Institute of Physics Publishing Journal of Physics: Conference Series 34 (2006) 500–505

doi:10.1088/1742-6596/34/1/082 International MEMS Conference 2006

A Comparison Simulation of Fixed-fixed Type MEMS

Switches

G Rezazadeh1, H Sadeghian, E Malekpour

Mech. Eng. Dept., Urmia University, Urmia, Iran

G.rezazadeh@mail.urmia.ac.ir, st_h.sadeghian@mail.urmia.ac.ir,
Ehsan.malekpour@gmail.com

Abstract. In the present work pull-in voltage of fixed-fixed end type MEMS switches with
variative electrostatic area has been calculated using a distributed model and applying a full
nonlinear finite difference discretizing method. The governing nonlinear differential equation
has been derived using of the variational principle for multi domain electromechanical coupled
system. The numerical results of the beam with variative electrostatic area with the results of
Coupled-Domain Finite Element method have been compared and very good agreement has
been achieved.

1. Introduction
In recent years, the field of Microelectromechanical systems (MEMS) has grown rapidly and has
entered into many defense and communication applications. Electrostatically actuated membranes or
beams have been widely used and studied by the MEMS community. Electrostatically actuated
devices form a broad class of MEMS devices due to their simplicity as they require few mechanical
components, and small voltage levels for actuation. Electrostatically actuated microbeams are used in
many MEMS devices such as capacitive MEMS switches and resonant sensors. Manufacturing and
design of these devices are, to some extent, in a more mature stage than some other MEMS devices.
MEMS microbeams are liable to an instability known as the pull-in instability. When the applied
voltage is increased beyond a critical value, called pull-in voltage, stable equilibrium positions of the
microbeam cease to exist. Pull-in instability, greatly limits the stable range of operation of
microbeams. In most cases it would be highly desirable to delay the onset of pull-in for better
performance of the device. The pull-in voltage depends on the interaction of the electrostatic forces
generated by the applied voltage, and the structural stiffness of the microbeam. Typically, such
membranes or beams are used in switches [1-4] and microrelays [5]. One-dimensional (1D), quasi-
two-dimensional (quasi-2D), and three-dimensional (3D) simulations with various accuracy level have
been used to help understand and characterize these membranes, beams, and devices [1-7]. Previous
work addressing pull-in instability has taken both electrical and mechanical approaches; existing
analytical method has been applied by balancing elastic restoring force with electrical attractive force
to construct structural equation and estimate pull-in solution [8]. A more general method for
evaluating the pull-in parameters is based on 3D coupled-domains finite-elements (FEM)/boundary-
1
All authors with Mech. Eng. Dept. of Urmia University, Urmia, Iran (Phone: ++98-914-1451407-8475; fax:
+98-441-2777022 ;)

© 2006 IOP Publishing Ltd 500

 501

elements (BEM) algorithms [9]-[14]. Electrostatically actuated devices form a broad class of MEMS
devices due to their simplicity as they require few mechanical components, and small voltage levels
for actuation. The structural elements that are used in MEMS devices are typically simple elements
like beams, plates, and membranes. Electrostatically actuated microbeams are used in many MEMS
devices such as capacitive MEMS switches and resonant sensors. Manufacturing and design of these
devices are, to some extent, in a more mature stage than some other MEMS devices.
MEMS microbeams are liable to an instability known as the pull-in instability. When the applied
voltage is increased beyond a critical value, called pull-in voltage, stable equilibrium positions of the
microbeam cease to exist. Pull-in instability, greatly limits the stable range of operation of
microbeams. In most cases it would be highly desirable to delay the onset of pull-in for better
performance of the device. The pull-in voltage depends on the interaction of the electrostatic forces
generated by the applied voltage, and the structural stiffness of the microbeam. In this paper the
nonlinear differential equation has been solved by finite difference and finite element methods
(ANSYS software). The results of two methods have been compared and a good agreement has been
achieved. Three methods are used for decreasing the pull-in voltage in MEMS devices such as MEMS
switches [15], [16]. In the present work the pull-in voltage of a beam with variative electrostatic area
has been investigated. Also the fringing effects of electrical field and residual stress have been studied.

2. Model Description
The schematic view of a fixed-fixed end type MEMS switch with variative electrostatic area as
described in this paper is shown in fig. 1. The device consists of a thin film metal membrane, called
the “beam”, suspended over a dielectric film deposited on top of the center conductor and fixed at both
ends to the ground conductors. When a voltage is applied between the beam and substrate, the
attractive electrostatic pressure pulls the beam down towards the dielectric film. The dielectric film
serves to prevent stiction between the beam and substrate, and yet provides a low impedance path
between the two contacts. The switch can be fabricated using surface micromachining techniques,
electroplating techniques and dry releasing technique compatible to the (millimetre wave integrated
circuit) MMIC fabrication processes. In theoretical analysis, the parameters of MEMS switches
involve materials constants (such as elastic modulus, Poisson’s ratio and so on), geometrical
dimensions of the fixed-fixed beam, moment of inertia of the cross-section, the gap between the
movable and the fixed ground plates, and electrostatic area.

Figure 1. Schematic view of a fixed-fixed end type MEMS switch with variative electrostatic area

3. Nonlinear Electromechanical Coupled Equation

502

When a driving voltage is applied between the electrodes, the electrostatic pressure deflects the beam.
The mechanical bending strain energy U m of the beam and the electrical co-energy U e* stored between
the upper and lower electrode of the beam are given by:
~ 2 ~ 2
L
E§ du · L
EI § d 2 u ·
Um ³ ³A 2 ¨  z ¸
dx ¹
dA dx ³ 2 ¨¨ dx 2 ¸¸ dx (1)
0 © 0 © ¹
1 L H 0 wV 2
U e* ³ dx (2)
2 0 g  u ( x) 
where z is the coordinate in the load direction with origin in the centroid of the cross section. H 0 , w,
and g represent the permittivity of air, the width of the beam, and the initial gap between the upper and
lower electrodes. u(x) is the deflection function and V is the applied voltage between the
movable/ground plates on the fixed substrate. The total potential energy U of the system can be
expressed as
U Um Ue (3)
The variation of total energy is zero at the equilibrium position.
GU G U m  G U e GU m  GU e* 0 (4)

~ 2
EI § d 2 u ·
L L
H wV 2
GU G ³ ¨¨ 2 ¸¸ dx  G ³ 0 dx
0 2 © dx ¹ 0 2 g  u ( x ) 
(5)
L­
° d2 § ~ d 2u · H 0 wV 2 ½°
¨ EI ¸
³ ® dx 2 ¨ dx 2 ¸ 2g  u ( x) 2 ¾°G u dx 0
0 °̄ © ¹ ¿
The įu is an arbitary function, so to satisfying equation (5):

d 2 § ~ d 2u · H 0 wV 2
¨ EI ¸
dx 2 ¨© dx 2 ¸¹ 2 g  u ( x) 2
(6)
~
E is dependent on the beam width w and film thickness t [14]. A beam is considered wide when w•5t.
~
Wide beams exhibit plane-strain conditions, and therefore, E becomes the plate modulus E /(1 Q 2 ) ,
~
where E is the young’s modulus. A beam is considered narrow when w<5t. In this case, E simply
becomes the young’s modulus, E. I is the effective moment of inertia of the cross-section and
is wt 3 12 which is wide relative to thickness and width. A uniform magnetic field cannot drop
abruptly to zero at an edge. In actual situation, there is always a ‘‘fringing field’’ existing, and a more
realistic Situation including ‘‘fringing field’’ modification is executive. By considering the fringing
effect of electrical field, the governing equation would be as follow;
d2 § ~ d 2u · H 0 wV 2 § g  u ( x) ·
¨ EI ¸ ¨1  0.65 ¸
dx 2 ¨ dx 2 ¸ 2 g  u ( x)  2
© w ¹
© ¹ (7)
Residual stress can be expressed as
Tr Vwt , V V 0 (1  Q ) (8)
where V 0 is the biaxial residual stress [9]. The nonlinear differential equation would be;
~ d 4u d 2u H 0 wV 2 § g  u ( x) ·
EIˆ 4  Tr 2 ¨1  0.65 ¸ (9)
2 g  u ( x ) 
2
dx dx © w ¹
 503

Three methods have been used to decrease the actuation voltage [15], [16]. One of them is using of the
variative electrostatic area .In our nonlinear distributed model as it is shown in Fig. 1 the electrostatic
pressure is applied only to a certain part of the beam and in this part the section is different from the
other section. In proposed model, we study the effect of variative electrostatic area on the pull-in
voltage.
The governing differential equation for beam with varying section, incorporating the first order
fringing-field correction and residual stress effect as:
~ d 4u d 2u H 0V 2 w2 § g  u ( x) ·
EIˆ2 4  Tr 2  2
¨¨1  0.65 0 ¸¸
dx dx 2[ g 0  u ( x)] © w2 ¹ (10)
u H ( x  L1 / 2)  H ( x  L1 / 2) 
And in other section the differential equation is as bellow
~ d 4u d 2u
EIˆ1 4  Tr 2 0
dx dx (11)
ˆ ˆ
where H(x) is the Heaviside function. I 1 , I 2 the effective moment of inertia of the cross-section which
w1t 3 w t3
are wide relative to their thickness and with, and equals to and 2 . w1 and w2 is the width of
12 12
each part of the beam.

4. Results
Due to the nonlinearity in the electrostatic pressure, an analytical solution is impractical to obtain and
a numerical solution is sought. Numerical finite difference method (FDM) and finite element method
(ANSYS software) are been used for evaluating the pull-in voltage from equation (10), (11).
Parameters used in this article are: Young’s modulus E is 169 GPa, the Poisson’s ratio u is .3, the
length of the beam is 800 µm, the width of the fixed-fixed beam is 50 µm, the thickness t is 14.4 µm
and the initial gap g is 1 µm, the Permittivity of air is 8.8541878u10-12 (F/m), the residual stress about
4 MPa and the width of electrostatic area is 100 µm and the length of electrostatic area is 100µm. As
shown in table 1, by considering the variative electrostatic area, the pull-in voltage is decreased and is
lower than pull-in voltage of simple beam model (the width of simple beam is equal to width of
variative electrostatic area). In Fig. 2 the beam deflection modeled by ANSYS is shown. The
deflection of pull-in phenomenon is demonstrated in fig. 3, when the voltage has equalled to its pull-in
value, the system will be unstable and the center gap suddenly decreased till zero. The center gap
versus voltage by FEM is shown in fig. 4.

Table 1. Pull-in voltage and center gap in different model

Type Model Pull-In Voltage (V) Center Gap(µm)
Lumped Model 35.34 0.6667
Simple beam(FDM) 40.42 0.6023
Simple beam(ANSYS) 40.90 0.6033
Varying section beam(FDM) 37.50 0.6030
Varying section beam( ANSYS) 37.80 0.6016
504

Figure 2. Deflection of fixed-fixed beam Figure 3. Deflection of fixed-fixed beam

Simulated by ANSYS Calculated by FDM

Figure 4. The beam deflection modeled by ANSYS

5. Conclusion
The design and simulation of fixed-fixed end types MEMS switches with simple and varying section
was implemented. The FDM and FEM (ANSYS software) was used to evaluate the pull-in
phenomenon. The results of FDM are closely agreed with ANSYS results. The fringing effect of
electrical field and residual stress effect is accounted in model. It is shown that by considering the
variative electrostatic area, the pull-in voltage is decreased, and it will be useful for design of complex
MEMS devices that will be worked on specific voltage.

References
[1] Goldsmith C, Randall J, Eshelman S, Lin T H, Denniston D, Chen S and Norvell B 1996
Characteristics of micromachined switches at microwave frequencies: Proceedings of the
Digest IEEE MTT-S International Microwave Symposium (San Francisco, USA) Vol. 2, ,
pp. 1141-1144.
[2] Chan E K, Garikipati K and Dutton R W 1999 Characterization of contact electromechanics
through capacitance-voltage measurements and simulations (J. Microelectromech) Syst.
8(2) 208-217.
[3] Yao Z J, Chen S, Eshelman S, Denniston D and Goldsmith C 1999 Micromachined low –loss
 505

microwave switches (J. Microelectromech) Syst. 8(2) 129-134.

[4] Osterberg P M and Senturia S D 1997 M-test: a test chip for MEMS material property
measurement using electrostatically actuated test structures (J. Microelectromech) Syst. 6(2)
107-118.
[5] Gretillat M A, Gretillat F and Rooij N f De 1999 Micromechanical relay with electrostatic
actuation and metallic contacts (J. Micromech) Microeng. 9 324-331.
[6] Osterberg P, Yie H, Cai X, White J and Senturia S 1994 Self-consistent simulation and
modeling of electrostatically deformed diaphragms: Proceeding of the Micro Electro
Mechanical Systems Workshop, MEMS’94, Oiso, Japan, pp. 28-32.
[7] Hung E S and Senturia S D 1999 Generating efficient dynamical models for
Microelectromechanical systems from a few finite-element simulation runs (J.
Microelectromech) Syst. 8 (3) 280-289.
[8] Degani O, Socher E and Lipson A et al 1998 (IEEE J. Microelectomech. Sys., V.7, No.4, 327-
379).
[9] Gupta R K 1997 Electrostatic pull-in test structure design for in situ mechanical property
measurements of Microelectromechanical systems (MEMS), Ph.D. Thesis, Supervised by
Senturia S D (MIT: Massachooset Institute of technology, USA) june.
[10] MEMCAD. http://www.conventor.com
[11] Intellisuite. http://www.intellisense.com
[12] Fischer M, Giousouf M, Schaepperle J, Eichner D, Weinmann M, Munch W Von and Assmus F
1998 Electrostatically deflectable polysilicon micromirrors-dynamic behaviour and
comparison with results from FEM modeling with ANSYS (Sens. Actuat) A 67 89-95.
[13] Chan E K, Garikipati K and Dutton R W 1999 Characterization of contact electromechanics
through capacitance-voltage measurements and simulation (J.MEMS) 8 (2) 208-217.
[14] Chan E K and Dutton R W 2000 Electrostatic micromechanical actuator with extended range of
travel (J.MEMS) 9 (3) 321-328.
[15] Mukherjee T, Fedder G K, and White J 2000 Emerging simulation approaches for
micromachined devices IEEE Trans. Comput.-Aided Design Integr.Circuits Syst., vol. 19,
pp. 1572–1589, Dec.
[16] CoventorWave for Cadence. Coventor Inc., Corporate development center, (Cambridge: UK),
MA. [Online]. Available: http://www.conventor.com
[17] Ark JY, Kim G H, Chung K W and Bu J U 2001 Monolithically integrated micromachined RF
MEMS capacitive switches. (Sensors and Actuators) A (89): 88–94