1 Copyright 2011 by ASME

Proceedings of the ASME 2011 International Design Engineering Technical Conferences & Computers and
Information in Engineering Conference
IDETC/CIE 2011
August 28-31, 2011, Washington, DC, USA
DETC2011-MNS 47543

ON A CONTROL DESIGN TO AN AFM MICROCANTILEVER BEAM, OPERATING IN
A TAPPING-MODE, WITH IRREGULAR BEHAVIOR


Kleber dos Santos Rodrigues
UNESP: Univ Estadual Paulista, Department of
Mechanical Engineering, FEB, UNESP, Bauru,
So Paulo, Brazil
Jos Manoel Balthazar
UNESP: Univ Estadual Paulista
, Department of Statistics, Applied Mathematics
and Computation, DEMAC, Rio Claro, So Paulo,
Brazil

Angelo Marcelo Tusset
Department of Engineering Science, UTFPR
Ponta Grossa, Paran, Brazil
Bento Rodrigues Pontes Jnior
UNESP: Univ Estadual Paulista,Department of
Mechanical Engineering, FEB, UNESP, Bauru, So
Paulo, Brazil
ABSTRACT

In last decades, control of nonlinear dynamic systems became
an important and interesting problem studied by many authors,
what results the appearance of lots of works about this subject
in the scientific literature. In this paper, an Atomic Force
Microscope micro cantilever operating in tapping mode was
modeled, and its behavior was studied using bifurcation
diagrams, phase portraits, time history, Poincare maps and
Lyapunov exponents. Chaos was detected in an interval of time;
those phenomena undermine the achievement of accurate
images by the sample surface. In the mathematical model,
periodic and chaotic motion was obtained by changing
parameters. To control the chaotic behavior of the system were
implemented two control techniques. The SDRE control (State
Dependent Riccati Equation) and Time-delayed feedback
control. Simulation results show the feasibility of the both
methods, for chaos control of an AFM system.


INTRODUCTION

Atomic Force Microscope (AFM) is a powerful tool in scanning
probe microscope, its application includes manipulation of
carbon nanotubes, DNA, imaging and actuation in nano-
electronics, etc. (Rtzel et al., 2003). In AFM, a micro-
cantilever with a tip at its free end vibrates and sends a signal to
a photo detector, the acquired images come from this movement
and the scanning process starts.



Figure 1 A Schematic of tapping mode atomic force
microscope (Zhao and Danckowitz, 2006)

The micro-cantilever has three modes of operation: non-contact,
contact and tapping mode operation. Tapping mode is the most
common type of operation, where the contact between tip and
sample occur. To generate the images, the micro-cantilever
vibrates near from its resonance. By this fact, the micro
cantilever may exhibit chaotic behavior under contains
circumstances (Yau et al, 2009).According to the literature,
Sebastian et al. (2001) used a one-degree-of-freedom model,
with linear coefficients to describe attractive and repulsive
forces. Garcia and San Paulo (2000) used the spring-mass
equation to describe the micro-cantilever behavior and DMT,
and van der Waals theories to describe tip-sample interactions.
After years, Misra and Danckowits (2007) proposed a model
and used event-driven method to control the resonant behavior.
Chaotic behavior is common in AFM. When this type of
behavior occurs, the images are affected, what is not a desirable
outcome. Mathematical modeling is commonly used to
Proceedings of the ASME 2011 International Design Engineering Technical Conferences &
Computers and Information in Engineering Conference
IDETC/CIE 2011
August 28-31, 2011, Washington, DC, USA
DETC2011-4
2 Copyright 2011 by ASME
understand the behavior of AFM micro-cantilever. Numerical
simulation and analysis of the obtained images are essential to
project control methods and to stabilize the system. This paper
utilizes the mathematical model proposed by Rtzel et al (2003)
which consists of using Bernoulli equations to describe the
micro-cantilever dynamics behavior and Lennard Jones
potential to describe the tip-sample interactions. In modeling
process, a partial-differential equation occurs and is discretized
and nondimensionalized via Galerking method. By using
bifurcation diagrams, phase portrait, time history, Lyapunov
exponents and Poincare maps, was possible to observe the
chaotic and periodic behavior. Chaos is observed in a limited
interval of time, what hardly interferes on the analysis of
obtained images. Nonlinear systems can be identified as weakly
non-linear and strongly non-linear. In AFM, weakly
nonlinearities appear in non-contact mode and traditional
perturbation methods like the average method and multiple
scale method can be used (Nozaki et al, 2010; Aime et al.
1999; Boisgard et al. 1999; Nony et al. 2001; Couturier et al.
2002). Time delayed feedback control method was originally
proposed by Pyragas for stabilizing unstable periodic orbits
embedded in chaotic attractors. A key feature of the control
method is that control input depends on only the difference
signal between the current and past outputs of a nonlinear
system. The control method, therefore, can be implemented
without any identification of the model, system parameters, and
underlying dynamics from experimental data (Yamasue and
Hikihara, 2006), (Salarieh and Alasty, 2007).
Tapping mode operation is identified as a strong nonlinear
system (Rtzel, et al), then the use of those perturbation
methods become a complex job. First, because analytical
solutions for nonlinear system are unknown in most cases, and
second, the perturbation methods become more difficult to
implement (Guran, 1997). To stabilize the system, the State-
Dependent Riccati Equation (SDRE) control was used to make
the synchronization between the nonlinear systems, driven the
chaotic system (slave system), to the periodic system (master
system). At least, feedfoward control was used to maintain the
system controlled in periodic motion.
The paper is organized as follows: modeling of micro cantilever
deflection and the tip-sample interactions; discretization of the
systems equation; numerical simulations to find chaotic and
periodic behavior; synchronization and stabilization of the
system using SDRE control method; consider the stability of
time delayed feedback control.
USED NOMENCLATURE

L micro cantilever length
mass density
E Youngs modulus
I area moment of inertia
A cross section
* equilibrium gap
z distance between tip and sample
w*(L) deflection of micro cantilever tip
piezoelectric frequency
( ) q x piecewise manner
Diracs delta
R Probe tip radius
( )
n
x eingenfunction approximation
y micro cantilever tip displacement
Y vibration amplitude nondimensionalized
x state vector
A dependent state matrix
F nonlinear vector
U control term
u
f
feedfoward control
u feedback control
e error between the orbits (master and slave)
B two dimensional constant vector concerning
coupling between the control input and the
state variables.
P(x) Riccati state dependent solution
Q, R LQR matrices
C
11,
C
12,
d
1,
E
1
non-dimensional parameters acquired
by the Si-Si system under consideration.
non-dimensional amplitude of dither
piezoelectric actuator.

1
,
2
number densities of molecules

1
c ,
2
c interaction coefficients of intermolecular
pair Potentials


MATHEMATICAL MODEL

The Euler-Lagrange equation is used to describe the micro
cantilever dynamics, and Lennard-Jones potential is used to
represents the interactions between tip and sample. L is the
cantilever length, is mass density, E is the Youngs modulus,
and area moment of inertia I of the cross-section area A is
chosen in the analysis. The beam is clamped at x=0 and free at
x=L. The micro cantilever deflection is u(x,t*),
* *( ) Z w L = is the equilibrium gap between probe tip and
sample, and the piezoelectric is modeled by ( *) y t Ysen t =
(Rtzel et al, 2003).

3 Copyright 2011 by ASME

Figure 2 A schematic of the (a) initial, (b) intermediate,
and (c) current configurations associated with the micro
cantilever deformation (Qin-Quan, Chen, 2006)


MODELING OF THE CANTILEVER DEFLECTION.
In this case is assumed that cantilever behavior is similar with
the beam behavior, than EulerBernoulli equation is used to
describe the relationship between the beams deflection and the
applied load (Witmer, 1991-1992):

4 4
4 4
( , *) *( )
( )
u x t w x
EI EI q x
x x
| | | |
+ =
| |

\ \
(1)
The dynamic beam equation is the Euler-Lagrange beam
equation:

2 2 2
2 2
2 2
0
1 ( , *) 1 ( , *) *( )
( ) ( , *)
2 * 2
L
u x t u x t x
S A EI q x u x t dx
t x x

( | |
| | | | | |
( | = + +
| | |
| ( \
\ \
\

(2)
The first term represents the kinetic energy, the second one
represents the potential energy due to internal forces and the
third term represents the potential energy due to the external
load. The Euler-Lagrange equation is used to determine the
function that minimizes the functional S. For a dynamic Euler-
Bernoulli beam, the Euler-Lagrange equation is
2 2 2 2
2 2 2 2
( , *) *( ) ( , *)
( )
*
u x t x u x t
EI A q x
x x x t

| | | | | | | | | |
+ = + | | | | |
| |

\ \ \ \ \
(3)
Simplifying (3):
2 4 4
2 4 4
( , *) ( , *) *( )
( )
*
u x t u x t x
A EI q x
t x x

| | | | | | | |
+ + =
| | | |
|

\ \ \ \
(4)
Where, q(x) represents the distributed load over the micro-
cantilever lenght.
Boundary conditions:
Equation 4 is a fourth order partial differential equation and
boundary conditions are needed to find solution ( , ) u x t . In this
case, the load ( ) q x is represented in a piecewise manner, load
isnt a continuous function. Representation of point load as a
distribution using the Dirac function in this model results:
( )
2 4 4
2 4 4
( , *) ( , *) *( )
*
( )
d u x t d u x t d x
A EI
dt dx dx
q x x L

| | | | | | | |
+ + =
| | | |
|
\ \ \ \

With:
0
( , ) 0
x
u x t
=
= ,
0
( , *)
0
x
du x t
dx
=
= ,
2
2
( , *)
0
x L
d u x t
dx
=
=

The partial differential equation is:
( ) ( , *) ( '''' *''''( )) ( ) Au x t EI u x q x x L + + = &&

(5)
In the next section, ( ) q x is modeled as attraction/repulsion
force derived from Lennard-Jones interaction potential
(Salarieh and Alasty, 2007).

MODELING OF THE TIP SURFACE INTERACTIONS.
Lennard-Jones does not model the real contact mechanics
encountered in tapping mode AFM, but it represents a generic
tip-surface-interaction potential (Ruetzel, Lee, Raman, 2003).
According to Basso et al (2000) and Israelachvili (1991), the
representation of interaction forces between the probe tip of
radius R and the sample surface, with a gap z is:

1 2
, 7
6 1260
L J
A R A R
U
z z
=
, 1 2
, 8 2
( )
180 6
L J
L J
U AR A R
P z
z z z

= = +

Where A
1
and A
2
are the Hamaker constants, and represents
repulsive and attractive potentials respectively. The Hamaker
constants are given by
2
1 1 2 1
A c = and,
2
2 1 2 2
A c = .The
gap z is represented with * ( , *) sin( *) z u L t Y t = , than the
Lennard-Jones potential is:
1 2
, 8 2
180( * ( , *) sin( *)) 6( * ( , *) sin( *))
L J
A R A R
P
u L t Y t u L t Y t
= +

(6)
4 Copyright 2011 by ASME
Finally, using equation 6 in equation 5, and substitution of q(x)
by P
L,J
, is possible to obtain the governing equation of the
system:
( , *) ( ( , *) * ( )) Au x t EI u x t x + + = &&

1 2
8 2
.
180( * ( , *) sin *) 6( * ( , *) sin *)
A R A R
u L t Y t u L t Y t
| |
= +
|

\

2
. ( ) sin * x L A Y t +

(7)
DISCRETIZATION
Next consider the situation when the excitation frequency is
close to the lowest frequency of the micro-cantilever. Under
near-resonant forcing, and in the absence of additional internal
resonances, only one mode of the micro cantilever is assumed
to participate in the response (Qin-Quan and Chen, 2007):

1 1
( , *) ( ) ( *) u x t x q t = (8)

Where
1
( ) x is the first approximate eigenfunction.
Substitution of (8) into (7), multiplication of (7) by
1
( ) x ,
subsequent integration over the domain, and the introduction of
a modal damping consistent with the Q factors listed in table
table1 (appendix 2) yields the single-degree-of-freedom model:

2 11 12
1 1 1 8 2
(1 ) (1 )
C C
y d y y B E sin t
y sin t y sin t


= + + + +

&& &
(9)
A positive y is the micro cantilever tip displacement towards the
sample, nondimensionalized by the equilibrium gap between the
tip and the sample,
11
c ,
12
c ,
1
d and
1
E , are non-dimensional
parameters acquired by the Si-Si system under consideration.
In this case, is the vibration amplitude of the dither
piezoelectric actuator nondimensionalized by the equilibrium
gap width.
Considering the following substitutions: y x =
1
, y x & =
2
. The
differential equation (9) can also be represented in state space:
) sin(
)) sin( 1 (
)) sin( 1 (
1
2
8
1
6
1 12 11
1 2 1 1 2
2 1
t E
t x
t x C C
B x d x x
x x
+

+
+ + =
=

&
&

(10)
SYSTEM WITH CHAOTIC BEHAVIOR

Numerical simulations with (10) using software Matlab,
integrator ODE45, step length h=0.001, parameters;
1
0.01 d = ,
1
0.148967 B = ,
5
11
4.59118 10 C

= ,
1
1.57367 E =
12
0.149013 C = , with initial conditions (0) 0.2 x = and (0) 0 x = &
follows:
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
x
n

Figure 3 Bifurcation diagram


In figure 3, its possible to observe that with =0.9 the first
bifurcation is observed. With =0.9, the following simulations
are obtained: In figures 4, 5 and 6, with 900 1000 t , chaotic
motion appears in the system, with
5
1.2 10 t = , Lyapunov
exponents obtained the results:
1 2 3
0.010154, 0.020146, 0 = = = .With one of the
exponent is positive, the system is chaotic for
5
10 t .

900 910 920 930 940 950 960 970 980 990 1000
-3.5
-3
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
t
x
1

Figure 4 Time history of chaotic motion.

5 Copyright 2011 by ASME


Figure 5 Poincare map

-3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5
-4
-3
-2
-1
0
1
2
3
4
5
x
1
x
2

Figure 6 Phase portrait of chaotic motion
0 2 4 6 8 10 12
x 10
4
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
t
L
y
a
p
u
n
o
v

E
x
p
o
n
e
n
t
s

Figure 7 Positive Lyapunov exponents


SYSTEM WITH PERIODIC MOTION

Numerical simulations carried with Eq.(10), using software
Matlab, integrator ODE45, steplenght h=0.001, and parameters;
1
0.01 d = , 0.01 = ,
1
0.148967 B = ,
5
11
4.59118 10 C

= ,
12
0.149013 C = ,
1
50 E = , we will obtain through figures 8 and
9, with 900 1000 t , periodic motion.
:

900 910 920 930 940 950 960 970 980 990 1000
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
t
x
1

Figure 8 Time history of periodic motion

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
-1.5
-1
-0.5
0
0.5
1
1.5
x
1
x
2

Figure 9 phase portrait of periodic motion



SYNCHRONIZATION OF NONLINEAR SYSTEMS

In oscillators, synchronization can occurs spontaneously, be
induced or even forced by control techniques. Two or more
oscillators can be taken to artificially synchronize their
trajectories; such technique is justified in practical applications,
when the synchronization is necessary. In synchronization, the
desired trajectory of a slave system is the trajectory of the
master system, and the control signal (feedback) at a given
moment is the function of difference between the trajectories an
instant before. In equation 10, with parameters of periodic
motion being the master system, and chaotic motion as the slave
system, the objective is to take the chaotic motion of slave
6 Copyright 2011 by ASME
system to periodic motion of the master system. Next, figure 9
shows both systems in same phase portrait:

-3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5
-4
-3
-2
-1
0
1
2
3
4
5
x
1
x
2


periodic
chaotic
Figure 10 Phase portrait of periodic motion and chaotic
motion


STATE DEPENDENT RICATTI EQUATION (SDRE) AND
FEEDFOWARD CONTROL

Since the mid-90s, State-Dependent Riccati Equation (SDRE)
strategies have emerged as general design methods that provide
a systematic and effective means of designing nonlinear
controllers, observers, and filters. One of the reasons to choose
SDRE control method is the fact that many others control
techniques developed have very limited applicability because of
the strong conditions imposed on the system (Tayfun, 2008).
In this paper, two control principles are used : the feedfoward to
maintain the slave system in the desired periodic orbit of the
master system, and the state feedback control that is used to
stabilize the behavior of dynamic systems. Consider the
nonlinear system (10) in this form:

( ) x A x x F = + & (11)
Where
n
x R is the state vector,
n n
A R

is a dependent state
matrix, F is the nonlinear vector of terms not state dependent.
With control, the system (11) is:

( ) x A x x F U = + + & (12)
With:
f
U u u = + (13)
f
u is the feedfoward control:
f
u F = (14)
With
0
(0) x x = , and substitution of (13) in (12), the system
(12) is represented in matrix form:

( ) x A x x Bu = + & (15)
Where
n n
B R

is matrix control, and u is the linear state
feedback control. Find u using State-Dependent Riccati
Equation Method. Writing the dynamic system defined by (10):

1 2
2 1 1 2
x x
x x d x F
=
= +
&
&
(16)
Where:
6
2 11 12 1
1 1
8
1
(1 sin( ))
sin( )
(1 sin( ))
C C x t
F B E t
x t

+
= + +


(17)
By introducing (13) in (16), follows (Shawky et al., 2007):

u
x
x
d x
x
(

+
(

=
(

1
0
1
1 0
2
1
1 2
1
&
&
(18)

The system in (17), implies that the origin of the system is an
equilibrium point, condition that allows using the SDRE control
to calculate u.
Considering the system (18) as (15) the matrix A, and matrix B,
are represented by:

(

=
1
1
1 0
d
A ,
0
1
B
(
=
(

(19)
The feedback control u, is:

1
( ) ( ) ( ) ( )
T
u R x B x P x K x e

= = (20)

with:
*
( ) e x x = , x is the slave system signal and * x is the
master system signal, P(x) is the Riccati state dependent
solution:
1
( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) 0
T T
P x A x A x P x P x B x R x B x P x Q x

+ + =
(21)
By using:

(

=
3
3
10 0
0 10
Q , [ ]
-5
10 = R

The functional to be minimized by the SDRE control is given
by:
0
( ( ) ( ) )
T T
r r
J e Q e e u R u u dt

= +

(22)

The synchronization of these two systems can be observed in
figure 11, the error
*
( ) e x x = in figure 10 and the phase
portrait in figure 11, after application of control (13) in (10).

7 Copyright 2011 by ASME
0 10 20 30 40 50 60 70 80 90 100
-2
-1.5
-1
-0.5
0
0.5
1
t
x
1


master
slave

(a)
6.5012 6.5013 6.5014 6.5015 6.5016 6.5017 6.5018 6.5019 6.502
-1.25
-1.25
-1.2499
-1.2498
-1.2498
-1.2498
-1.2497
-1.2496
-1.2496
-1.2496
-1.2495
t
x
1


master
slave

(b)

Figure 11 a: Synchronization of master and slave systems,
b: zoom of the figure (a)

-2 -1.5 -1 -0.5 0 0.5 1
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
x
1
x
2


master
slave

(a)
0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
0.0384
0.0385
0.0386
0.0387
0.0388
0.0389
x
1
x
2


master
slave

(b)

Figure 12 (a): Phase portrait of two systems, master and
slave controlled, (b): zoom of the figure (a)

We observed from figures 10 and 11 that control application,
was efficient, driven the slave system to the master system, with
0003 . 0 max
*
= x x for 100 t .

TIME DELAYED FEEDBACK CONTROL
So far, this control method has been successfully applied to
various experimental systems including the AMF (Yamasue and
Hikihara, 2006). As originally suggested by Pyragas,
continuous control input u stabilizing a chaotic oscillation is
given by the difference between the current output and the past
one as follows (Yamasue and Hikihara, 2006), (Pyragas, 1992,
2002):

( ) ( ) [ ] ( ) ( ) [ ] { } t x t x g t x t x g K u
2 1 2 1
, , = (23)

where: is the time delay and k the feedback gain.

The ( ) ( ) [ ] t x t x g
2 1
, and ( ) ( ) [ ] t x t x
2 1
, imply a scalar output
signal measured at the current time t and the past time ( ) t ,
respectively. Since the control input (1) only depends on the
output signal. The time delay is adjusted to the period of a
target unstable periodic orbit we intend to stabilize in a chaotic
attractor, and the control input therefore converges to null after
the controlled system is stabilized to the target orbit (Yamasue
and Hikihara, 2006).Assuming that the velocity of oscillation is
measured as an output of the nonlinear system (10), the control
signal u is given as follows:

( ) ( ) [ ] t x t x K u
2 2
= (24)

8 Copyright 2011 by ASME
The system (10), with control (24) m can be represented in the
following way:

( ) ( ) [ ] t x t x K t E
t x
t x C C
B x d x x
x x
2 2 1
2
8
1
6
1 12 11
1 2 1 1 2
2 1
) sin(
)) sin( 1 (
)) sin( 1 (
+
+

+
+ + =
=

&
&
(25)

The time delay and feedback gain K are important control
parameters which substantially affecting of the control
performance. The time delay is adjusted to

2
2
=

= to
stabilize an orbit with the same frequency as the external force
oscillating the microcantilever, and the feedback gain is here
adjusted to be 2 . 0 = K , as proposed in Yamasue and Hikihara
(2006). Figures 13 and 14 shows the application of control (24).


-3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5
-5
-4
-3
-2
-1
0
1
2
3
4
5
x
1
x
2


without control
control

(a): t=[1:600]
-3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5
-5
-4
-3
-2
-1
0
1
2
3
4
5
x
1
x
2


without control
control

(b): t=[10:600]
Figure 13 - Chaotic attractor (black) and target unstable
periodic orbit (red)

0 10 20 30 40 50 60 70 80 90 100
-3.5
-3
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
t
x
1


without control
control

Figure 14 - Displacement of microcantilever

Considering the above results, we can see that time-delayed
feedback control is an efficient method for stabilizing unstable
periodic orbits of chaotic systems.

CONCLUSIONS

Control and stabilization of chaotic and irregular behavior of
micro-cantilevers in AFM are essential, because when this kind
of phenomena occurs, theres a lost of quality in image
detection process.
Modeling of a control method in tapping-mode is very difficult
due to strong nonlinearities. In this case, is possible to observe
in figures 9, 10 and 11 that the method of synchronization, with
the proposed control method was efficient with easy
implementation.
SDRE was effective, the implementation was very simple, and
the synchronization used, can be implemented in practical
situations without problems, with satisfactory results.
The numerical simulation shows that error can be minimized to
acceptable values by adjusting matrices Q and R of LQR
control, with a global stability of controlled system.
In this paper with the given Q and R matrices, the obtained
synchronization has an error 0003 . 0 max
*
= x x , after the
synchronization, feedfoward control kept the slave system into
master system condition, which ensured the efficiency of
methods, used to control the system.
In figures 13 and 14 we can observe the feasibility in applying
of the time delayed feedback control technique for both chaos
control of AFM system (10). Analyzing of the application of the
two controls, we can verify that The Possibility of application
using time delayed feedback depends on dynamical properties
of target periodic orbits under grazing bifurcation, and the
system takes time to synchronize with the 2 = t desired orbit,
while the application of the SDRE control enables its
application to stabilize any desired orbits, with much less time
needed to 2 = t as observed in the time delayed feedback
control.
9 Copyright 2011 by ASME
ACKNOWLEDGMENTS
Authors thank CNPq and Fapesp for financial support. K.S.
Rodrigues thanks CAPES for fellowship support.

REFERENCES

Aime, J. P., Boisgard, R., Nony, L. and Couturier, G., 1999,
Nonlinear dynamic behavior of an oscillating tipmicro lever
system and contrast at the atomic scale, Phys. Rev. Lett. Vol 82,
pp. 33883391.

Basso, M., Giarre, L., Dahleh, M. and Mezic, I., 2000,
Complex dynamics in a harmonically excited Lennard-Jones
Oscillator: micro-cantilever sample interaction in scanning
probe microscopes, J. Dynam. Syst. Meas. Control Vol 122,
pp. 240245.

Boisgard, R., Michel, D. and Aime, J. P., 1999, Hysteresis
generated by attractive interaction: oscillating behavior of a
vibrating tipmicro-cantilever system near a surface, Surf. Sci.
Vol 401, pp. 199205

Couturier, G., Nony, L., Boisgrad, R. and Aime, J.-P., 2002,
Stability of an oscillating tip in noncontact atomic force
microscopy: theoretical and numerical investigations, J. Appl.
Phys. Vol 91, pp. 25372543

Dankowicks H. and Zhao, X., 2006, Characterization of
intermittent contact in tapping mode force microscopy, J.
Comput. Nonlinear Dynam. - Vol 1, pp. 109.

Fenili, A. and Balthazar, J.M., 2011, The rigid-flexible
nonlinear robotic manipulator: Modeling and control ,
Communications in Nonlinear Science and Numerical
Simulation Vol 16, pp. 2332-2341.

Garcia, R. and San Paulo, A., 2000, Dynamics of a vibrating
tip near or in intermittent contact with a surface,
PhysicalReview B Vol 61, pp. 20.

Israelachvili, J., 1991, Intermolecular and surface forces, 2nd
ed. London

Misra, S., Dankowicks, H. and Paul, M.R., 2007, Event-driven
feedback tracking and control of tapping-mode atomic force
microscopy, Proc. R. Soc. A ,vol. 464 no. 2096, pp. 2113-2133

Nony, L., Boisgrad, R. and Aime, J.-P., 2001, Stability
criterions of an oscillating tip-cantilever system in dynamic
force Microscopy, Eur. Phys. J. B Vol 24, pp. 221229.

Nozaki, R., Pontes Jr, B.R, Tusset, A. M. and Balthazar, J.M.,
2010, Optimal linear control to an atomic force microscope:
(AFM) problem with chaotic behavior,Dincon10- 9
th
Brazilian conference on dynamics, control and their
Applications (http://www.sbmac.org.br/dincon/ )

Nozaki, R., 2010, Nonlinear dynamics, chaos and control in
Atomic Force Microscope, |Master Degree dissertation, mech
Enginering College (FEB), UNESP, SP, Brazil.

Pyragas, K., 1992, Continuous control of chaos by self-
controlling feedback, Phys. Lett. A, 170, pp.421-428.

Pyragas, K., 2002, Analytical properties and optimization of
time-delayed feedback control, Physical Review E 66, 026207,
pp. 1-9.

Qin-Quan H. and Chen L., 2007,Bifurcation and chaos in
atomic force microscope. Elsevier, Chaos, Solutions and
potentials, Proc R Soc London A ; 459:192548.

Rutzel, S., Lee, S.I. and Raman, A., 2003, Nonlinear dynamics
of atomic-force-microscope probes driven in Lenard-Jones,
Fractals Vol 33, pp. 711715.

Salarieh, H, A. Alasty, 2007, Control of chaos in atomic force
microscopes using delayed feedback based on entropy
minimization, Communications in Nonlinear Science and
Numerical Simulation Vol 14, pp. 637- 644

Sebastian, A., Salapaka, M. V., Chen, D. J., and Cleveland J. P.,
2001, Harmonic and power balance tools for tapping mode
atomic force microscope, Journal of Applied Physics Vol 89,
pp. 64736480.

Shawky, A. M., Ordys, A. W., Petropoulakis, L. and Grimble,
M. J., 2007, Position control of flexible manipulator using
non-linear with state-dependent Riccati equation. Proc.
IMechE, 221 Part I: J. Systems and Control Engineering, pp.
475- 486.

Witmer, E.A., 1992. "Elementary Bernoulli-Euler Beam
Theory". MIT Unified Engineering Course Notes. pp. 114-164

Yamasue, K. and Hikihara, T., 2006, Control of
microcantilevers in dynamic force microscopy using time
delayed feedback, Review of Scientific Instruments 77,
053703, pp. 1-6.

Yau , H., Wang C., Chou, Y., 2009. Design of Sliding Mode
Controller for AFM system by Backstepping Design Proc.
IEEE International Conference on Networking, Sensing and
Control, pp. 840-845


10 Copyright 2011 by ASME

ANNEX A

DISCRETIZATION PROCESS
Consider the partial differential equation;
1 2
8 2
2
( , *) ( ( , *) * ( ))
.
180( * ( , *) sin *) 6( * ( , *) sin *)
. ( ) sin *.
Au x t EI u x t x
A R A R
u L t Y t u L t Y t
x L A Y t



+ + =
| |
= +
|

\
+
&&



Now, consider the situation when the excitation frequency is
close to the lowest frequency of the micro-cantilever. Under
near-resonant forcing, and in the absence of additional internal
resonances, only one mode of the micro cantilever is assumed
to participate in the response (Qin-Quan and Chen, 2007):

1 1
( , ) ( ) ( ) u x t x q t =

Where
1
( ) x is the first approximate eigenfunction.
Substitution of (8) into (7), multiplication of (7) by
1
( ) x ,
subsequent integration over the domain, and the introduction of
a modal damping consistent with the Q factors listed in table 1,
annex B yields the single-degree-of-freedom model (Rtzel,
Lee and Raman, 2003):

2 11 12
1 1 1 8 2
(1 ) (1 )
C C
y d y y B E sin t
y sin t y sin t


= + + + +

&& &

Where

1
( *)
*
x t
y

= ,
1 1 1
( *) ( ) ( *) x t L q t = , * *( ) Z w L = ,
0
* t t = ,
1 1
(1 ) B =
,

*
Y

= ,
3
3EI
k
L
= ,
0

= ,
1
1
2
0 1
0
L
c
d
A dx
=

,
*
Z

= ,
1
11 1 9
180 ( *)
A R
C
k
= ,
2
12 1 3
6 ( *)
A R
C
k
= ,
1 1
1 0
0
2
1
0
L
L
EI dx
A dx

,
2
1
1
2 2
0 1
0
( )
L
k L
A dx

1 1
0
1
2
1
0
( )
L
L
L dx
E
dx

and a positive y s the microcantilever tip displacement
towards the sample, non-dimensionalized by the equilibrium
gap between the tip and the sample.
1
is the vibration
amplitude of the dither piezoelectric actuator non-
dimensionalized by the equilibrium gap width, and
0
is a
reference frequency for subsequent non-dimensionalization.

Note that B
1
+ C
1l
+ C
12
= 0, so that, on the
substitution of y = u = 0 in (8), equation (8) is identically
satisfied. The dotted quantities now represent derivatives
with respect to rescaled time t.

ANNEX B

TABLE 1 - PROPERTIES AND DIMENSIONS OF THE
CANTILEVERS (RTZEL, LEE AND RAMAN, 2006).



Descriptions symbol SiSi(111) Sipolystyrene
length L 449 m 154 m
width b 46 m 13.7 m
thickness h 1.7 m 6.9 m
tip radius R 150 nm 20 nm
material density 2 330 kg m
3
2 330 kg m
3

static stiffness k 0.11 N m
1
40 N m
1

elastic modulus E 176 GPa 130GPa
1st resonance f
1
11.804 kHz 350.0 kHz
Q factor (air) Q 100 100
Hamaker (att.) A
2
1.86510
19
J 1.15010
19
J
Hamaker (rep.) A
1
1.359610
-70
Jm
6
0.83810
-70
J