. Hsieh, J. D. Adams, S. C. Minne, C. F. Quate, and D. M. Adderton, Rev. Sci. Instrum. 71(5), 2097 (2000). 21 O. H. Willemsen, M. M. E. Snel, A. Cambi, J. Greve, B. G. De Grooth, and C. G. Figdor, Biophys. J. 79(6), 3267 (2000).
2 M.

1 G.

Downloaded 24 Aug 2012 to 130.207.50.37. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions