Você está na página 1de 42

Author's Accepted Manuscript

Studying the Effects of small scale and Casimir


force on the nonlinear pull-in instability and
Vibrations of FGM microswitches under elec-
trostatic Actuation
R. Gholami, R. Ansari, H. Rouhi

www.elsevier.com/locate/nlm

PII: S0020-7462(15)00145-6
DOI: http://dx.doi.org/10.1016/j.ijnonlinmec.2015.08.007
Reference: NLM2532

To appear in: International Journal of Non-Linear Mechanics

Received date: 2 February 2015


Revised date: 16 May 2015
Accepted date: 11 August 2015

Cite this article as: R. Gholami, R. Ansari, H. Rouhi, Studying the Effects of
small scale and Casimir force on the nonlinear pull-in instability and
Vibrations of FGM microswitches under electrostatic Actuation, International
Journal of Non-Linear Mechanics, http://dx.doi.org/10.1016/j.ijnonlinmec.2015.08.007

This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal
pertain.
Studying the Effects of Small Scale and Casimir Force on the Nonlinear Pull-In Instability
and Vibrations of FGM Microswitches under Electrostatic Actuation

R. Gholami*,a, R. Ansarib, H. Rouhib


a
Department of Mechanical Engineering, Lahijan Branch, Islamic Azad University, P.O. Box 1616, Lahijan, Iran
b
Department of Mechanical Engineering, University of Guilan, Rasht, Iran, P.O. Box 41635-3756

Abstract
The size-dependent nonlinear pull-in instability and free vibration of electrostatically actuated
microswitches with the consideration of Casimir force effect are studied using a numerical
solution approach. To this end, a non-classical nonlinear beam model is developed based on
Mindlin’s strain gradient elasticity and the Timoshenko beam theory. The geometric nonlinearity
is taken into account according to the von Kármán hypothesis. Also, the microswitches are
assumed to be made of functionally graded materials (FGMs). To obtain the size-dependent
governing equations and boundary conditions, the virtual work principle is applied. The
presented equations can be simply reduced to those on the basis of modified versions of strain
gradient and couple stress theories (MSGT and MCST) as well as the classical elasticity theory.
For solving the problem, the generalized differential quadrature (GDQ) method and the pseudo
arc-length continuation technique are employed. In the numerical results, the influences of
different parameters such as length scale parameter, Casimir force, material gradient index and
geometrical properties on the pull-in instability and free vibration of actuated microswitches are
examined.

Keywords: Electrostatically actuated microswitch; Nonlinear pull-in instability; Mindlin’s strain


gradient theory; Timoshenko beam theory; Casimir force; Functionally graded material

1. Introduction
Electrostatically actuated micro-systems such as micro-switches are widely used in different
applications. Microswitches can be used in energy conversion mechanisms, automotive

*
Corresponding author. Tel. /fax: +98 1412222906.

E-mail address: gholami_r@liau.ac.ir (R. Gholami).

1
applications, information processing, gyroscopes, toys, accessories, microsensors such as
accelerometers, etc. [1-5]. The pull-in instability is one of the important phenomena related to
such systems. The pull-in instability happens when the input voltage exceeds a critical value,
called pull-in voltage, and as a result the microstructure suddenly collapses. This phenomenon
was first observed by Nathanson et al. [6] and Taylor [7] experimentally.
In recent years, the mathematical modeling of pull-in instability of electrostatically actuated
microstructures has attracted a lot of attention from the research community. There are a number
of works on this topic in the literature based on the classical elasticity theory (e.g. [8-18]).
Several experimental studies have shown that the mechanical behavior of microstructures is size-
dependent [19-23]. Hence, for better understanding the mechanical responses of structures at
microscale, size-dependent elasticity theories such as the strain gradient theory (SGT), the couple
stress theory (CST) and the nonlocal theory (NT) should be used. In CST [24, 25], two material
length scale parameters (in addition to two classical material constants) are employed for
capturing the size effects. Also, in SGT [26, 27], the first and second derivatives of strain tensor
effective on the strain energy density are taken into account. As compared to CST, SGT includes
some higher-order stresses as well as classical and couple stresses. Moreover, in the modified
versions of CST and SGT (MCST & MSGT) [28, 29], one and three length scale parameters are
respectively used to consider the size effects. Also, the theory of nonlocal continuum mechanics
initiated by Eringen [30, 31] has been widely accepted in analyzing small-scale structures for
which the influence of small length scales can no longer be neglected. The application of
aforementioned higher-order theories to the pull-in problem of microstructures has been reported
by many researchers. The reader is referred to [32-42], [43-49] and [50-52] for SGT-, CST- and
NT-based papers, respectively.
Using the strain gradient theory, Wang et al. [33] studied the pull-in phenomenon in
electrostatically actuated microbeam-based micro-electro-mechanical systems (MEMS).
Rahaeifard and Ahmadian [36] analyzed the pull-in instability of electrostatically actuated
micro-cantilevers using SGT. Moreover, Seyyed Fakhrabadi et al. [37] provided an analytical
solution to investigate the nonlinear behaviors of electrostatically actuated carbon nanotubes
(CNTs) based on strain gradient theory. Sedighi and his associates [38] studied the static and
dynamic pull-in instabilities of cantilever nanoactuators using SGT. They revealed the
considerable effect of small scale on the pull-in behavior of nanoactuators with very small sizes.

2
Sedighi [39] also investigated the size-dependent dynamic pull-in instability of vibrating
electrically actuated microbeams using SGT. The dynamic behavior of micro- and nanobeams
subjected to electrostatic actuation was studied by Maani Miandoab et al. [40] within the
framework of SGT. They also evaluated Young's modulus and length scale parameters of poly
silicon nanobeams through fitting the static pull-in voltages predicted based on SGT to the
experimental data [41]. On the basis of MSGT, Tadi Beni [42] studied the pull-in instability of
electrostatically actuated nano-bridges and nano-cantilevers.
Based on MCST, Rahaeifard et al. [44] investigated the static pull-in instability of
microcantilevers. Rokni et al. [48] developed analytical solutions for size-dependent static pull-
in instability of electrostatic micro-actuators. Based upon MCST, Askari and Tahani [49] studied
the dynamic pull-in of micro-electro-mechanical beam subjected to shock acceleration pulse.
In the context of nonlocal elasticity theory, Mousavi et al. [50] studied the pull-in instability
of nano-switches under electrostatic and intermolecular forces. Seyyed Fakhrabadi et al. [51]
investigated the nonlocal pull-in instability of electrostatically actuated CNTs. Moreover, size-
dependent nonlinear static and dynamic responses of capacitive nanoactuators subjected to a DC
voltage were analyzed using NT by Najar et al. [52].
Recently, the effects of Casimir force and functionally graded material properties on the
instability and vibrational behaviors of microbeams have been considered in some research
works. For example, Jia et al. [53] considered the Casimir force effect on the forced vibration of
electrically actuated micro-switches near resonance region. Based on MCST, Abbasnejad et al.
[54] investigated the pull-in instability of electrostatically actuated microbeams made of FGMs.
Sedighi and his co-workers [55] took the influences of Casimir and van der Waals attractions
into account for the analysis of dynamic pull-in instability of vibrating nano-actuators. The
Casimir force effect was also considered in the analysis of electrostatically actuated
nanoelectromechanical systems (NEMS) [56]. Furthermore, Sedighi et al. [57] studied the
dynamic instability of FGM nano-bridges with considering Casimir attraction and electric filed
actuation.
It should be mentioned that Witvrouw and Mehta [58] proposed a new non-homogeneous
functionally graded polycrystalline-SiGe (poly-SiGe) layer in MEMS so as to obtain the desired
electrical and mechanical properties. Moreover, Gromova et al. [59] presented an experimental
investigation on the characterization and strain gradient optimization of PECVD poly-SiGe

3
layers for MEMS applications. Hasanyan et al. [60] analyzed the pull-in instability of
functionally graded MEMS as a result of heat generated by an electric current.
In the present article, within the framework of most general form of strain gradient theory
(Mindlin’s SGT [26]) that encompasses MSGT and MCST, the nonlinear pull-in instability and
vibrational behavior of functionally graded Timoshenko microswitches subjected to the
electrostatic actuation are investigated with considering the effect of Casimir force. The virtual
work principle is utilized to derive the size-dependent governing equations and boundary
conditions which are then discretized based upon the generalized differential quadrature (GDQ)
method. The resulting nonlinear algebraic parameterized equations are finally solved by means
of the pseudo arc-length continuation algorithm.

2. Derivation of Governing Equations and Boundary Conditions


According to Mindlin’s strain gradient theory [26], the stored strain energy Π s in a continuum
made of linear elastic materials occupying region ∀ with negligible deflection is given by

1
2 ∫∀
Πs = ( σ : ε + τ M ξ ) dV (1)

where σ , τ , ε and ξ denote stress, double stress tensors, strain and third-order strain gradient
tensors, respectively. The strain tensors are expressed as

1
ε = ∇u + ( ∇u ) + ( ∇u )( ∇u )  ,
T T
(2a)
2 

1
ξ = ∇ε = ∇ ∇u + ( ∇u ) + ( ∇u )( ∇u )  .
T T
  (2b)
2

in which ui denotes the components of the displacement vector u . The stresses are also defined
as [26]

(3a
σ ij = σ ji = λε kk δ ij + 2 µε ij
)

(3b
1
τ ijk = τ ikj = a1 (ξ kppδ ij + 2ξ ppiδ jk + ξ jppδ ik ) + 2a2ξippδ jk + a3 (ξ ppk δ ij + ξ ppjδ ik ) + 2a4ξijk + a5 (ξ jki + ξ kji
2 )

where δ ij is the Kronecker delta. The physical description of τ ijk can be found in [26].

4
Also, λ and µ are Lamé’s parameters defined as

Eν E
λ= , µ= . (4)
(1 + ν) (1 − 2 ν) 2 (1 + ν)

As shown in Fig. 1, an FGM microbeam made from a mixture of ceramic and metal with the
length L and thickness h subjected to the transverse force q is considered. A coordinate system
( x, y , z ) is selected on the central axis of the beam, where x , y and z axes are taken along the
length, width and height directions of beam, respectively. It is assumed that the microbeam at
bottom surface ( z = − h / 2) and top surface ( z = h / 2) is ceramic-rich and metal-rich,
respectively. The volume fractions of ceramic (denoted by subscript c ) and metal (denoted by
subscript m ) phases are defined by the following power-law functions
k k
1 z 1 z (5)
Vc ( z ) =  +  , Vm ( z ) = 1 −  + 
 2 h   2 h 
where k stands for the material gradient index. Effective Young’s modulus (E) and Poisson’s
ratio (ν) can be written as

E ( z ) = E c Vc + E m Vm , ν ( z ) = ν cVc +ν mVm , ρ ( z ) = ρcVc + ρmVm . (6)


Based on the Timoshenko beam theory, the displacements of an arbitrary point in the
microswitch along x, y and z axes are introduced as
u1 = U ( x ) − zΨ ( x ) , u2 = 0, u3 = W ( x ) (7)
in which U ( x ) , W ( x ) and Ψ ( x ) are the axial displacement of center of cross sections,
lateral deflection, and rotation angle of the cross section with respect to the vertical direction,
respectively. For a Timoshenko microbeam under tiny slopes after deflection and possible finite
transverse deflection, the nonlinear strain-displacement relations can be approximated by the von
Kármán hypothesis as follows
2 2
∂u1 1  ∂W  ∂U ∂Ψ 1  ∂W  1  ∂W 
ε xx = +   = −z +   , ε xz =  −Ψ (8)
∂x 2  ∂x  ∂x ∂x 2  ∂x  2  ∂x 

By substituting Eqs. (7) and (8) into Eq. (2b) one can arrive at the following nonzero constituents
of ξ

∂ 2U ∂W ∂ 2W ∂ 2Ψ 1  ∂ 2W ∂Ψ  ∂Ψ (9)
ξ xxx = 2
+ 2
− z 2
, ξ xxz = ξ xzx =  2 −  , ξ zxx = − .
∂x ∂x ∂x ∂x 2  ∂x ∂x  ∂x
The main components of symmetric section of the stress tensor are expressed by means of the
kinematic parameters as

5
 ∂U 1  ∂W 2 ∂Ψ   ∂W 
σ xx = ( λ + 2 µ )  +   −z , σ xz = µ  − Ψ . (10)
 ∂x 2  ∂x  ∂x   ∂x 

Inserting Eq. (9) into (3b) leads to the following nonzero constituents of the higher-order
stresses
 ∂ 2U ∂W ∂ 2W ∂2Ψ 
τ xxx = 2 ( a1 + a2 + a3 + a4 + a5 )  2
+ − z ,
 ∂x ∂x ∂x 2 ∂x 2 
1 ∂ 2W 1 ∂Ψ (11)
( a3 + 2a4 + a5 ) 2 − ( a1 + a3 + 2a4 + 3a5 ) ,
τ xxz = τ xzx =
2 ∂x 2 ∂x
2
1 ∂W 1 ∂Ψ
τ zxx = ( a1 + 2a5 ) 2 − ( a1 + 4a2 + 4a4 + 2a5 ) .
2 ∂x 2 ∂x
Symbols Π C and Π NC are used to denote the strain energies corresponding to the classical and
strain gradient theories, respectively. By introducing Eqs. (8-11) into (1), the strain energy due to
the variations of classical and higher-order stresses with respect to the initial configuration can
be obtained as

1   ∂U 1  ∂W  2   
L
1 ∂Ψ  ∂W
Π C = ∫σ : ε dV = ∫  N xx  +    − M xx + Qx − Ψ   dx , (12a)
2V 2 0   ∂x 2  ∂x   ∂x  ∂x  

1
2 V∫
N NC = τ M ξdV

(12b)
  ∂ U ∂W ∂ W   ∂ 2W ∂Ψ  ∂Ψ 
L 2 2
∂2Ψ
= ∫ Txxx  2 + 2 
− M xxx 2
+ Txxz  2
−  − Tzxx  dx
0   ∂x ∂x ∂x  ∂x  ∂x ∂x  ∂x 

where the normal resultant force N xx , shear force Q x , bending moment M xx , and other higher-
order resultant forces and higher-order moments ( Txxx , M xxx , Txxz , Tzxx ), in relation with higher-
order stresses acting on a section are given as

{N xx , Q, Txxx , Txxz , Tzxx } = ∫ {σ xx , κ s σ xz , τ xxx , τ xxz , τ zxx } dA,


A

(13)
{M xx , M xxx } = ∫ {σ xx , τ xxx } zdA
A

The work done by the transverse force q is written as

6
L
Π w = ∫qWdx (14)
0

where the external transverse force can be the electrostatic force acting on the deformable
microswitch and Casimir forces. With neglecting the fringing field correction, the electrostatic
force acting on the deformable microswitch is approximated by the parallel capacitor formula as
[33]
ε0V2
Fe ( x ) = 2
, (15)
2 (G − W )

in which ε 0 is the permittivity in vacuum and V is the applied voltage. Moreover, the Casimir
force is calculated by the following relation [61]

π 2 C
Fc ( x ) = 4
, (16)
240 ( G − W )

where  = 1.055 × 10−34 Js is Planck’s constant divided by 2π ; and C = 3 × 108 ms −1 is the speed
of light.

The kinetic energy of microswitch, Π T , is also obtained as

1
L
  ∂U ∂Ψ   ∂W  
2 2

ΠT = ∫ ∫ ρ   −z  +   dAdx
2 0 A   ∂t ∂t   ∂t  
(17)
1   ∂U   ∂W  
L 2 2 2
∂U ∂Ψ  ∂Ψ 
= ∫  I0   − 2 I1 + I2   + I0    dx
2 0   ∂t  ∂t ∂t  ∂t   ∂t  

where ρ and A are the mass density and cross-sectional area of microswitch, respectively. The
inertia terms appeared in Eq. (17) are defined by
h2

{ I 0 , I1 , I 2 } = ∫ ρ ( z ) { 1, z, z 2 } dz.
−h 2 (18)
By using the virtual work principle, taking the variations of U , W and Ψ , integrating by parts
and setting the coefficients of δ U , δ W and δΨ equal to zero; the governing equations and
associated boundary conditions are derived as

∂N xx ∂ 2 Txxx ∂ 2U ∂2Ψ
− = I 0 2 − I1 2 , (19a)
∂x ∂x 2 ∂t ∂t

7
∂  ∂Txxx  ∂W  ∂Q x ∂ 2Txxz ∂ 2W
N
 xx −  + − + q = I ,
∂x  ∂x  ∂x  ∂x (19b)
0
∂x 2 ∂t 2

∂ (M xx + Txxz + Tzxx ) ∂ 2 M xxx ∂ 2Ψ ∂ 2U


Qx − + = I 2 2 − I1 2 . (19c)
∂x ∂x 2 ∂t ∂t

 ∂Txxx 
 N xx −  =0 or δU x = 0, L = 0, (19d)
 ∂x  x = 0,L

 ∂Txxx  dW ∂Txxz 
 N xx − ∂x  dx + Q x + ∂x  = 0 or δW x = 0,L = 0, (19e)
   x = 0,L

 ∂M xxx 
 M xx + Txxz + Tzxx −  =0 or δΨ x = 0,L = 0, (19f)
 ∂x  x = 0,L

 ∂U 
Txxx x = 0,L
= 0 or δ   x = 0,L = 0, (19g)
 ∂x 

 ∂W   ∂W 
 Txxx + Txxz  = 0 or δ   x = 0, L = 0, (19h)
 ∂x  x = 0,L  ∂x 

 ∂Ψ 
M xxx x = 0,L
=0 or δ   x = 0, L = 0. (19i)
 ∂x 

3. Dimensionless Form of Governing Equations and Boundary Conditions


By defining the following parameters
h /2 h /2
1
{ A11 , B11 , D11} = ∫ ( λ + 2µ ) {1, z, z 2 } dz, A22 = − ∫ ( a1 + a3 + 2a4 + 3a5 ) dz,
− h /2
2 − h /2

h /2

{ A33 , B33 , D33 } = 2 ∫ ( a1 + a2 + a3 + a4 + a5 ) {1, z, z 2 } dz, (20)


− h /2

h /2 h /2
1
A44 = ∫
2 − h /2
( a5 + 2a4 + a3 ) dz,{ A55 , B55 , D55 } = ∫ µ 1, z, z 2 dz, { }
− h /2

8
h /2
1
2 − h∫/2
A66 = ( 2a1 + 5a5 + a3 + 6a4 + 4a2 ) dz.

the following dimensionless parameters are introduced

x L (U ,W ) G t A110
ζ = , η = , ( u, w ) = ,ψ = Ψ , g 0 = ,τ = , (21a)
L h h h L I 00

 A11 A55 B11 B55 D11 D55 


( a11 , a55 , b11 , b55 ,, d11 , d55 ) =  , , , , 2
, 2 
,
 A110 A110 A110 h A110 h A110 h A110 h 

 A22 A A   A B D 
( a22 , a44 , a66 ) =  2
, 44 2 , 66 2  , ( a33 , b33 , d33 ) =  33 2 , 33 3 , 33 4  , (21b)
 A110 h A110 h A110 h   A110 h A110 h A110 h 

   I0 I1 I2  qL2 2 ε 0 V 2 L2 4 π 2  CL2
 0 1 2 
I , I , I = , , 2 
, q0 = , α = ,β =
   I 00 I 00 h I 00 h  hA110 h 3 A110 240h5 A110

in which A110 and I 00 are the values of A11 and I 0 for a homogeneous microswitch made of
metal. Using the dimensionless parameters, the dimensionless governing equations together with
the dimensionless boundary conditions can be written as

(2
 ∂ u 1 ∂w ∂ w 
2
∂ ψ a33  ∂ u 1 ∂w ∂ w 3 ∂ w ∂ w  b33 ∂ ψ
2 2 4 4
∂u 2 3
∂ 2ψ
2a
4 ¯ 2 ¯
a11  2 +  − b −  + +  + = I 0 − I 1
 ∂ζ η ∂ζ ∂ζ 2  11 ∂ζ 2 η 2  ∂ζ 4 η ∂ζ ∂ζ 4 η ∂x 2 ∂x3  η 2 ∂ζ 4 ∂τ 2 ∂τ 2
)
(2
 a44 ∂ 4 w a22 ∂ 3ψ a11  ∂u ∂ 2 w ∂ 2u ∂w 3 ∂ 2 w  ∂w  
2
 ∂2w ∂ψ
k s a55  2 − η − 2 − +  + + 
 ∂ζ ∂ζ  η ∂ζ
4
η ∂ζ 3 η  ∂ζ ∂ζ 2 ∂ζ 2 ∂ζ 2η ∂ζ 2  ∂ζ   2b
)

b11  ∂ψ ∂ 2 w ∂w ∂ 2ψ  a33  ∂ 4u ∂w ∂ 3u ∂ 2 w  b33  ∂w ∂ 4ψ ∂ 2 w ∂ 3ψ  a33  ∂w


−  + −  + +  + − 4
η  ∂ζ ∂ζ 2 ∂ζ ∂ζ 2  η 3  ∂ζ 4 ∂ζ ∂ζ 3 ∂ζ 2  η 3  ∂ζ ∂ζ 4 ∂ζ 2 ∂ζ 3  η 4  ∂ζ

(2
 ∂w   ∂ u 1 ∂w ∂ w 
2 2
∂ ψ d33 ∂ ψ a22 ∂ w 2
∂ ψ b33  ∂ u 2c1 ∂w ∂ 4 w 3 ∂
4 3 2 4
k s a55η  − ηψ  − b11  2 +  + d − + + a +  + +
 ∂ζ   ∂ζ η ∂ζ ∂ζ 2  11 ∂ζ 2 η 2 ∂ζ 4 η ∂ζ 3 66 ∂ζ 2 η 2  ∂ζ 4 ) η ∂ζ ∂ζ 4 η ∂

(2
2d

9
  ∂u 1  ∂w 2  )
∂ψ a33  ∂ 3u 1 ∂w ∂ 3 w 1  ∂ 2 w   b33 ∂ 3ψ 
2

δ u ζ = 0,1 = 0 or a11  +    − b11 − 2  3+ 3


+  2  + 2  =0
  ∂ζ 2η  ∂ζ  
 ∂ζ η 

∂ζ η ∂ζ ∂ζ η  ∂ζ   η ∂ζ 3 

 ∂u    ∂ 2u 1 ∂w ∂ 2 w  ∂ 2ψ  (2
δ  ζ = 0, L = 0 or  33  2
a +  − b  = 0,
 ∂ζ    ∂ζ η ∂ζ ∂ζ 2  33 ∂ζ 2  2e
ζ = 0,1
)
(2
 a11  ∂u ∂w 1  ∂w   b11 ∂ψ ∂w a44 ∂ 3 w
3
 ∂w
δ w ζ = 0,1 = 0 or ks a55  − ηψ  +  +   − + 2 3 2f)
 ∂ζ  η  ∂ζ ∂ζ 2η  ∂ζ   η ∂ζ ∂ζ η ∂ζ

a ∂ 2ψ a33  ∂ 3u ∂w 1  ∂w  2 ∂ 3 w 1 ∂w  ∂ 2 w 2  b ∂ 3ψ ∂w
+ 22 −  3 +   +  2  + 3
33
= 0,
η ∂ζ 2 η 3  ∂ζ ∂ζ η  ∂ζ  ∂ζ
3
η ∂ζ  ∂ζ   η ∂ζ 3 ∂ζ
 ζ = 0,1

(2
 ∂w   ∂2w ∂ψ a33  ∂ 2u 1 ∂w ∂ 2 w  ∂w b33 ∂ 2ψ ∂w  2g
δ  = 0 or  44
a 2
+ a22η +  2+ 2 
− 2  = 0,
 ∂ζ  ζ = 0,1  ∂ζ ∂ζ η  ∂ζ η ∂ζ ∂ζ  ∂ζ η ∂ζ ∂ζ 
ζ = 0,1 )

∂ψ b33  ∂ 3u 1 ∂w ∂ 3 w 1  ∂ 2 w  
 ∂u 1  ∂w  2  2 (2
δψ ζ =0,1 = 0 or b11  +    − d11 −  + +   
∂ζ η 2  ∂ζ 3 η ∂ζ ∂ζ 3 η  ∂ζ 2   2h
 ∂ζ 2η  ∂ζ    
)
d33 ∂ 3ψ a22 ∂ 2 w ∂ψ
+ 2 3
− 2
− a66 =0
η ∂ζ η ∂ζ ∂ζ ζ = 0,1

 ∂ψ   ∂ 2 u 1 ∂w ∂ 2 w  ∂ 2ψ (2
δ  = 0 or b33  2
+ −
 33d = 0.
 ∂ζ  ζ = 0,1  ∂ζ η ∂ζ ∂ζ 2  ∂ζ 2 ζ = 0,1
2i)

The simply-supported boundary condition is expressed as


 ∂ 2u 1 ∂w ∂ 2 w  ∂ 2ψ
u = a33  2 + −b = 0,
 ∂ζ η ∂ζ ∂ζ 2  33 ∂ζ 2

∂2w ∂ψ a33  ∂ 2u 1 ∂w ∂ 2 w  ∂w b33 ∂ 2ψ ∂w (23)


w = a44 + a22η +  +  − = 0,
∂ζ 2 ∂ζ η  ∂ζ 2 η ∂ζ ∂ζ 2  ∂ζ η ∂ζ 2 ∂ζ

10
 ∂u 1  ∂w  2  ∂ψ b33  ∂ 3u 1 ∂w ∂ 3 w 1  ∂ 2 w   d33 ∂ 3ψ a22 ∂ 2 w
2
∂ψ  ∂ 2u 1
b11  +    − d11 −  + +    + − − a66 = b33  +
 ∂ζ 2η  ∂ζ   ∂ζ η 2  ∂ζ 3 η ∂ζ ∂ζ 3 η  ∂ζ 2   η 2 ∂ζ 3 η ∂ζ 2 ∂ζ  ∂ζ 2 η
 

Also, for clamped boundary condition one has


∂u ∂u ∂ψ
u= =w= =ψ = = 0. (24)
∂ζ ∂ζ ∂ζ

It should be noted that by defining the following parameters


 4   1 1   4 1 
a1 = µ  l22 − l12  , a2 = µ  l02 − l12 − l22  , a3 = − µ  l12 + l22  ,
 15   15 2   15 2 
(25)
1  2 
a4 = µ  l12 + l22  , a5 = µ  l12 − l22  .
3  3 

the obtained equations so far can be reduced to those based on MSGT [29]. In Eq. (25), l0 , l1 and
l2 show independent material length scale parameters, which are associated with the dilatation
gradients, deviatoric stretch gradients and symmetric rotation gradients, respectively. In addition,
when l0 = l1 = 0 , the model of MCST is retrieved [28]. The relations of classical elasticity theory
(CT) are also recovered when all material length scale parameters become equal to zero.

4. Solution
The GDQ method [62] is used to discretize the governing equations and boundary conditions.
The shifted Chebyshev–Gauss–Lobatto grid points are employed to generate grid points in ξ i
direction

1 i −1  (26)
ξi =  1 − cos π , i = 1, 2, … , N
2 N −1 

Consider column vectors U, W and Ψ as follows

T T
U = {U1 , U 2 , … , U N } , W = {W1 , W2 , … , WN } , Ψ = {Ψ1 , Ψ 2 , … , Ψ N } ,
T (27)

with N entries equal to the number of grid points Ui = u ( ξ i ) , Wi = w ( ξ i ) and Ψi = ψ ( ξ i ) .


Now, one can write the discretized form of governing equations Eqs. (22a-22c) as
Ä (28)
M X+ KX + N ( X ) + Q ( W, α , β ) = 0

in which X , K and N ( X) are field variables vector, stiffness matrix, geometric stiffness matrix,
mass matrix and nonlinear part vector, respectively, which are given as

11
 0  (29a)
U     Nu ( X ) 
   α2 β4   
X = W  , Q ( W, α , β ) =  2
+ 4
, N ( X ) =  N w ( X ) ,
 
Ψ   2 ( g 0 − W ) 240 ( g 0 − W )   N ( X) 
 ψ 
 0 

 ¹ ( 0) ¹
0  (29b)
 − I 1 Dξ 0 I 2 Dξ( ) 
 ¹ 
− I 1 Dξ( )
0
M= 0 0 
¹ ( 0)
¹ 
− I 3 Dξ( ) 
0
 I 2 Dξ 0
 

 (2) a 33 (4) b33 (4) 


 a11Dξ − 2 Dξ 0 − b11D(2)
ξ + Dξ 
 η η2 
 a 44 (4) a 22 (3) 
K= 0 k s a 55 D(2)
ξ − Dξ − k s a 55 ηD(1)
ξ − Dξ 
 η2 η 
 (2) b33 (4) a 22 (3) d 
 − b11Dξ + 2 Dξ k s a 55 ηD(1)
ξ + Dξ − k s a 55 η2 D(0) (2)
ξ + (a 66 + d11 ) Dξ −
33
2
D(4)
ξ 
 η η η 

(29c)
1 (1) a 1 3 
N u (X) = a11
η
( 2
)(
η η
1
) 4

η
2
(
Dξ W o Dξ( ) W − 332  Dξ( ) W o Dξ( ) W + Dξ( ) W o Dξ( ) W 
3


)( ) ( )( )
a11  (1) 3 
Nw ( X ) =
η 
( ( 2)
)(
( 2) ( 3)
 Dξ U o Dξ W + Dξ U o Dξ W +

) ( )(
Dξ( ) W o Dξ( ) W o Dξ( ) W 
2 1 1


) ( )( )( )


b11
η
(( D Ψ ) o ( D W ) + ( D W ) o ( D Ψ )) − aη (( D U ) o ( D W ) + ( D U ) o ( D W ))
(1)
ξ
( 2)
ξ
(1)
ξ
( 2)
ξ
33
3
( 4)
ξ
(1)
ξ
( 3)
ξ
( 2)
ξ

a 33 b

η 4 (( )( )( ) ( )( )(
4 Dξ(1) W o Dξ( 2) W o Dξ( 3) W + Dξ( 4) W o Dξ(1) W o Dξ(1) W + Dξ( 2) W o Dξ( 2) W o Dξ( 2) W + 333
η
) ( )( )( )) Dξ1 W

1 (1) b 1 3 
N ψ (X) = − b11
η
( 2
)(
η η
1
)4

η
2
(
Dξ W o Dξ( ) W + 332  Dξ( ) W o Dξ( ) W + Dξ( ) W o Dξ( ) W 
3


)( ) ( )( )
(29d)
where o signals the Hadamard product. Furthermore, the weighting coefficients of the first
derivative and higher-order derivatives are calculated by

12
 Ix , r = 0
 P (ξ i )
 , , i ≠ j and i, j = 1, …, N and r = 1
 ( ξi − ξ j ) P (ξ j )

 D(ξr )  = Wij(r) =   (1) ( r −1) Wij( r −1) 
  ij (30)
r Wij Wii − ξ − ξ  , i ≠ j and i, j = 1, …, N and r = 2,3, … N − 1
  i j

 N
 − ∑ Wij(r) , i = j and i, j = 1, …, N and r = 1, 2,3, … N − 1
 j=1; j≠ i

N
where I x is a N × N identity matrix and P ( x i ) = ∏ (ξ
j=1; j≠ i
i − ξj) .

A similar procedure can be used to discretize the boundary conditions. For instance, the
discretized form of clamped boundary condition becomes
U = W = Ψ = D(1) (1) (1)
ξ U = Dξ W = Dξ Ψ = 0 at ξ = 0,1 (31)
By inserting the boundary conditions for boundary grid points in the discretized form of
domain equations, a set of nonlinear equations will be achieved as follows
F ( W, U, V, α , β ) = 0 ; F : R3 N + R 2 → R3N (32)
In order to solve this set of parameterized nonlinear equations, one parameter such as β is fixed
and another one such as α is considered as an active parameter; afterwards, the set of
F : R 3N +1 → R 3N can be solved using the pseudo arc-length method [63] which enables us to
follow the unstable solution branches as well as the stable ones.

5. Results and Discussion


In this section, selected numerical results are presented on the size-dependent pull-in instability
and free vibration of microswitches under simply supported-simply supported (SS-SS), simply
supported-clamped (SS-C) and clamped-clamped (C-C) boundary conditions.
First, in order to check the validity and accuracy of the present analysis, the linear and nonlinear
pull-in voltages of a clamped-clamped silicon microbeam obtained from the present analysis are
compared with those of the numerical solutions given in [64] and experimental data reported in
[65], as shown in Table 1. The parameters used in the verification are as L = 1mm ,
b(width) = 30 µ m , h = 2.4 µ m , g 0 = 10.1 µ m and l0 = l1 = l2 = l = 0 . It is seen that there is a
reasonable agreement between the present results and those of [64] and [65].

Also, in order to verify the convergence of the present numerical method, Table 2 is given in
which the linear and nonlinear pull-in voltages of homogenous clamped-clamped microbeams

13
with different lengths corresponding to various total numbers of the grid points are presented.
Furthermore, in this table, two other comparisons are made between the present results and those
obtained using a numerical solution procedure given in [66] and the experimental data of [67]. It
is assumed that b = 100 µ m, g 0 = 1.18 µ m and h =1.5 µ m . The converging trend of results is
clearly observed. It is also seen that there is an excellent agreement between the present results
and those of [66] and [67]. According to Table 2, N = 31 is utilized for all of the following
numerical calculations.

As stated, the FGM microswitch is assumed to be made of a ceramic phase and a metal one.
In this paper, the FGM is considered to be a mixture of aluminum (Al) and ceramic (SiC) and all
of the following numerical results are obtained for an FGM microswitch with the following
material properties
E m = 70 GPa , ν m = 0.3 and ρ m = 2702 kg / m 3 for Al

E c = 427 GPa , ν c = 0.17 and ρ c = 3100 kg / m 3 for SiC

Moreover, as the values of material length scale parameters l0 , l1 and l2 for the FGMs have not
been reported in the literature up to now, these parameters are approximately assumed to be
equal to l0 = l1 = l2 = 15 µ m .

Figs. 2 and 3 show the variation of nondimensional center gap (Non. Dim. Center Gap) of FGM
microswitches versus nondimensional applied voltage (Non. Dim. Voltage) and nondimensional
Casimir parameter (Non. Dim. Casimir Parameter) based on MSGT. These figures are given for
different values of thickness-to-length scale parameter ratios (nondimensional length scale
parameter, h / l ). The solid circles indicate the beginning of pull-in instability. The upper branch
is stable whereas the lower branch is unstable. The pull-in instability happens where these
branches of solution collide at a certain applied voltage/Casimir force at which the slope of curve
approaches infinity. The size effect on the pull-in behavior of microswitches is clearly seen from
the figures. It is observed that as the nondimensional length scale parameter decreases, the pull-
in voltage and pull-in Casimir parameter increase. A considerable difference is seen between the
prediction of MSGT and that of its classical counterpart when the thickness of microswitch
becomes comparable with the length scale parameter. However, two sets of results tend to
converge at sufficiently large values of h / l .

It should be remarked that the intermolecular force can play a significant role in the
performance of microswitches when the only reason of pull-in phenomenon is the Casimir force
[13]. For a given initial gap, as the magnitude of Casimir force becomes equal or larger than the

14
critical Casimir force, the pull-in may occur even without an applied voltage. Hence, the Casimir
force may considerably reduce the pull-in voltage particularly for small initial gaps.
Figs. 4 and 5 provide a comparison between the results of different elasticity theories and
models on the pull-in voltage and pull-in Casimir parameter of microswitches. The results of
these figures are generated using MSGT, MCST and CT based on both linear and nonlinear
Timoshenko beam models. One can find that with increasing the nondimensional length scale
parameter, the curves of all models tend to converge. This converging trend can reveal the
reliability of the present solution approach. The discrepancy between the curves is significant at
small scales: the nondimensional critical voltages and Casimir parameters of MSGT are larger
than those of MCST, and the latter ones are also larger than those of CT. In addition, Figs. 4 and
5 illustrate that the influence of geometric nonlinearity becomes more prominent as the
nondimensional length scale parameter increases, especially for SS-SS microswitches.
The effects of material gradient index on the pull-in characteristics of microswitches are
highlighted in Figs. 6 and 7. It is observed that as the material gradient index increases, the pull-
in voltage and pull-in Casimir parameter become smaller.
The effect of length-to-thickness ratio ( L / h ) is also shown in Figs. 8 and 9. These figures
indicate the importance of modeling short microswitches using the Timoshenko beam theory.
According to the figures, as expected, the pull-in instability is postponed as the microswitch
becomes shorter.

The effect of nondimensional initial center gap on the nondimensional critical pull-in voltage can
be studied in Fig. 10. The results of this figure are generated corresponding to both linear and
nonlinear models based upon MSGT for different values of h / l . The classical results are also
given for the comparison purpose. It is seen that the pull-in voltage of both linear and nonlinear
models gets larger as the initial center gap increases. Furthermore, it is observed that the pull-in
voltage is underestimated when the mid-plane stretching effect is ignored. Fig. 10 shows that the
difference between the linear and nonlinear models becomes more significant when the initial
center gap increases. The nonlinear effects are also dependent on the scale, so that the difference
between the solid (nonlinear) and dashed (linear) curves is intensified by increasing the
nondimensional length scale parameter. Also, one can find that the nonlinear effects are more
pronounced for microswitches subject to softer end conditions.

Fig. 11 shows the variation of nondimensional natural frequency against dimensioness applied
voltage corresponding to various values of h / l . The results of this figure are obtained using the
nonlinear model based on Eq. (28). First, it is observed that for all types of boundary conditions,
when the applied voltage is so large and accordingly the dynamic pull-in occurs, the
nondimensional frequency quickly tends to zero. Another finding is that for microswitches with

15
C-C and SS-C end conditions, when the applied voltage increases, the frequency monotonously
decreases. This trend is similar for all nondimensional length scale parametes. But the free
vibration behavior of microswitches with SS-SS boundary conditions is different from that
observed for microswitches under other types of boundary conditions. In this case, the frequency
monotonously decreases with increasing the voltage for small values of h / l ; but the variation of
frequency with the voltage at large values of h / l is not monotonous. This maye be attributed to
the effect of nonlinear strain hardening for SS-SS boundary conditions at larger values of h / l .

Moreover, Fig. 12 is given for the linear frequencies. By comparing this figure with Fig. 11
one can realize that at small nondimensional length scale parameters, the difference between the
results of linear and nonlinear models is insignificant especially for C-C boundary conditions.
But, it is observed that for microbeams with SS-SS and SS-C end conditions at large
nondimensional length scale parameters, there is a considerable difference between the results of
two models. It can be thus concluded that the nonlinearity effects are more important when the
nondimensional length scale parameter gets larger and the boundary conditions become softer.
The variation of nondimensional frequency versus nondimensional Casimir parameter is
indicated in Fig. 13. As the Casimir force increases, the frequencies monotonously decreases for
all kinds of end supports and all values of h / l . The size effect can also be seen in this figure.

6. Conclusion
In this article, the nonlinear pull-in and free vibration behaviors of electrostatically actuated
microswitches made of FGMs including the Casimir force effect were investigated. In order to
capture the size effects, the most general form of strain gradient elasticity was utilized. The
developed elastic model was based on the Timoshenko beam theory and von Kármán’s
geometric nonlinearity. A variational approach was employed to derive the governing equations
which were then numerically solved using the GDQ and pseudo arc-length methods. In the
numerical results, the nonlinear pull-in instability and free vibration characteristics of SS-SS, C-
C and SS-C microswitches were analyzed. The results were obtained based on different elasticity
theories including MCST, MSGT and CT using both linear and nonlinear beam models. The
effects of length scale parameter, geometric nonlinearity, length-to-thickness ratio, material
gradient index and initial center gap were shown. Some important conclusions are as follows.
The size has an important effect on the pull-in voltage and pull-in Casimir parameter of
microswitches when their thickness is comparable with the length scale parameter. The pull-in

16
stability of microswitches improves as h / l decreases. Also, by increasing the aforementioned
ratio, the results of size-dependent model tend to those of classical model.

• The importance of geometric nonlinearity effect increases by increasing the


nondimensional length scale parameter and also by softening the boundary conditions.
Moreover, the pull-in voltage is underestimated when the mid-plane stretching effect is
neglected.
The results of considered elasticity theories tend to converge at large values of h / l . At small
sizes, the results of MSGT are larger than those of MCST and CT.

• The pull-in voltage increases as the initial center gap becomes larger. In addition, by
increasing the initial center gap, the nonlinearity is intensified.

• The free vibration response of actuated SS-SS microswitches is different from that of
C-C and SS-C ones.

Acknowledgments
The corresponding author gratefully acknowledges the support provided by Lahijan Branch,
Islamic Azad University.

References
[1] Zimmermann, L., Ebersohl, J. Ph., Le Hung, F., Berry, J. P., Baillieu, F., Rey, P., Diem, B.,
Renard, S., and Caillat, P., 1995, “Airbag Application: A Microsystem Including a Silicon
Capacitive Accelerometer, CMOS Switched Capacitor Electronics and True Self-Test
Capability,” Sens. Actuat. A: Phys., 46, pp. 190–195.

[2] Noetzel, J., Tønnesen, T., Benecke, W., Binder, J., and Mader, G., 1996, “Quasianalog
Accelerometer Using Microswitch Array,” Sens. Actuat. A: Phys., 54, pp. 574–578.

17
[3] Tønnesen, T., Lüdtke, O., Noetzel, J., Binder, J., and Mader, G., 1997, “Simulation, Design
and Fabrication of Electroplated Acceleration Switches,” J. Micromech. Microeng., 7, p. 237.

[4] Michaelis, S., Timme, H. J., Wycisk, M., and Binder, J., 2000, “Additive Electroplating
Technology as a Post-CMOS Process for the Production of MEMS Acceleration-Threshold
Switches for Transportation Applications,” J. Micromech. Microeng., 10, p. 120.

[5] Jeong,Y. J., Park, J. H., and Lee, D. W., 2014, “Pressure Level Sensor Using a Conductive
Diaphragm and Microswitch Arrays,” Sens. Actuat. A: Phys., 218, pp. 154–161.

[6] Nathanson, H. C., Newell, W. E., Wickstrom, R. A., Davis, and J. R., 1967, “The Resonant
Gate Transistor,” IEEE T. Electron Dev., 14, pp. 117-133.

[7] Taylor, G. I., 1968, “The Coalescence of Closely Spaced Drops When They Are at Different
Electric Potentials,” Proc R. Soc. London A, 306, pp. 423-434.

[8] Bochobza-Degani, O., and Nemirovsky, Y., 2002, “Modeling the Pull-In Parameters of
Electrostatic Actuators with a Novel Lumped Two Degrees of Freedom Pull-In Model,” Sens.
Actuat. A: Phys., 97–98, pp. 569-578.

[9] Zhao, X. P., Abdel-Rahman, E. M., and Nayfeh, A. H., 2004, “A Reduced-Order Model for
Electrically Actuated Microplates,” J. Micromech. Microeng., 14, pp. 900-906.

[10] Rong, H., Huang, Q. A., Nie, M., and Li, W., 2004, “An Analytical Model for Pull-In
Voltage of Clamped–Clamped Multilayer Beams,” Sens. Actuat. A: Phys., 116, pp. 15-21.

[11] Soleymani, P., Sadeghian, H., Tahmasebi, A., and Rezazadeh, G., 2006, “Pull-In Instability
Investigation of Circular Micro Pump Subjected to Nonlinear Electrostatic Force,” Sensors
Transducers J., 69, pp. 622-628.

[12] Chao, P. C. P., Chiu, C. W., and Tsai, C. Y., 2006, “A Novel Method to Predict the Pull-In
Voltage in a Closed Form for Micro-Plates Actuated by a Distributed Electrostatic Force,” J.
Micromech. Microeng., 16, pp. 986-998.

18
[13] Batra, R. C., Porfiri, M., and Spinello, D., 2008, “Reduced-Order Models for
Microelectromechanical Rectangular and Circular Plates Incorporating the Casimir Force,”
Int. J. Solids Struct., 45, pp. 3558-3583.

[14] Nabian, A., Rezazadeh, G., Haddad-derafshi, M., and Tahmasebi, A., 2008, “Mechanical
Behavior of a Circular Micro plate Subjected to Uniform Hydrostatic and Non-Uniform
Electrostatic Pressure,” Microsyst. Technol., 14, pp. 235-240.

[15] Rezazadeh, G., Talebian, S., and Fathalilou, M., 2009, “Dynamic Pull-in Phenomenon in the
Fully Clamped Electrostatically Actuated Rectangular Microplates Considering Damping
Effects,” Sensor Transducer J., 103, pp. 122-131.

[16] Mojahedi, M., Moghimi Zand, M., and Ahmadian, M. T., 2010, “Static Pull-In Analysis of
Electrostatically Actuated Microbeams Using Homotopy Perturbation Method,” Appl. Math.
Model., 34, pp. 1032-1041.

[17] Talebian, S., Rezazadeh, G., Fathalilou, M., and Toosi, B., 2010, “Effect of Temperature on
Pull-In Voltage and Natural Frequency of an Electrostatically Actuated Microplate,”
Mechatronics, 20, pp. 666-673.

[18] Wang, Y. G., Lin, W. H., Li, X. M., and Feng, Z. J., 2011, “Bending and Vibration of an
Electrostatically Actuated Circular Microplate in Presence of Casimir Force,” Appl. Math.
Model., 35, pp. 2348-2357.

[19] Fleck, N. A., Muller, G. M., Ashby, M. F., and Hutchinson, J. W., 1994, “Strain Gradient
Plasticity: Theory and Experiment,” Acta Metall. Mater., 42, pp. 475-487.

[20] Stölken, J. S., and Evans, A. G., 1998, “A Microbend Test Method for Measuring the
Plasticity Length Scale,” Acta Mater., 46, pp. 5109-5115.

[21] Nix, W. D., and Gao, H., 1998, “Indentation Size Effects in Crystalline Materials: A Law
for Strain Gradient Plasticity,” J. Mech. Phys. Solids, 46, pp. 411-425.

[22] Chasiotis, I., and Knauss, W. G., 2003, “The Mechanical Strength of Polysilicon Films: Part
2. Size Effects Associated with Elliptical and Circular Perforations,” J. Mech. Phys. Solids,
51, pp. 1551-1572.

19
[23] McFarland, A. W., and Colton, J.S., 2005, “Role of Material Microstructure in Plate
Stiffness with Relevance to Microcantilever Sensors,” J. Micromech. Microeng., 15, pp.
1060-1067.

[24] Mindlin, R. D., and Tiersten, H. F., 1962, “Effects of Couple-Stresses in Linear Elasticity,”
Arch. Ration. Mech. Anal., 11, pp. 415-448.

[25] Koiter, W. T., 1964, “Couple Stresses in the Theory of Elasticity,” Proceedings of
the Koninklijke Nederlandse Akademie van Wetenschappen (B), 67, pp. 17-44.

[26] Mindlin, R. D., 1964, “Micro-Structure in Linear Elasticity,” Arch. Ration. Mech. Anal., 6,
pp. 51-78.

[27] Mindlin, R. D., 1965, “Second Gradient of Strain and Surface Tension in Linear Elasticity,”
Int. J. Solids Struct., 1, pp. 417-438.

[28] Yang, F., Chong, A. C. M., Lam, D. C. C., and Tong, P., 2002, “Couple Stress Based Strain
Gradient Theory for Elasticity,” Int. J. Solids Struct., 39, pp. 2731-2743.

[29] Lam, D. C. C., Yang, F., Chong, A. C. M., Wang, J., Tong, P., 2003, “Experiments and
Theory in Strain Gradient Elasticity,” J. Mech. Phys. Solids, 51, pp. 1477-1508.

[30] Eringen, A. C., 1983, “On Differential Equations of Nonlocal Elasticity and Solutions of
Screw Dislocation and Surface Waves,” J. Appl. Phys., 54, pp. 4703-4710.

[31] Eringen, A. C., 2002, Nonlocal Continuum Field Theories, Springer, New York.

[32] Wang, B., Zhou, S., Zhao, J., and Chen, X., 2011, “Pull-in Instability Analysis of
Electrostatically Actuated Microplate with Rectangular Shape,” Int. J. Precision Eng.
Manufact., 12, pp. 1085-1094.

[33] Wang, B., Zhou, S., Zhao, J., and Chen, X., 2011, “Size-Dependent Pull-In Instability of
Electrostatically Actuated Microbeam-Based MEMS,” J. Micromech. Microeng., 21, p.
027001.

20
[34] Wang, B., Zhou, S., Zhao, J., and Chen, X., 2012, “Pull-In Instability of Circular Plate
MEMS: A New Model Based On Strain Gradient Elasticity Theory,” Int. J. Appl. Mech., 04,
p. 1250003.

[35] Ansari, R., Gholami, R., Mohammadi, V., and Faghih Shojaei, M., 2013, “Size-Dependent
Pull-In Instability of Hydrostatically and Electrostatically Actuated Circular Microplates,”
ASME J. Comput. Nonlinear Dyn., 8, p. 021015.

[36] Rahaeifard, M., and Ahmadian, M. T., 2015, “On Pull-In Instabilities of Microcantilevers,”
Int. J. Eng. Sci., 87, pp. 23-31.

[37] Seyyed Fakhrabadi M. M., Rastgoo, A., Ahmadian, M. T., 2014, “Non-linear Behaviors of
Carbon Nanotubes under Electrostatic Actuation Based on Strain Gradient Theory,” Int. J.
Non-Linear Mech., 67, pp. 236-241.

[38] Sedighi, H. M., Koochi, A., and Abadyan, M., 2014, “Modeling the Size Dependent Static
and Dynamic Pull-In Instability of Cantilever Nanoactuator Based on Strain Gradient
Theory,” Int. J. Appl. Mech., 6, 1450055.

[39] Sedighi, H. M., 2014, “Size-Dependent Dynamic Pull-In Instability of Vibrating Electrically
Actuated Microbeams Based on the Strain Gradient Elasticity Theory,” Acta Astronautica,
95, pp. 111–123.

[40] Maani Miandoab, E., Nejat Pishkenari, H., and Yousefi-Koma, A., 2014, “Dynamic
Analysis of Electrostatically Actuated Nanobeam Based on Strain Gradient Theory,” Int. J.
Struct. Stab. Dyn., DOI: 10.1142/S021945541450059X.

[41] Maani Miandoab, E., Yousefi-Koma, A., and Nejat Pishkenari, H., 2014, “Poly Silicon
Nanobeam Model Based on Strain Gradient Theory,” Mech. Res. Commun., 62, pp. 83–88.

[42] Tadi Beni, Y., Karimipour, I., and Abadyan, M., 2015, “Modeling the Instability of
Electrostatic Nano-Bridges and Nano-Cantilevers Using Modified Strain Gradient Theory,”
Appl. Math. Model., 39, pp. 2633–2648.

21
[43] Zhao, J. F., Zhou, S. J., and Wang, B. L., 2011, “Size Dependent Pull-in Phenomena in
Electro-Statically Actuated Micro-Beam Based on the Modified Couple Stress Theory,” Adv.
Mater. Res., 335-336, pp. 633-640.

[44] Rahaeifard, M., Kahrobaiyan, M. H., Asghari, M., and Ahmadian, M. T., 2011, “Static Pull-
In Analysis of Microcantilevers Based on the Modified Couple Stress Theory,” Sensors
Actuators A, 171, pp. 370-374.

[45] Zhang, J., and Fu, Y., 2012, “Pull-In Analysis of Electrically Actuated Viscoelastic
Microbeams Based on a Modified Couple Stress Theory,” Meccanica, 47, pp. 1649-1658.

[46] Rahaeifard, M., Kahrobaiyan, M. H., Ahmadian, M. T., and Firoozbakhsh, K., 2012, “Size-
Dependent Pull-In Phenomena in Nonlinear Microbridges,” Int. J. Mech. Sci., 54, pp. 306-
310.

[47] Baghani, M., 2012, “Analytical Study on Size-Dependent Static Pull-In Voltage of
Microcantilevers Using the Modified Couple Stress Theory,” Int. J. Eng. Sci., 54, pp. 99-105.

[48] Rokni, H., Seethaler, R. J., Milani, A. S., Hosseini-Hashemi, S., and Li, X. F., 2013,
“Analytical Closed-Form Solutions for Size-Dependent Static Pull-In Behavior in
Electrostatic Micro-Actuators via Fredholm Integral Equation,” Sens. Actuat. A: Phys. 190,
pp. 32-43.

[49] Askari, A. R., and Tahani, M., 2015, “Size-Dependent Dynamic Pull-In Analysis of Beam-
Type MEMS under Mechanical Shock Based on the Modified Couple Stress Theory,” Appl.
Math. Model., 39, pp. 934-946.

[50] Mousavi, T., Bornassi, S., and Haddadpour, H., 2013, “The Effect of Small Scale on the
Pull-In Instability of Nano-Switches Using DQM,” Int. J. Solids Struct., 50, pp. 1193–1202.

[51] Seyyed Fakhrabadi, M. M., Rastgoo, A., and Ahmadian, M. T., 2014, “Size-Dependent
Instability of Carbon Nanotubes under Electrostatic Actuation Using Nonlocal Elasticity,”
Int. J. Mech. Sci., 80, pp. 144–152.

[52] Najar, F., El-Borgi, S., Reddy, J. N., and Mrabet, K., 2015, “Nonlinear Nonlocal Analysis of
Electrostatic Nanoactuators,” Compos. Struct., 120, pp. 117–128.

22
[53] Jia, X. L., Yang, J., Kitipornchai, S., and Lim, C. W., 2012, “Resonance Frequency
Response of Geometrically Nonlinear Micro-Switches under Electrical Actuation,” J. Sound
Vib., 331, pp. 3397–3411.

[54] Abbasnejad, B., Rezazadeh, G., and Shabani, R., 2013, “Stability Analysis of a Capacitive
FGM Micro-Beam Using Modified Couple Stress Theory,” Acta Mechanica Solida Sinica,
26, pp. 427–440.

[55] Sedighi, H. M., Daneshmand, F., and Zare, J., 2014, “The Influence of Dispersion Forces on
the dynamic Pull-In Behavior of Vibrating Nano-Cantilever Based NEMS Including
Fringing Field Effect,” Arch. Civil Mech. Eng., 14, pp. 766–775.

[56] Liang, B., Zhang, L., Wang, B., and Zhou, S., 2015, “A Variational Size-Dependent Model
for Electrostatically Actuated NEMS Incorporating Nonlinearities and Casimir Force,”
Physica E, 71, pp. 21–30.

[57] Sedighi, H. M., Daneshmand, F., and Abadyan, M., 2015, “Dynamic Instability Analysis of
Electrostatic Functionally Graded Doubly-Clamped Nano-Actuators,” Compos. Struct., 124,
pp. 55–64.

[58] Witvrouw, A., and Mehta, A., 2005, “The Use of Functionally Graded Poly-SiGe Layers for
MEMS Applications,” Mater. Sci. Forum, 492–493, pp. 255–260.

[59] Gromova, M., Mehta, A., Baert, K., and Witvrouw, A., 2006, “Characterization and Strain
Gradient Optimization of PECVD Poly-SiGe Layers for MEMS Applications,” Sens.
Actuat. A: Phys., 130-131, pp. 403–410.

[60] Hasanyan, D. J., Batra, R. C., and Harutyunyan, S., 2008, “Pull-In Instabilities in
Functionally Graded Microthermoelectromechanical Systems,” J. Therm. Stresses, 31, pp.
1006–1021.

[61] Lamoreaux, S. K., 2005, “The Casimir Force: Background, Experiments, and Applications,”
Rep. Prog. Phys., 68, pp. 201–236.

[62] Shu, C., 2000, “Differential Quadrature and Its Application in Engineering,” Springer,
London.

23
[63] Keller, H. B., 1977, “Numerical Solution of Bifurcation and Nonlinear Eigenvalue
Problems, Applications of Bifurcation Theory,” (Proc. Advanced Sem., Univ. Wisconsin,
Madison, Wis., 1976), Academic Press, New York, pp. 359-384.

[64] Das, K., and Batra, R. C., 2009, “Pull-In and Snap-Through Instabilities in Transient
Deformations of Microelectromechanical Systems,” J. Micromech. Microeng., 19, 035008.

[65] Krylov, S., Seretensky, S., and Schreiber, D., 2008, “Pull-In Behavior and Multistability of
a Curved Microbeam Actuated by a Distributed Electrostatic Force,” IEEE 21st Int. Conf. on
Micro Electro Mechanical Systems, Ed. Seretensky,S., pp. 499–502

[66] Jia, X. L., Zhang, S. M., Ke, L. L., Yang, J., and Kitipornchai, S., 2014, “Thermal Effect on
the Pull-In Instability of Functionally Graded Micro-Beams Subjected to Electrical
Actuation,” Compos. Struct., 116, pp. 136-146.

[67] Tilmans, H. A. C., and Legtenberg, R., 1994, “Electrostatically Driven Vacuum-
Encapsulated Polysilicon Resonators, Part II: Theory and Performance,” Sens. Actuat. A:
Phys., 45, pp. 67–84.

List of Captions

Table 1: Linear and nonlinear pull-in voltages (V) and pull-in displacements ( µ m ) of an
electrostatically actuated homogenous clamped-clamped microbeam

Table 2: Convergence of the linear and nonlinear pull-in voltages (V) of electrostatically
actuated homogenous clamped-clamped microbeams with different lengths
Figure 1: Schematic view of an actuated microswitch subjected to the transverse force

24
Figure 2: Nondimensional center gap versus nondimensional applied voltage for different
nondimensional length scale parameters based on MSGT ( L / h = 12, k = 0.2, g 0 = 1.2 )

Figure 3: Nondimensional center gap versus nondimensional Casimir parameter for different
nondimensional length scale parameters based on MSGT ( L / h = 12, k = 0.2, g 0 = 1.2 )

Figure 4: Nondimensional critical voltage versus nondimensional length scale parameter based
on different models ( L / h = 12, k = 0.2,, g 0 = 1.2 )

Figure 5: Nondimensional critical Casimir parameter versus nondimensional length scale


parameter based on different models ( L / h = 12, k = 0.2,, g 0 = 1.2 )

Figure 6: Nondimensional center gap versus nondimensional applied voltage for different
material gradient indexes based on MSGT ( L / h = 12, h / l = 2, g 0 = 1.2 )

Figure 7: Nondimensional center gap versus nondimensional Casimir parameter for different
material gradient indexes based on MSGT ( L / h = 12, h / l = 2, g 0 = 1.2 )

Figure 8: Nondimensional center gap versus nondimensional applied voltage for different lenth-
to-thickness ratios based on MSGT ( h / l = 2, k = 0.2, g 0 = 1.2 )

Figure 9: Nondimensional center gap versus nondimensional Casimir parameter for different
lenth-to-thickness ratios based on MSGT ( h / l = 2, k = 0.2, g 0 = 1.2 )

Figure 10: Nondimensional critical pull-in voltage versus initial center gap ( g0 ) for different
nondimensional length scale parameters based on linear and nonlinear models (
L / h = 12, k = 0.2 )

Figure 11: Nondimensional frequency versus nondimensional applied voltage for different
nondimensional length scal parameters based on MSGT and the nonlinear model (
L / h = 12, k = 0.2, g0 = 1.2 )

Figure 12: Nondimensional frequency versus nondimensional applied voltage for different
nondimensional length scal parameters based on MSGT and the linear model (
L / h = 12, k = 0.2, g 0 = 1.2 )

Figure 13: Nondimensional frequency versus nondimensional Casimir parameter for different
nondimensional length scal parameters based on MSGT ( L / h = 12, k = 0.2, g 0 = 1.2 )

25
Table 1: Linear and nonlinear pull-in voltages (V) and pull-in displacements ( µ m ) of an
electrostatically actuated homogenous clamped-clamped microbeam

Linear analysis Nonlinear analysis


Case Voltage Displacemen Voltage Displacement (
(V) t ( µm ) (V) µm )

Present 65.3832 4.0062 135.537 6.3515

Numerical solution [64]


49 4.5 118.5 6.5
Experimental [65] Pull-in voltage = 100V , Pull-in displacement = 5 µ m

Table 2: Convergence of the linear and nonlinear pull-in voltages (V) of electrostatically
actuated homogenous clamped-clamped microbeams with different lengths
L Numerical simulation Experi
N = 11 N = 21 N = 31
[66] mental
( µm) Linear Nonlinear Linear Nonlinear Linear Nonlinear Linear Nonlinear [67]

210 27.4604 28.2888 27.4614 28.2904 27.4614 28.2904 27.506 28.352 27.95
310 13.8175 14.1569 13.8179 14.1576 13.8179 14.1576 13.834 14.182 13.78
410 8.7619 8.9329 8.7621 8.9333 8.7621 8.9333 8.769 8.945 9.13

26
510 6.2967 6.3939 6.2968 6.3941 6.2968 6.3941 6.300 6.401 6.57

27
Figure 1: Schematic view of an actuated microswitch subjected to the transverse force

(a) C-C (b) SS-C


1 1
0.9 0.9
α cr = 1.594 α cr = 1.164
0.8 0.8
α cr = 1.275 α cr = 0.926
Non. Dim. Center Gap

Non. Dim. Center Gap

0.7 0.7
α cr = 1.114 α cr = 0.815
0.6 0.6
α cr = 1.023 α cr = 0.755
0.5 α cr = 0.931 0.5
α cr = 0.698
0.4 α cr = 0.874 0.4 α cr = 0.664
0.3 α cr = 0.826 0.3 α cr = 0.638
0.2 α cr = 0.8 0.2 α cr = 0.624

0.1 0.1
0 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 0.2 0.4 0.6 0.8 1 1.2
Non. Dim. Voltage ( α ) Non. Dim. Voltage ( α )

28
(c) SS-SS
1
0.9

0.8 h/l = 2, α cr = 0.803


h/l = 3, α cr = 0.649
Non. Dim. Center Gap

0.7
h/l = 4, α cr = 0.585
0.6
h/l = 5, α cr = 0.554
0.5
h/l = 7, α cr = 0.525
0.4 h/l = 10, α cr = 0.51
0.3 h/l = 20, α cr = 0.499
0.2 Classic, α cr = 0.494

0.1

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Non. Dim. Voltage ( α )

Figure 2: Nondimensional center gap versus nondimensional applied voltage for different
nondimensional length scale parameters based on MSGT ( L / h = 12, k = 0.2, g0 = 1.2 )

(a) C-C (b) SS-C


1 1
0.9 β cr = 3.926 0.9 β cr = 3.335
0.8 β cr = 3.496 0.8 β cr = 2.946
Non. Dim. Center Gap

β cr = 3.255
Non. Dim. Center Gap

0.7 0.7 β cr = 2.739


0.6 β cr = 3.109
0.6 β cr = 2.619
β cr = 2.953
0.5 0.5 β cr = 2.495
β cr = 2.851
0.4 0.4 β cr = 2.418
β cr = 2.763
0.3 β cr = 2.356
β cr = 2.716
0.3
β cr = 2.324
0.2 0.2
0.1 0.1
0 0
0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2 3.6 4 0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2 3.6
Non. Dim. Casimir Parameter ( β ) Non. Dim. Casimir Parameter ( β )

29
(c) SS-SS
1
0.9 h/l = 2, β cr = 2.732
0.8 h/l = 3, β cr = 2.407
Non. Dim. Center Gap

0.7 h/l = 4, β cr = 2.247

0.6 h/l = 5, β cr = 2.159


h/l = 7, β cr = 2.074
0.5
h/l = 10, β cr = 2.025
0.4
h/l = 20, β cr = 1.989
0.3
Classic, β cr = 1.975
0.2

0.1
0
0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7
Non. Dim. Casimir Parameter ( β )

Figure 3: Nondimensional center gap versus nondimensional Casimir parameter for different
nondimensional length scale parameters based on MSGT ( L / h = 12, k = 0.2, g0 = 1.2 )

30
(a) C-C (b) SS-C
2.4 1.8
Non. Dim. Critical Voltage (α cr )

Non. Dim. Critical Voltage (α cr )


2.2
1.6
2
1.4
1.8
1.6 1.2
1.4 1
1.2
0.8
1
0.8 0.6

0.6 0.4
2 4 6 8 10 12 14 16 18 20 2 4 6 8 10 12 14 16 18 20
Non. Dim. Length Scale Parameter ( h/l ) (c) SS-SS Non. Dim. Length Scale Parameter ( h/l )
1.4
Non. Dim. Critical Voltage (α cr )

1.2 MSGT, Nonlinear


MCST, Nonlinear
1 CT, Nonlinear
MSGT, Linear
0.8 MCST, Linear
CT, Linear
0.6

0.4

0.2
2 4 6 8 10 12 14 16 18 20
Non. Dim. Length Scale Parameter ( h/l )

Figure 4: Nondimensional critical voltage versus nondimensional length scale parameter based
on different models ( L / h = 12, k = 0.2, g0 = 1.2 )

31
(a) C-C (b) SS-C
4.2

Non. Dim. Critical Casimir Parameter (βcr )


5
Non. Dim. Critical Casimir Parameter ( βcr )

4.8 4
4.6 3.8
4.4 3.6
4.2 3.4
4
3.2
3.8
3
3.6
3.4 2.8
3.2 2.6
3 2.4
2.8 2.2
2.6 2
2 4 6 8 10 12 14 16 18 20
2 4 6 8 10 12 14 16 18 20
Non. Dim. Length Scale Parameter ( h/l )
Non. Dim. Length Scale Parameter ( h/l ) (c) SS-SS
Non. Dim. Critical Casimir Parameter (βcr )

3.6
3.4 MSGT, Nonlinear
MCST, Nonlinear
3.2
CT, Nonlinear
3 MSGT, Linear
2.8 MCST, Linear
CT, Linear
2.6
2.4
2.2
2
1.8
2 4 6 8 10 12 14 16 18 20
Non. Dim. Length Scale Parameter ( h/l )

Figure 5: Nondimensional critical Casimir parameter versus nondimensional length scale


parameter based on different models ( L / h = 12, k = 0.2, g0 = 1.2 )

32
(a) C-C (b) SS-C
1 1

0.9 0.9
α cr = 1.735
0.8 0.8 α cr = 1.265
α cr = 1.594
0.7 α cr = 1.164

Non. Dim. Center Gap


0.7
Non. Dim. Center Gap

α cr = 1.408
α cr = 1.031
0.6 α cr = 1.244 0.6
α cr = 0.914
0.5 α cr = 1.121 0.5
α cr = 0.826
α cr = 0.94
0.4 0.4 α cr = 0.693
α cr = 0.845
0.3 0.3 α cr = 0.623
α cr = 0.687
α cr = 0.506
0.2 0.2

0.1 0.1

0 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 0 0.2 0.4 0.6 0.8 1 1.2
Non. Dim. Voltage ( α ) Non. Dim. Voltage ( α )

(c) SS-SS
1
Ceramic, α cr = 0.865
0.9
k = 0.2, α cr = 0.803
0.8 k = 0.6, α cr = 0.721
Non. Dim. Center Gap

0.7 k = 1.2, α cr = 0.648


0.6 k = 2.0, α cr = 0.591

0.5 k = 5.0, α cr = 0.498


k = 10, α cr = 0.443
0.4
Metal, α cr = 0.352
0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Non. Dim. Voltage ( α )

Figure 6: Nondimensional center gap versus nondimensional applied voltage for different
material gradient indexes based on MSGT ( L / h = 12, h / l = 2, g 0 = 1.2 )

33
(a) C-C (b) SS-C
1 1
0.9 β cr = 4.097 0.9 β cr = 3.477
0.8 β cr = 3.926 β cr = 3.335
0.8
β cr = 3.688 β cr = 3.136
Non. Dim. Center Gap

Non. Dim. Center Gap


0.7 0.7
β cr = 3.466 β cr = 2.951
0.6 0.6
β cr = 3.29 β cr = 2.803
0.5 β cr = 3.012 0.5
β cr = 2.569
0.4 β cr = 2.856 0.4 β cr = 2.435
0.3 β cr = 2.573 0.3 β cr = 2.191

0.2 0.2

0.1 0.1

0 0
0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2 3.6 4 0 0.5 1 1.5 2 2.5 3 3.5
Non. Dim. Casimir Parameter ( β ) Non. Dim. Casimir Parameter ( β )

(c) SS-SS
1
0.9 Ceramic, β cr = 2.842
0.8 k = 0.2, β cr = 2.732
Non. Dim. Center Gap

0.7 k = 0.6, β cr = 2.583

0.6 k = 1.2, β cr = 2.444


k = 2.0, β cr = 2.331
0.5
k = 5.0, β cr = 2.139
0.4
k = 10, β cr = 2.019
0.3
Metal, β cr = 1.799
0.2
0.1

0
0 0.4 0.8 1.2 1.6 2 2.4 2.8
Non. Dim. Casimir Parameter ( β )

Figure 7: Nondimensional center gap versus nondimensional Casimir parameter for different
material gradient indexes based on MSGT ( L / h = 12, h / l = 2, g0 = 1.2 )

34
(a) C-C (b) SS-C
1 1

0.9 0.9

0.8 0.8
α cr = 2.144
α cr = 1.618

Non. Dim. Center Gap


0.7
Non. Dim. Center Gap

0.7 α cr = 1.832
α cr = 1.357
0.6 α cr = 1.594 0.6
α cr = 1.164
α cr = 1.256 0.5
0.5 α cr = 0.901
α cr = 1.029
0.4 0.4 α cr = 0.732

0.3 0.3

0.2 0.2

0.1 0.1

0 0
0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Non. Dim. Voltage ( α ) Non. Dim. Voltage ( α )

(c) SS-SS
1
L/h = 8, α cr = 1.158
0.9
L/h = 10, α cr = 0.95
0.8 L/h = 12, α cr = 0.803
Non. Dim. Center Gap

0.7 L/h = 16, α cr = 0.611


0.6 L/h = 20, α cr = 0.492
0.5
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1 1.2
Non. Dim. Voltage ( α )

Figure 8: Nondimensional center gap versus nondimensional applied voltage for different length-
to-thickness ratios based on MSGT ( h / l = 2, k = 0.2, g0 = 1.2 )

35
(a) C-C (b) SS-C
1 1

0.9 0.9
β cr = 4.543 β cr = 3.921
0.8 0.8
β cr = 4.205 β cr = 3.597
0.7 0.7
Non. Dim. Center Gap

Non. Dim. Center Gap


β cr = 3.926 β cr = 3.335
0.6 β cr = 3.487 0.6 β cr = 2.936
0.5 β cr = 3.159 0.5 β cr = 2.647

0.4 0.4

0.3 0.3

0.2 0.2
0.1 0.1
0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0 0.5 1 1.5 2 2.5 3 3.5 4
Non. Dim. Casimir Parameter ( β ) Non. Dim. Casimir Parameter ( β )

(c) SS-SS
1
0.9
L/h = 8, β cr = 3.274
0.8
L/h = 10, β cr = 2.969
Non. Dim. Center Gap

0.7
L/h = 12, β cr = 2.732
0.6
L/h = 16, β cr = 2.386
0.5 L/h = 20, β cr = 2.143
0.4
0.3

0.2
0.1

0
0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2
Non. Dim. Casimir Parameter ( β )

Figure 9: Nondimensional center gap versus nondimensional Casimir parameter for different
length-to-thickness ratios based on MSGT ( h / l = 2, k = 0.2, g0 = 1.2 )

36
(a) C-C (b) SS-C
4
2.8
3.5

Non. Dim. Critical Pull-in Voltage


Non. Dim. Critical Pull-in Voltage

2.4
3
2
2.5
1.6
2
1.2
1.5

1 0.8

0.5 0.4

0 0
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Non. Dim. Initial Gap ( g0 ) Non. Dim. Initial Gap ( g0 )

(c) SS-SS
2
h/l = 2, Nonlinear
1.8
h/l = 2, Linear
Non. Dim. Critical Pull-in Voltage

1.6 h/l = 3, Nonlinear


h/l = 3, Linear
1.4 h/l = 5, Nonlinear
1.2 h/l = 5, Linear
Classic, Nonlinear
1 Classic, Linear
0.8
0.6

0.4

0.2

0
0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Non. Dim. Initial Gap ( g0 )

Figure 10: Nondimensional critical pull-in voltage versus initial center gap ( g0 ) for different
nondimensional length scale parameters based on linear and nonlinear models (
L / h = 12, k = 0.2 )

37
(a) C-C (b) SS-C
2.2 1.6

2 h/l = 2
1.4 h/l = 3
1.8 h/l = 4
1.2

Non. Dim. Frequency (ω )


Non. Dim. Frequency (ω )

1.6 h/l = 5
h/l = 7
1.4 1 h/l = 10
1.2 h/l = 20
0.8
Classic
1
0.8 0.6

0.6 0.4
0.4
0.2
0.2
0 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 0 0.2 0.4 0.6 0.8 1 1.2
Non. Dim. Voltage ( α ) Non. Dim. Voltage ( α )

(c) SS-SS
1
h/l = 2
0.9 h/l = 3
h/l = 4
0.8
h/l = 5
Non. Dim. Frequency (ω )

0.7 h/l = 7
h/l = 10
0.6
h/l = 20
0.5 Classic

0.4
0.3

0.2

0.1

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Non. Dim. Voltage ( α )

Figure 11: Nondimensional frequency versus nondimensional applied voltage for different
nondimensional length scal parameters based on MSGT and the nonlinear model (
L / h = 12, k = 0.2, g 0 = 1.2 )

38
(a) C-C (b) SS-C
2.2 1.6

2
1.4
1.8
1.2
1.6

Non. Dim. Frequency (ω )


Non. Dim. Frequency (ω )

1.4 1
1.2
0.8
1

0.8 0.6

0.6 0.4
0.4
0.2
0.2

0 0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2
Non. Dim. Voltage ( α ) Non. Dim. Voltage ( α )

(c) SS-SS
1.1
h/l = 2
1 h/l = 3
0.9 h/l = 4
h/l = 5
Non. Dim. Frequency (ω )

0.8
h/l = 7
0.7 h/l = 10
0.6 h/l = 20
Classic
0.5
0.4
0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Non. Dim. Voltage ( α )

Figure 12: Nondimensional frequency versus nondimensional applied voltage for different
nondimensional length scal parameters based on MSGT and the linear model (
L / h = 12, k = 0.2, g 0 = 1.2 )

39
(a) C-C (b) SS-C
2.2 1.6
2
1.4
1.8
Non. Dim. Frequency (ω )

1.2

Non. Dim. Frequency (ω )


1.6
1.4 1
1.2
0.8
1
0.8 0.6
0.6
0.4
0.4
0.2
0.2
0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
Non. Dim. Casimir Parameter ( β ) Non. Dim. Casimir Parameter ( β )

(c) SS-SS
1.1
h/l = 2
1 h/l = 3
0.9 h/l = 4
h/l = 5
Non. Dim. Frequency (ω )

0.8
h/l = 7
0.7 h/l = 10
0.6 h/l = 20
Classic
0.5
0.4
0.3
0.2
0.1
0
0 0.5 1 1.5 2 2.5 3
Non. Dim. Casimir Parameter ( β )

Figure 13: Nondimensional frequency versus nondimensional Casimir parameter for different
nondimensional length scal parameters based on MSGT ( L / h = 12, k = 0.2, g 0 = 1.2 )

40
Highlights
• Developing a non-classical nonlinear beam model for FGM microswitches
• Considering size effects based on the most general form of SGT
• Considering the shear deformation effects based on the Timoshenko beam theory
• Deriving the governing equations including the Casimir force effect using a variational approach
• Solving the pull-in and free vibration problems of electrostatically actuated microswitches
• Studying the effects of length scale parameter, Casimir force, types of boundary conditions and other
parameters in the results

41