Author’s Accepted Manuscript

Modified von Kármán equations for elastic

nanoplates with surface tension and surface
elasticity

Y.M. Yue, C.Q. Ru, K.Y. Xu

www.elsevier.com/locate/nlm

PII: S0020-7462(16)30282-7
DOI: http://dx.doi.org/10.1016/j.ijnonlinmec.2016.10.013
Reference: NLM2725
To appear in: International Journal of Non-Linear Mechanics
Received date: 9 June 2016
Revised date: 26 October 2016
Accepted date: 26 October 2016
Cite this article as: Y.M. Yue, C.Q. Ru and K.Y. Xu, Modified von Kármán
equations for elastic nanoplates with surface tension and surface elasticity,
International Journal of Non-Linear Mechanics,
http://dx.doi.org/10.1016/j.ijnonlinmec.2016.10.013
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.
 Modified von Kármán equations for elastic nanoplates with surface
tension and surface elasticity

Y. M. Yue1,2, C. Q. Ru2, K. Y. Xu1,2*

1
Shanghai Institute of Applied Mathematics and Mechanics, Shanghai Key Laboratory of Mechanics in
Energy Engineering, Department of Mechanics, Shanghai University, Shanghai 200072, China
2
Department of Mechanical Engineering, University of Alberta, Edmonton, Canada T6G 3G8

*
Corresponding author. kyxu@shu.edu.cn

Abstract
In this paper, modified von Kármán equations are derived for Kirchhoff nanoplates with surface
tension and surface tension-induced residual stresses. The simplified Gurtin-Murdoch model
which does not contain non-strain displacement gradients in surface stress-strain relations is
adopted, so that the von Kármán strain-compatibility equation can be expressed in terms of the
stress function and deflection. The modified von Kármán equations derived here are different
than the existing related models especially for elastic plates with in-plane movable edges. Unlike
the existing models which predict a surface tension-induced tensile pre-stress for an elastic plate
with in-plane movable edges, the present model predicts that this tensile pre-stress is actually
cancelled by the surface tension-induced residual compressive stress. Our this result is consistent
with recent clarification on similar issue for cantilever beams with surface tension, which implies
that the existing models have incorrectly predicted an unvalid tensile pre-stress for an elastic
plate with in-plane movable edges which leads to significant overestimation of postbuckling load
and free vibration frequencies. In addition, our numerical examples indicated that surface
stresses can moderately increase or decrease postbuckling load and free vibration frequency of
Kirchhoff nanoplate with all in-plane movable edges, depending on the surface elasticity
parameters and the geometrical dimensions of nanoplates.

1
Keywords: Kirchhoff plate, surface tension, surface stress, von Kármán equations, postbuckling,
free vibration

1. Introduction
Over the last decades, beam- and plate-like elastic nanostructures have been widely used in
MEMS/NEMS [1-5]. Owing to the large ratio of surface to volume, the effects of surface tension
and surface elasticity on the mechanical behavior of such elastic nanostructures have attracted
considerable attention. As experimental and atomistic simulation methods for nanoscale
materials are complex, expensive and time-consuming, effective theoretical methods, such as
continuum elastic models, have been widely used to investigate the mechanical behavior of
elastic nanostructures. In an effort to study surface elasticity of small-scale elastic materials,
Gurtin and Murdoch (GM) [6, 7] developed a theoretical framework of surface elasticity in the
1970s, and the related surface elasticity parameters were estimated, among others, by Miller and
Shenoy [8] using atomistic simulation. In the past decade, the linearized GM model has been
widely used to study the influence of surface elasticity on static bending, compressed buckling
and vibration of nanobeams, nanofilms or nanoplates [9-14].
The GM model treats the initial surface stress  0 , called “surface tension”, as a finite value,

while the surface tension  0 -induced residual stress is treated as infinitesimal. Consequently, the
original form of the GM model has two features: 1) the surface stresses depend not only on
surface strains but also on some displacement gradients which cannot be expressed in terms of
surface strains; 2) the surface tension  0 appears as a coefficient in the linear equation of

motion/equilibrium while the  0 -induced residual stress is absent, which predicts an unbalanced
tensile pre-stress for thin nanobeams and nanoplates. Consequently, for instance, for one-
dimensional cantilever nanobeams, the original form of GM model could have given a non-zero
tensile axial stress caused by  0 . Actually, Gurtin et al [15] have correctly stated that the  0 -
induced compressive residual stress must be treated as a finite value and added to eliminate the
tensile axial force caused by  0 , and then there will be no an unbalanced axial force in a
cantilever beam. Park and Klein [16] and Yun and Park [17] investigated surface stress effects on
vibration and bending behavior of metal nanowires using the surface Cauchy-Born model. Their

2
results indicated that whether a non-zero axial force exists depends on the end conditions. Song
et al [18] used a continuum model to analyze mechanical behaviors of nanowires with surface
tension and surface tension-induced residual stresses. By comparing surface Cauchy-Born model
and generalized Young–Laplace model with experimental data, Song et al [18] concluded that
surface tension-induced residual stress is essential for correctly predicting overall mechanical
behavior of cantilever nanobeams/nanowires.
Recently, some modified versions of GM model have been proposed by some authors [19-
21] to address the above-mentioned issues. In the present paper, a strain-consistent model
proposed in Ru [22] is employed to study large deflection behavior of nanoplates with surface
tension and surface tension-induced residual stresses. In particular, modified Von Kármán
equations are derived for the Kirchhoff nanoplates with surface tension and surface tension-
induced residual stresses, and the derived equations are used to study the large deflection
mechanical behaviors of plate-like nanostructures.

2. An elastic nanoplate with surface tension and surface elasticity

An elastic isotropic plate of uniform thickness h is considered here. Rectangular Cartesian
coordinates (x, y, z) are introduced where the xy-plane coincides with the geometric mid-plane
of the plate and the z-coordinate taken positive downward. According to Kirchhoff’s hypothesis,
the displacement field can be represented by
w  x, y, t  w  x, y, t 
u1 ( x, y, z, t )  u*  u  z , u 2 ( x, y , z , t )  v *  v  z , u3 ( x, y, z , t )  w( x, y, t ). (1)
x y
where u, v and w denote the displacements of a material point at (x, y, 0) on the mid-plane
caused by applied mechanical loadings, and u* and v* are the in-plane residual displacements
induced by the initial surface stress  0 .

For the plate with surface tension  0 and  0 -induced residual stresses, the nonlinear strains
of von Kármán type are given by
2w 2w 2w
 xx      z 2 ,  yy   yy   yy  z 2 ,  xy  2 xy   xy   xy  2 z
* 0 * 0 * 0
, (2)
x y xy
xx xx

where  xx* ,  xx* ,  xy*  are the in-plane residual strains induced by surface tension  0 , and

3
 2
u 1  w  v 1  w  v u w w
2

 0
    ,  yy 
0
   ,  xy 
0
  , (3)
x 2  x  y 2  y  x y x y
xx

are the in-plane strains of the mid-plane, called “membrane strains”. The membrane strains
satisfy St. Venant’s compatibility condition

 2 xx0   yy   xy   2 w   2 w  2 w
2 0 2 0 2

     . (4)
y 2 x 2 xy  xy  x 2 y 2
Based on Hooke’s law, the constitutive equations of the bulk plate can be written as
Ez   2 w 2w 
 xx 
E
1  2
 xx  yy   1  2  xx  yy   1  2  xx  yy   1  2  x2  y 2  ,
E * * E 0 0

 
Ez   2 w 2w 
 yy 
E
1  2
 yy xx  1  2  yy xx  1  2  yy xx  1  2  y 2 x2  ,
   
E
 *
  *

E
 0
  0
   (5)
 
E E E Ez 2w
 xy   xy   xy 
*
 xy 
0
1   ,
2 1   2 1   2 1   1  2 xy
where E and  are the elastic modulus and Poisson’s ratio of the bulk plate, respectively.
For surface constitutive relations, the surface strain energy adopted in Ru [22] is
 1 s 2
   0 1   xxs   yys   0  xxs   yys   0  xxs    yys    xy   ,
1 2 2 2
(6)
2  4

where  s
xx ,  xxs , and  xys  are surface strains given by (2) with z=±h/2,

0  s   0 and 0  s   0 and s and s are two surface elastic constants. Just as for the GM
model, the displacements between surface and bulk material are continuous and the surface
strains are equal to the values of bulk strains on the surface. It is verified that the surface stresses
can be expressed as [22]

 xxs   0  0   xxs   yys   20 xxs   0   0  20   xx  0 yy

 2w 2w 
  0   0  20   xx*  0 *yy    0  20   xx0  0 yy0    0  20  2  0 2  z,
 x y 
 yys   0  0   xxs   yys   20 yys   0   0  20   yy  0 xx (7)

 2w 2w 
  0   0  20   *yy  0 xx*    0  20   yy0  0 xx0    0  20  2  0 2  z ,
 y x 
 w
2
 xys  0 xys  0 xy  0 xy*  0 xy0  20 z.
xy

4
As we known, surface stresses in the original GM model depend on surface strains and some
gradients of in-plane displacements (u, v), and they cannot be expressed in terms of surface
strains and the deflection w. The present model, as a simplified GM model, makes the surface
stresses in Eq. (7) depend on surface strains and the deflection w only, independent of any non-
strain gradients of in-plane displacements (u, v). It is this feature that makes it possible to express
the von Kármán compatibility equation (4) in terms of the stress function φ and deflection w (see
Eq. (17)).
Thus, the resultant mid-plane membrane forces are given by
h /2
N xx    xx dz    xxs dc  2 0  N xx*  C1 xx0  C2 yy0 ,
 h /2 c
h /2
N yy    yy dz    yys dc  2 0  N *yy  C1 yy0  C2 xx0 , (8)
 h /2 c
h /2
N xy    xy dz    xys dc  N xy*  C3 xy0 ,
 h /2 c

Eh Eh Eh
C1   20  40 , C2   20 , C3   2 0 , (9)
1  2
1  2
2 1  

where c is the perimeter of the cross-section, h is the plate thickness, N xx* , N *yy , and N xy* are the

residual mid-plane membrane forces caused by surface tension-induced residual stresses, defined
by

2  xx
 *  *yy   2  0  20   xx*  0 *yy   C1 xx*  C2 *yy ,
Eh
N xx* 
1 

2  yy
 *  xx*   2  0  20   *yy  0 xx*   C1 *yy  C2 xx* ,
Eh
N *yy  (10)
1 
 E 
N xy*    20   xy*  C3 xy* .
 2 1   
Thus the membrane forces (8) are given in terms of the in-plane membrane strains (3) and the
deflection w. So the in-plane membrane strains given by the simplified GM model can be written
in term of in-plane membrane forces as
C1  N xx  N xx*  2 0   C2  N yy  N *yy  2 0 
  0
,
C12  C22
xx

C1  N yy  N *yy  2 0   C2  N xx  N xx*  2 0 
 yy0  , (11)
C12  C22
N xy  N xy*
  0
xy .
C3

5
In addition, the total resultant moments can be expressed as
 2w  2 w  h2  2w 2w 
 
h /2
M xx    xx zdz    xxs zdc   D          
y 2  2  y 2 
2 ,
 x x 2
 h /2 c 2 0 0 0

 2w  2 w  h2  2w 2w 

 
h /2
M yy    yy zdz    yys zdc   D 
         (12)
x 2  2  x 2 
2 ,
 y y 2
 h /2 c 2 0 0 0

2w 2w
   xy zdz    xys zdc   D 1  
h /2
M xy  0 h 2 .
 h /2 c xy xy

where D  Eh3 12 1  2  is the flexural rigidity (also called “bending stiffness”) of the plate.

Equations of motion/equilibrium in the direction x, y, z, yield

N xx N xy
  0, (13)
x y
N xy N yy
  0, (14)
x y

 2 M xx  2 M xy  2 M yy  N xx N xy  w  N xy N yy  w
2       
x 2 xy y 2  x y  x  x y  y (15)
2w 2w 2w 2w
 q  N xx 2  N yy 2  2 N xy  h 2 .
x y xy t
where the in-plane inertia of the plate is ignored. Based on Eqs. (8), (13) and (14), the resultant
mid-plane membrane forces can be assumed as
 2  2  2
N xx  2 0  N  h 2 , N yy  2 0  N yy  h 2 , N xy  N xy  h
* * *
, (16)
y x xy
xx

where  (x, y) is the stress function. Thus, substituting Eqs. (11) and (16) into Eq. (4), the
compatibility equation (4) can now be written in terms of the stress function  and the
deflection w as
  2 w  2  2 w  2 w 
   E 
4 *
  2 , (17)
 xy  x y 
2

where
  
 Eh   40  40 1     E  0 1   
C 2
C 2
 h ,
E*  1
 2
(18)
C1h  Eh   20  40  1   
2
 

6
 4 4 4
and 4   2  .
x 4 x 2y 2 y 4
It is stated that the compatibility equation (17) could not be obtained if the original form of the
GM model is followed [12-14, 23, 24] because, in that case, surface stresses depend on some
non-strain gradients of in-plane displacements (u, v) and therefore the membrane forces cannot
be given in terms of the in-plane membrane strains and the deflection w. It is necessary to apply
the simplified GM model.
Utilizing Eqs. (12) and (16) into Eq. (15), the governing equation for large deflection of a
Kirchhoff plate with surface tension and surface elasticity can be finally obtained as
 h2  4 2w  *  2   2 w
 D     2  
0   w   h  q  2  2
w   xx
N  h 
t 2 y 2  x 2
0 0
 2   (19)
  2   2 w  *  2   2 w
  N *yy  h 2  2  2  N xy h  .
 x  y  xy  xy

By defining
 2  2 w  2  2 w  2  2 w
L  w,      2 , (20)
y 2 x 2 x 2 y 2 xy xy

h2
D*  D   0  20  , (21)
2
the governing equation (19) can be rewritten as
2w *  w
2
*  w
2
*  w
2
D*4 w   h  q  2  2
w  N  N  2 N  hL  w,   . (22)
t 2 x 2 y 2 xy
0 xx yy xy

Eqs. (22) and (17) are the modified von Kármán equations for a Kirchhoff elastic plate with
surface tension and surface elasticity. When all surface effects are neglected (i.e.,  0  0,

s =s  0 ), it is readily seen that the above modified von Kármán equations reduce to the
classical von Kármán equations.
In the present modified von Kármán equations, the transverse forces  N xx* , N yy
*
, and N xy* 

depend on the in-plane boundary conditions. For an elastic plate with all edges in-plane
immovable, in-plane residual strains/stresses cannot be developed, i.e.  xx
*
= *yy   xy*  0 . Then it

can be obtained from Eq. (10) that N xx*  N *yy  N xy*  0. Thus for an elastic plane with in-plane

immovable edges, Eq. (22) reduces to

7
 D* 4  2 w q 2 0 2
 w  2    w  L  w,   , (23)
h t h h
Eq. (23) is the same as the deflection governing equation shown in Wang and Wang [13, 14],
while the compatibility equation (17) is still different from that of Wang and Wang [13, 14].
However, when all edges of the plate are in-plane movable, residual stresses in bulk plate
induced by the initial surface tension  0 are nonzero, and the present equation (22) is essentially
different than the previous related models [12-14, 23, 24].
For an elastic plane with in-plane movable edges, in the absence of applied external loads,
the resultant mid-plane membrane forces Eq. (8) can be rewritten as
N xx  2 0  N xx*  0,
N yy  2 0  N *yy  0, (24)
N xy  N xy*  0.

Then the residual mid-plane membrane forces  N xx* , N yy

*
, and N xy*  can be determined as

N xx*  N yy
*
 =  2 0 , N xy*  0. (25)

Then Eq. (22) simplifies into

D* 4 2w q
 w   2   L  w,   . (26)
h t h
Eqs. (26) and (17) are the modified von Karman equations for the plate with all edges in-plane
movable. In Section 3, Eqs. (26) and (17) are used to investigate large deflection buckling and
free vibration of an elastic plate with all in-plane movable edges.
It should be mentioned that Eqs. (23) and (26) are just the governing equations of the plate
with all edges in-plane immovable and the plate with all edges in-plane movable, respectively.
For other boundary conditions Eqs. (23) and (26) are not valid, and Eq. (22) should be used
without any simplifications.

3. Numerical results and discussions

As the previous plate models neglected the surface tension-induced residual stresses, in the
following calculations, the influence of surface tension-induced residual stresses on the
postbuckling and vibration behaviors of nanoplate all in-plane movable edges is investigated,
and compared with the results obtained by the models without surface tension-induced residual
stresses. For the numerical calculations, silver nanoplates with different surface material

8
parameters are analyzed. The bulk material parameters of silver nanoplates are
E  76 109 N m2 and   0.3 [25]. Based on atomistic calculations [26, 27] and experiment
verification [28] of the silver surface properties, it is found that the surface modulus Es and

surface tension  0 of silver beams or plates are in the order 1-10N/m. Thus, we use two sets of
surface material parameters for the calculations shown below [29]:
Set I: 0  20  0

 0  1N m, s  20 N m, s  15 N m, Es  s  2s  10 N m,
Set II: 0  20  0

 0  1N m, s  20 N m, s  5 N m, Es  s  2s  10 N m.

The lengths of the nanoplate in the x and y directions are a and b, respectively. In the following
calculations for both postbuckling and free vibration, we consider a square plate with a=b=20h.
All edges of the plate are simply supported and the boundary conditions about deflection are
given as
w  w, xx  0 for x  0, a
(27)
w  w, yy  0 for y  0, b

(1) Postbuckling under uniaxial compressive load

Let us consider postbuckling of an elastic plate under uniaxial compressive load (-P) per unit
length at the edges x=0 and x=a. Assume that all edges are free of in-plane shear stresses. In
view of Eqs. (16) and (25), the force boundary conditions become
N xx  h, yy   P for x = 0, a
(28)
N xy  h, xy  0 for x  0, a and y  0, b
To study post-buckling of the hinged elastic plate, the nonlinear deflection associated with the
lowest-order buckling mode is of the form
x   y 
w  x, y   f sin   sin  , (29)
 a   b 
where f is the amplitude. Substituting Eq. (29) into Eq. (17), the compatibility equation can be
rewritten as
E* f 2  4  2 x 2 y 
4  2 2 
cos  cos . (30)
2 ab  a b 

9
Referring to [30], the Airy stress function can be expressed as

f 2  a  2 x  b  2 y  P 2
2 2

  E*   cos    cos  y . (31)

32  b  a a b  2h

It should be stated that because Eq. (29) is an approximate solution, the stress function in Eq. (31)
just approximately satisfies the boundary conditions by [14]
b b
0
N xx dy   h, yy dy  bP for x = 0, a
0
(32)

To get the relation between amplitude f and the applied uniform load P, the Galerkin’s method is
applied to solve Eq. (26), which gives
 D* 4  x y
0  h  w  L  w,  sin a sin b dxdy  0.
a b
0
(33)

Solving Eq. (33), the postbuckling relation is obtained as

2  1  E  hf  1
2
2
* 2 2
P  D 2   
*
 2   , (34)
b   16b  
2

where   a / b. Replacing D* and E* by D and E, respectively, Eq. (34) becomes the
postbuckling relation of a Kirchhoff nanoplate without surface effects [30].
Wang and Wang [14] followed the original GM model and did not include surface tension-
induced residual compressive stresses in Eq. (22). Consequently, instead of Eq. (34), the
postbuckling relation they predicted was

2  1  2  s  2s    2 hf 2  1
2
 
P  D*     0        2 .
2
2   2 1  E  2  2
(35)
b    h  16b   
Clearly, Eq. (35) has overestimated the buckling force P. A similar problem for some existing
solutions of 1D nano-cantilevers has been pointed out and clarified by some authors [20, 31, 32].
Our present result for elastic plates with in-plane movable edges is consistent with those of [20,
31, 32] for nano-cantilevers.
To illustrate the influence of surface tension-induced residual stresses, the normalized
postbuckling load P/P0 vs the dimensionless amplitude f/h are shown in Fig. 1 for two sets of

surface material parameters, where P0  D 2 1   2  b2 2 is the critical force for linearized
2

buckling of an otherwise identical elastic plate without any surface stress effect. In Fig. 1(a) for a
square nanoplate with Set I of surface material parameters ( 0  20  0 ), because the surface

10
elasticity enhances the effective bending stiffness, both the present model and Wang and Wang’s
model [14] predict that the postbuckling load is bigger than that of a plate without surface effects.
If the plate thickness h is 5nm, compared to the results without surface effect, the normalized
postbuckling load predicted by Wang and Wang [14] can increase by 130% because of the
unbalanced surface tension  0 effect, but the largest increase given by the present model is only
about 13%. Increasing the plate thickness h to 10nm, the normalized postbuckling load predicted
by both models decreases, as expected.
In Fig. 1(b) for a square nanoplate with Set II of surface material parameters ( 0  20  0 ),
because the surface effect has a softening effect on the effective bending stiffness, the present
model predicts the postbuckling load is a little smaller than that of the plate without surface
stress effect. However, Wang and Wang’s model [14] predicts that the postbuckling load is
greatly bigger than that of the plate without surface stress effect. This clearly indicates that the
model of Wang and Wang [14] considerably overestimates the postbuckling load due to the
unbalanced surface tension  0 effect. Clearly, both the present model and Wang and Wang’s
model [14] predict that the surface stress effect on the normalized postbuckling load is size-
dependent.

(a) (b)

Fig.1 Normalized postbuckling load P/P0 vs normalized amplitude f/h for different plate thickness: (a) for surface
material parameters of Set I, (b) for surface material parameters of Set II. (a=b=20h)

(2) Free vibration

Next, the effect of surface tension-induced residual stresses on free vibration of a
rectangular nanoplate with in-plane movable edges is studied. Assume the deflection function is
given as

11
 x y
w  x, y, t     t  sin sin , (36)
a b
where   t  is a function of time. Substituting Eq. (36) into Eq. (17), the compatibility equation

can be rewritten as
E * 2  t   4  2 x 2 y 
4  2 2 
cos  cos . (37)
2 ab  a b 
The approximate boundary conditions for the stress function  can be written as [13]
b

0 , yy dy  0 for x  0, a
(38)
a
0
, xx dx  0 for y  0, b

The expression of stress function  can be determined by Eqs. (37) and (38) as

E *  a  2 x  b  2 y  2
2 2

    cos    cos   t  . (39)

32  b  a a b 

Using the Galerkin’s method to determine the deflection w, i.e.,

 D* 4 
0  h  w  L  w,    wwdxdy  0,
a b
0
(40)

Thus the governing equation of the nonlinear vibration of the nanoplate is obtained as

d 2
2
    3  0, (41)
dt
in which

D*    E*   
4 4


 h  a 
1   
2 2
,     1    .
16   a 
4
(42)

Eq. (41) is a standard Duffing’s equation. With initial conditions   0  f and   0   0 , the

exact solution can be determined in terms of Jacobi elliptic function cn as [33] (when t=0, cn=1)
 f2 
  t   f cn    f t ,
 2
. (43)
 2    f 2  
 
The frequency  of nonlinear free vibration is obtained as

   f 2
 , (44)
2 K  k ,  2

12
where K is the complete elliptic integral of the first kind, and k  1 2 1     f 2  . It is

known that for an elastic plate without any membrane forces, the frequency of linearized small-
amplitude free vibration is given by [34]
2
0 
a 2 1   
2 D
h
. (45)

It is seen from Eq. (44) that unlike Eq. (45), the frequency of nonlinear correlated depends on the
amplitude f. In the absence of surface stresses, letting D*=D and E*=E in Eq. (42), we have

D   E  
4 4

  1    , 1    1    ,
2 2
1  4
(46)
h  a  16   a 
the frequency of nonlinear free vibration of the nanoplate without surface effects is obtained as,

 1  1 f
2

1  , (47)
2 K  k1 ,  2 

where k1  1 2 1  1  1 f 2  .

Wang and Wang [13] studied the influence of surface tension and surface elasticity on the
nonlinear free vibration of nanoscale plates. They adopted the original GM surface model and
did not include the surface tension-induced residual compressive stresses in Eq. (22). The
nonlinear free vibration frequency predicted by their model is

  2  2 f
2

2  , (48)
2 K  k2 ,  2 

with k2  1 2 1   2  2 f 2  , and

D*    2 0     2  s  2s   1   
4 2 4

  1       1    ,  2   E    1    . (49)
2 2
2  2

4

h  a  h  a   h  16   a 
Clearly, Eqs. (48) and (49) of Wang and Wang [13] overestimate the frequency of nonlinear
vibration due to the unbalanced surface tension  0 . Similar problems for some previous
solutions of 1D nano-cantilevers have been pointed out by some authors [31, 32]. Clearly, our
present result for elastic plates with in-plane movable edges is consistent with those of [31, 32]
for nano-cantilevers.

13
 Now let us examine the effects of surface tension and surface tension-induced residual
stresses by comparison of the frequencies predicted by Eq. (44) given by the present model, Eq.
(48) given by Wang and Wang’s model [13], and Eq. (47) given by the plate without surface
stress effects.
Fig. 2 shows the normalized free vibration frequency  0 against the normalized
amplitude f/h for two sets of surface material parameters. Whatever the plate is with surface
material parameters of Set I or Set II, it is clearly shown that when f/h>0.5,  0 increases with
the increasing normalized amplitude f/h. For the plate with small deflection (f/h<0.5), the change
of amplitude has small influence on the free vibration frequency. Actually, large amplitude f
results in large membrane forces, and thus the free vibration frequency increases. By comparing
the normalized frequencies of the plate with and without surface effects, it is found that the
difference between frequencies predicted by the plate models with and without surface stress
effects decrease with the increasing f/h. Indeed, it follows from Eqs. (47) and (48) that

2 K  k ,  2   2   2 f 2
 . (50)
1 K  k1 ,  2  1  1 f 2

With f increasing, 2 1 approaches l  1 (a constant determined by l  1 , which means

when the amplitude f increases, the influence of 1 and  2 on normalized frequency is weaker

and weaker). Based on the expressions of 1 and  2 , it can be seen that effective bending
stiffness has a decreasing effect on the free vibration frequency.
Compared to the normalized frequency of nanoplate without surface stress effect, it is
obvious that with surface material parameters of Set I ( 0  20  0 ) (Fig. 2(a)), due to the
enhanced bending stiffness, the normalized frequency predicted by the present model has a small
increase (the maximum increase is about 6% for h=5nm and about 3% for h=10nm). For the
plate with surface material parameters of Set II ( 0  20  0 ) (Fig. 2(b)), since the bending
stiffness is softened by surface stress effects, the normalized frequency predicted by the present
model has a considerable decrease (the maximum decrease is more than 8% for h=5nm and more
than 4% for h=10nm). However, the model of Wang and Wang [13] predicts that, because of the
unbalanced surface tension-induced tensile pre-stress, whatever the plate is with surface material
parameters of Set I or Set II, the normalized frequency predicted by [13] is always bigger than

14
that of the plate without surface effects. Moreover, the normalized frequency predicted by Wang
and Wang’s model [13] can have more than 40% increase for h=5m and more than 20% increase
for h=10m, as compared with results without surface stress effects. Clearly, when surface
tension-induced residual stresses are not included in the governing equation (22), free vibration
frequency of micro/nanoplates is greatly overestimated by the Wang and Wang’s model [13].

(a) (b)

Fig.2 Normalized frequency  0 vs normalized amplitude f/h for different plate thickness: (a) for surface material

parameters of Set I, (b) for surface material parameters of Set II. (a=b=20h)

4. Conclusions
Modified von Kármán equations for Kirchhoff nanoplates with surface tension and surface
tension-induced residual stresses are derived, which have similar elegant mathematical form as
the classical von Kármán equations without surface stress effect. Employing the modified von
Kármán equations, the influence of surface tension-induced residual stresses on postbuckling and
nonlinear free vibration behaviors is analyzed. Numerical results for nanoplates of thickness in
nanometers show that surface tension-induced residual stresses are crucial for the nanoplates
with all edges in-plane movable. The existing models which only accounted for surface tension
but not for the surface tension-induced residual stress can greatly overestimate the postbuckling
load and natural frequency of a Kirchhoff plate. The present model predicts that, depending on
specific material and surface stress parameters, postbuckling load and nonlinear frequencies of a
Kirchhoff nanoplate with surface stress effects may increase or decrease as compared to an
identical Kirchhoff nanoplate without surface stress effects.

15
Acknowledgements

Yue and Xu thank the support of China Scholarship Council (CSC), National Natural Science
Foundation of China [grant numbers 10772106, 11072138], Natural Science Foundation of
Shanghai [grant number15ZR1416100] and Shanghai Leading Academic Discipline Project
[grant number S30106], and Ru thanks the support of the Natural Science and Engineering
Research Council (NSERC) of Canada [grant number RGPIN204992].

References
[1] D. Akin, A. Gupta, R. Bashir, Detection of bacterial cells and antibodies using surface
micromachined thin silicon cantilever resonators, Journal of Vacuum Science & Technology
B 22 (2004) 2785–2791.
[2] R. Sandberg, W. Svendsen, K. Molhave, A. Boisen, Temperature and pressure dependence of
resonant in multi-layer microcantilevers, Journal of Micromechanics and Microengineering,
15 (2005) 1454–1458
[3] S. Roy, Z. Gao, Nanostructure-based electrical biosensors, Nano Today 4 (2009) 318–334.
[4] D. I. Caruntu, I. Martinez, Reduced order model of parametric resonance of electrostatically
actuated MEMS cantilever resonators, International Journal of Non-Linear Mechanics 66
(2014) 28-32.
[5] V. Yantchev, I. Katardjiev, Thin film Lamb wave resonators in frequency control and sensing
applications: a review, Journal of Micromechanics and Microengineering, 23 (2013) 043001.
[6] M. E. Gurtin, A. I. Murdoch, A continuum theory of elastic material surfaces, Archive for
Rational Mechanics and Analysis 57 (1975a) 291–323.
[7] M. E. Gurtin, A. I. Murdoch, Addenda to our paper: a continuum theory of elastic material
surfaces, Archive for Rational Mechanics and Analysis 59 (1975b) 389–390.
[8] R. E. Miller, V. B. Shenoy, Size-dependent elastic properties of nanosized structural elements,
Nanotechnology 11 (2000) 139–147.
[9] R. Ansari, S. Sahmani, Bending behavior and buckling of nanobeams including surface stress
effects corresponding to different beam theories, International Journal of Engineering Science
49 (2011) 1244–1255.
[10] M.J. Lachut, J.E. Sader, Effect of surface stress on the stiffness of thin elastic plates and
beams, Physical Review B 85 (2012) 085440.

16
[11] V.A. Eremeyev, H. Altenbach, N.F. Morozov, The influence of surface tension on the
effective stiffness of nanosize plates, Doklady Physics 54 (2009) 98–100.
[12] C.W. Lim, L.H. He, Size-dependent nonlinear response of thin elastic films with nano-scale
thickness, International Journal of Mechanical Sciences 46 (2004) 1715-1726.
[13] K.F. Wang, B.L. Wang, Effect of residual surface stress and surface elasticity on the
nonlinear free vibration of nanoscale plates, Journal of Applied Physics 112 (2012) 013520.
[14] K.F. Wang, B.L. Wang, Effect of surface energy on the nonlinear postbuckling behavior of
nanoplates, International Journal of Non-Linear Mechanics 55 (2013) 19-24.
[15] M.E. Gurtin, X. Markenscoff, R.N. Thurston, Effect of surface stress on the natural
frequency of thin crystals, Applied Physics Letters 29 (1976) 529-530.
[16] H.S. Park, P.A. Klein, Surface stress effects on the resonant properties of metal nanowires:
the importance of finite deformation kinematics and the impact of the residual surface stress,
Journal of Mechanics and Physics of Solids 56 (2008) 3144–3166.
[17] G. Yun, H.S. Park, Surface stress effects on the bending properties of fcc metal nanowires,
Physical Review B 79 (2009) 195421.
[18] F. Song, G.L. Huang, H.S. Park, X.N. Liu, A continuum model for the mechanical behavior
of nanowires including surface and surface-induced initial stresses, International Journal of
Solids and Structures 48 (2011) 2154–2163.
[19] S.G. Mogilevskaya, S.L. Crouch, A. LaGrotta, H.K. Stolarski, The effects of surface
elasticity and surface tension on the transverse overall elastic behavior of unidirectional nano-
composites, Composites Science and Technology 70 (2010) 427–434.
[20] C.Q. Ru, Simple geometrical explanation of Gurtin–Murdoch model of surface elasticity
with a clarification of its related versions, Science China Physics Mechanics Astronomy 53
(2010) 536–544.
[21] P. Chhapadia, P. Mohammadi, P. Sharma, Curvature-dependent surface energy and
implications for nanostructures, Journal of the Mechanics and Physics of Solids 59 (2011)
2103–2115.
[22] C.Q. Ru, A strain-consistent elastic plate model with surface elasticity, Continuum
Mechanics and Thermodynamics 28 (2016) 263-273.
[23]C.W. Wang, et al, Silver nanoplates formed at the air/water and solid/water interfaces,
Colloids and Surface A: Physcochemical and Engineering Aspects 340 (2009) 93–98.

17
[24] L. Huang, Y. Zhai, S. Dong, J. Wang, Efficient preparation of silver nanoplates assisted by
non-polar solvents, Journal of Colloid and Interface Science 331 (2009) 384–388.
[25] K.F. Wang, T. Katamura, B.L. Wang, Nonlinear pull-in instability and free vibration of
micro/nanoscale plates with surface energy – A modified couple stress theory model,
International Journal of Mechanical Sciences 99 (2015) 288–296.
[26]R. E. Miller, V.B. Shenoy, Size-dependent elastic properties of nanosized structural elements,
Nanotechonoly 11 (2000) 139-147.
[27] V.B. Shenoy, Atomistic calculations of elastic properties of metallic fcc crystal surfaces,
Physical Review B 71(2005) 094104.
[28] G.Y. Jing et al, Surface effects on elastic properties of silver nanowires: Contact atomic-
force microscopy, Physical Review B 73 (2006) 235409.
[29]J. He, M. Lilly, Surface stress effect on bending resonance of nanowires with different
boundary conditons, Applied Physics Letters 93 (2008) 263108.
[30] E. Ventsel, T. Krauthammer, Thin Plates and Shells, Marcel Dekker Inc., New York (2004)
[31] M.J. Lachut, J.E. Sader, Effect of surface stress on the stiffness of cantilever plates, Physical
Review Letters 99 (2007) 206102.
[32] R.K. Karabalin, L.G. Villanueva, M.H. Matheny, J.E. Sader, M.L. Roukes, Stress-induced
variations of the stiffness of micro- and nanocantilever beams, Physical Review Letters 108
(2012) 236101.
[33] A. Salas, J. Castillo, Exact Solution to Duffing Equation and the Pendulum Equation,
Applied Mathematical Sciences, 8 (2014) 8781-8789.
[34] A.W. Leissa, Vibration of plates, NASA, SP-160 (1969).

Highlights
 The surface model contains surface elasticity, surface tension and surface tension-induced
residual stresses.
 Modified von Kármán equations are derived for Kirchhoff nanoplates with surface effects.
 Distributions of surface tension-induced residual stresses in the plates depend on the
boundary conditions.

18
 Neglecting surface tension-induced residual stresses may lead to overestimate the surface
effect.

19