Advanced Control of Atomic Force Microscope for Faster Image ...

6 downloads 239160 Views 1MB Size Report
Dec 20, 2013 - Abstract. In atomic force microscopy (AFM), the dynamics and nonlinearities of its nanopositioning stage are major sources of image distortion, ...
Chapter 19

Advanced Control of Atomic Force Microscope for Faster Image Scanning M. S. Rana, H. R. Pota and I. R. Petersen

Abstract In atomic force microscopy (AFM), the dynamics and nonlinearities of its nanopositioning stage are major sources of image distortion, especially when imaging at high scanning speed. This chapter discusses the design and experimental implementation of an observer-based model predictive control (OMPC) scheme which aims to compensate for the effects of creep, hysteresis, cross-coupling, and vibration in piezoactuators in order to improve the nanopositioning of an AFM. The controller design is based on an identified model of the piezoelectric tube scanner (PTS) for which the control scheme achieves significant compensation of its creep, hysteresis, cross-coupling, and vibration effects and ensures better tracking of the reference signal. A Kalman filter is used to obtain full-state information about the plant. The experimental results illustrate the use of this proposed control scheme.

19.1 Introduction Nanotechnology is an area of modern science which deals with the control of matter at dimensions of 100 nm or less. In recent years, of all the available microscopy techniques, atomic force microscopy (AFM) has proved itself extremely versatile as an investigative tool in this field. It is becoming a driving technology in nanomanipulation and nanoassembly [1, 2]. M. S. Rana (B) · H. R. Pota · I. R. Petersen School of Engineering and Information Technology, The University of New South Wales, Canberra, ACT, 2600, Australia e-mail: [email protected] H. R. Pota e-mail: [email protected] I. R. Petersen e-mail: [email protected] L. Liu et al. (eds.), Applied Methods and Techniques for Mechatronic Systems, Lecture Notes in Control and Information Sciences 452, DOI: 10.1007/978-3-642-36385-6_19, © Springer-Verlag Berlin Heidelberg 2014

371

372

M. S. Rana et al.

Fig. 19.1 Schematic view of an AFM

The AFM’s attractive features are its fast and easy sample preparation, relatively low cost, and ability to operate in various environments. It provides a 3D profile of a surface on a nanoscale by measuring forces, such as the van der Waals, capillary, electrostatic, and magnetic, between a sharp probe ( Nc − 1. The F matrix with dimensions of (N P , n) and the Φ matrix with dimensions of (N P , Nc ) are: ⎤ ⎡ CA ⎢ C A2 ⎥ ⎥ ⎢ 3 ⎥ ⎢ F = ⎢ C A ⎥; ⎢ .. ⎥ ⎣ . ⎦ C ANp

⎡ ⎢ ⎢ ⎢ Φ=⎢ ⎢ ⎣

CB C AB C A2 B .. .

0 CB C AB .. .

··· ··· ··· .. .

··· ··· ··· .. .

0 0 0 .. .

⎤ ⎥ ⎥ ⎥ ⎥. ⎥ ⎦

C A N p −1 B C A N p −2 B · · · · · · C A N p −Nc B The control law is derived based on the minimization of the cost function J=

Np m=1

Q(y(k + m|k) − Rs (k + m))2 +

Nc

R(u(k + m − 1))2 ;

(19.12)

m=1

subject to the linear inequality constraints on the system inputs u min ≤ u(k + i − 1) ≤ u max , i = 1, . . . , Nc ; u min ≤ u(k + i − 1) ≤ u max , i = 1, . . . , Nc ;

(19.13a) (19.13b)

where Np is the prediction horizon, Nc is the control horizon, Q is the state weighting matrix, R is the control weighting, Rs is the reference signal, u min and u max are the low and high levels of the control action, respectively, and u min and u max are the low and high levels of the control increments, respectively. By considering the above equations, the constrained MPC problem can be expressed as a quadratic programming (QP) problem:

19 Advanced Control of Atomic Force Microscope for Faster Image Scanning

1 min( U T EU + U T f ); 2

381

(19.14)

s.t. MU ≤ γ ; where E = Φ T QΦ + R; f = Φ T Q F x(k + 1|k) − Φ T Q Rs ; M ∈ Rm c ×Nc and γ ∈ R Nc ×1 are computed using Eq. (19.13), m c is the number of constraints and Rs ∈ R N p ×1 is the reference signal.

19.4.2 Observer Design Kalman observer can be used as a state observer and noise filter [37]. The displacements of the PTSs are taken from the capacitive position sensors. Capacitive sensors are commonly used in nanopositioning systems because of their high-resolution measurement capability. However, it adds unwanted noise and disturbances to the output displacement which degrades an AFM’s scanning performance. To remove of which we design a Kalman state observer as a noise filter. The Kalman state observer estimates the states from the measured output. The Kalman observer dynamics is: x(k ˆ + 1) = (A − LC)x(k) ˆ + Bu(k) + L y(k); yˆ (k) = Cˆ x(k); ˆ

(19.15) (19.16)

where x(k) ˆ is the estimated states, yˆ (k) is the estimated state output, Cˆ is the identity matrix of dimension n ×n, and L is the observer gain which depends on the Gaussian white noise, process noise covariance, and measurement noise covariance.

19.5 Experimental Results 19.5.1 Creep Effect Compensation During the slow operation of the AFM in open-loop, the scanned image shows a creep effect as shown in Fig. 19.9a. Due to this effect the right side bottom edge of the vertical axis of the scanned image is rolled-off. In Fig. 19.9b, the creep effect is reduced significantly using the proposed controller.

382

M. S. Rana et al.

(a)

(b) 8

7

7

6

6

5

5 µm

µm

8

4

4

3

3

2

2

1

1

0

0 0

1

2

3

4 5 µm

6

7

8

0

1

2

3

4 µm

5

6

7

8

Fig. 19.9 Creep effect in the scanned image at open-loop (a) and compensation of creep effect using the proposed controller (b)

(a)

(b)

Fig. 19.10 Compensation of hysteresis effect at 31.25 Hz a using the AFM PI controller and b using the proposed controller

19.5.2 Hysteresis Effect Compensation The reduction of hysteresis effect at 31.25 Hz using the AFM PI controller and the proposed controller is shown in Fig. 19.10a and b, respectively. The image of the grating has a distinct curvature for the image taken by using the AFM PI controller. In the case of proposed controller, there have no hysteresis effect in the scanned image as shown in Fig. 19.10b.

19 Advanced Control of Atomic Force Microscope for Faster Image Scanning

0 −10 −20 1 10

10

2

10

3

5 0 −5 −10 1 10

Magnitude (dB)

(b)

10

Phase (rad)

Phase (rad)

Magnitude (dB)

(a)

383

Measured open−loop Meassured closed−loop 2

10

3

10

20 0

−20 −40 1 10

2

10

3

10

5 0 −5

Measured open−loop Measured closed−loop

−10 1 10

Frequency (Hz)

2

10

3

10

Frequency (Hz)

Fig. 19.11 Comparison of measured open-loop and closed-loop frequency responses of the X piezo (a) and Y piezo (b)

19.5.3 Vibration Effect Compensation The performance of the OMPC control scheme is evaluated by measuring its closedloop frequency responses. In Fig. 19.11a and b, the measured closed-loop frequency responses are plotted along with the open-loop frequency responses of the X and Y piezo displacements, respectively. By implementing this OMPC control scheme, we have achieved a reasonable damping in the closed-loop system for both axes, which in turn reduces the vibration significantly. Therefore, the closed-loop system has higher bandwidths for both the X and Y axes. The reduction of vibration effect by using the proposed controller is shown in Fig. 19.12. Figure 19.12 illustrates the results obtained by implementing the OMPC scheme which show significantly different images from those taken in the openloop condition. In the open-loop condition, the scanner displacement is quite poor because of the uncontrolled tube resonance and results in a sluggish image as shown in Fig. 19.12a, c, and e at 31.25, 62.50, and 125 Hz scanning speed and this problem is compensated by using the proposed controller as shown in Fig. 19.12b, d, and f. Fig. 19.12f demonstrates that the proposed controller can compensate the vibration effect at the higher scanning speed.

19.5.4 Cross-coupling Effect Compensation Figure 19.13a and b show comparisons of the open-loop and closed-loop crosscouplings for the 5.21 and 31.25 Hz input reference signals, respectively. They illustrate that there is significant improvement in cross-coupling in the closed-loop. To

384

M. S. Rana et al.

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 19.12 Compensation of vibration effect at 31.25, 62.50, and 125 Hz a,c, and e open-loop and b, d, and f using the proposed controller

19 Advanced Control of Atomic Force Microscope for Faster Image Scanning

(a)

(b)

0.02

−0.02 −0.04 −0.06 −0.08

0.02 0

dy(µm)

y

d (µm)

0

385

−0.02 −0.04 −0.06

0

0.1

0.2

0.3

−0.08

0.4

0

Time (sec)

0.02

0.04

0.06

0.08

Time (sec)

Fig. 19.13 Cross-coupling properties of the scanner at 5.21 and 31.25 Hz, respectively [closed-loop (red) and open-loop (blue)] 0.06

(b)

0.06

0.04

0.04

0.02

0.02

d x (µm)

dx (µm)

(a)

0 −0.02

0 −0.02 −0.04

−0.04

−0.06

−0.06 0

0.1

0.2

0.3

0.4

0

Time (sec)

0.02

0.04

0.06

Time (sec)

Fig. 19.14 Comparison of tracking performance of the proposed controller at 5.21 and 31.25 Hz, respectively, [reference signal (blue) and output signal (red)]

measure this cross-coupling, a reference triangular signal is applied to the X axis of the piezo and output taken from the Y position sensor.

19.5.5 Experimental Tracking Performance The experimental tracking performance of the proposed controller is shown in Fig. 19.14. In Fig. 19.14a and b the closed-loop tracking is presented at 5.21 and 31.25 Hz, respectively, which reflect the tracking performance of the proposed controller. The closed-loop tracking error are summarized in Table 19.1.

386 Table 19.1 RMS values of tracking error in closed-loop

M. S. Rana et al. Scan frequency (Hz)

RMS tracking error (nm)

5.21 31.25

0.369 0.689

19.6 Conclusion and Future Work The main contribution of this chapter was to compensate the nonlinear effect of the PTS by using the proposed controller for proper tracking and improved imaging at higher scanning speed of an AFM. The proposed OMPC controller was implemented to compensate the effect of creep, hysteresis, cross-coupling, and vibration of the PTS. The experimental results demonstrate that the tracking control performance is greatly improved in the high-speed application using the proposed controller. In this work, the plant considered was a single-input single-output (SISO) system. However, the author is interested in working with a multi-input multi-output (MIMO) system in future. Acknowledgments The authors would like to thank Mr. Shane Brandon, SEIT, UNSW, Canberra, Australia for his technical support during the experimental tests.

References 1. Yong YK, Ahmed B, Moheimani SOR (2010) Atomic force microscopy with a 12-electrode piezoelectric tube scanner. Rev Sci Instrum 81(3):033 701–10 2. Meyer E, Hug HJ, Bennewitz R (2004) Scanning probe microscopy. Springer, Berlin 3. Sarid D (1994) Scanning force microscopy: with applications to electric, magnetic and atomic forces. Oxford University Press, Oxford 4. Fleming AJ, Aphale SS, Moheimani SOR (2010) A new method for robust damping and tracking control of scanning probe microscope positioning stages. IEEE Trans Nanotechnol 9(4):438–448 5. Yong YK, Liu K, Moheimani SOR (2010) Reducing cross-coupling in a compliant XY nanopositioner for fast and accurate raster scanning. IEEE Trans Control Syst Technol 18(5):1172– 1179 6. Taylor ME (1993) Dynamics of piezoelectric tube scanners for scanning probe microscopy. Rev Sci Instrum 64(1):154–158 7. Adriaens H, De Koning W, Banning R (2000) Modeling piezoelectric actuators. IEEE/ASME Trans Mechatron 5(4):331–341 8. Rana MS, Pota HR, Petersen IR (2012) Improved control of atomic force microscope for high-speed image scanning. In: Australian control conference (AUCC). Sydney, pp 470–475 9. Bazaei A, Yong YK, Moheimani SOR, Sebastian A (2012) Tracking of triangular references using signal transformation for control of a novel AFM scanner stage. IEEE Trans Control Syst Technol 20(2):453–464 10. Jung H, Shim JY, Gweon D (2001) Tracking control of piezoelectric actuators. Nanotechnology 12(1):14–20 11. Croft D, Shedd G, Devasia S (2000) Creep, hysteresis, and vibration compensation for piezoactuators: atomic force microscopy application. Proc Am Control Conf 3:2123–2128

19 Advanced Control of Atomic Force Microscope for Faster Image Scanning

387

12. Jung H, Shim JY, Gweon D (2000) New open-loop actuating method of piezoelectric actuators for removing hysteresis and creep. Rev Sci Instrum 71(9):3436–3440 13. Croft D, Shedd G, Devasia S (2001) Creep, hysteresis, and vibration compensation for piezoactuators: atomic force microscopy application. J Dyn Syst Meas Control Trans ASME 123(1):35– 43 14. Leang K, Devasia S (2007) Feedback-linearized inverse feedforward for creep, hysteresis, and vibration compensation in afm piezoactuators. IEEE Trans Control Syst Technol 15(5):927– 935 15. Yi KA, Veillette RJ (2005) A charge controller for linear operation of a piezoelectric stack actuator. IEEE Trans Control Syst Technol 13(4):517–526 16. Chuang N, Petersen IR, Pota HR (2013) Robust H ∞ control in fast atomic force microscopy. Asian J Control 15(4):1–15 17. Cruz-Hernandez JM, Hayward V (2001) Phase control approach to hysteresis reduction. IEEE Trans Control Syst Technol 9(1):17–26 18. Mahmood IA, Moheimani SOR (2009) Making a commercial atomic force microscope more accurate and faster using positive position feedback control. Rev Sci Instrum 80(6):063 705063–705-8 19. Moheimani SOR, Vautier BJG (2005) Resonant control of structural vibration using chargedriven piezoelectric actuators. IEEE Trans Control Syst Technol 13(6):1021–1035 20. Aphale SS, Bhikkaji B, Moheimani SOR (2008) Minimizing scanning errors in piezoelectric stack-actuated nanopositioning platforms. IEEE Trans Nanotechnol 7(1):79–90 21. Pota HR, Moheimani SOR, Smith M (2002) Resonant controller for smart structures. Smart Mater Struct 11:1–8 22. Bhikkaji B, Ratnam M, Fleming AJ, Moheimani SOR (2007) High-performance control of piezoelectric tube scanners. IEEE Trans Control Syst Technol 15(5):853–866 23. Moheimani SOR, Vautier BJG, Bhikkaji B (2006) Experimental implementation of extended multivariable PPF control on an active structure. IEEE Trans Control Syst Technol 14(3):443–455 24. Kenton BJ, Fleming AJ, Leang KK (2011) Compact ultra-fast vertical nanopositioner for improving scanning probe microscope scan speed. Rev Sci Instrum 82(12):123 703-123–7038 25. Schitter G, Astrom K, DeMartini B, Thurner P, Turner K, Hansma P (2007) Design and modeling of a high-speed AFM-scanner. IEEE Trans Control Syst Technol 15(5):906–915 26. Kenton B, Leang K (2012) Design and control of a three-axis serial-kinematic high-bandwidth nanopositioner. IEEE/ASME Trans Mechatron 17(2):356–369 27. Fairbairn MW, Moheimani SOR, Fleming AJ (2011) Improving the scan rate and image quality in tapping mode atomic force microscopy with piezoelectric shunt control. In: Australian control conference (AUCC). pp 26–31 28. Grosswindhager S, Kozek M, Voigt A, Haffner L (2013) Fuzzy predictive control of district heating network. Int J Model Ident Control 19(2):161–170 29. Su B, Qi G, Van Wyk BJ (2012) Output feedback predictive control for uncertain non-linear switched systems. Int J Model Identif Control 17(3):195–205 30. Li D, Xi Y (2011) The synthesis of robust model predictive control with QP formulation. Int J Model Identif Control 13(1/2):1–8 31. Rana MS, Pota HR, Petersen IR (2012) Model predictive control of atomic force microscope for fast image scanning. In: 51st conference on decision and control (CDC). Hawaii, USA, pp 2477–2482 32. Devasia S, Eleftheriou E, Moheimani SOR (2007) A survey of control issues in nanopositioning. IEEE Trans Control Syst Technol 15(5):802–823 33. Privara S, Cigler J, Vana Z, Ferkl L (2012) Incorporation of system steady state properties into subspace identification algorithm. Int J Model Identif Control 16(2):159–167 34. Ljung L (2002) Prediction error estimation methods. Circ Syst Signal Process 21:11–21 35. Kabaila P (1983) On output-error methods for system identification. IEEE Trans Autom Control 28(1):12–23

388

M. S. Rana et al.

36. Wang L (2009) Model predictive control system design and implementation using MATLAB. Springer, London 37. Ray P, Panda G (2012) Harmonics estimation using KF-Adaline algorithm and elimination with hybrid active power filter in distorted power system signals. Int J Model Identif Control 16(2):149–158

Suggest Documents