Eshelby tensors and overall properties of nano-composites

considering both interface stretching and bending effects

Junbo Wang1, Peng Yan1, Leiting Dong 1*, Satya N. Atluri 2

1
School of Aeronautic Science and Engineering, Beihang University, Beijing 100191,

CHINA

2
Department of Mechanical Engineering, Texas Tech University, USA

Abstract: In this study, the fundamental framework of analytical micromechanics

is generalized to consider nano-composites with both interface stretching and

bending effects. The interior and exterior Eshelby tensors for a spherical nano-

inclusion, with an interface defined by the Steigmann–Ogden model, subjected to

an arbitrary uniform eigenstrain are derived for the first time. Correspondingly, the

stress/strain concentration tensors for a spherical nano-inhomogeneity subjected to

arbitrary uniform far-field stress/strain loadings are also derived. Using the

obtained concentration tensors, the effective bulk and shear moduli are derived by

employing the dilute approximation and the Mori-Tanaka method, respectively,

which can be used for both nano-composites and nano-porous materials. An

equivalent interface curvature parameter reflecting the influence of the interface

bending resistance is found, which can significantly simplify the complex

expressions of the effective properties. In addition to size-dependency, the closed

* Corresponding author: ltdong@buaa.edu.cn (L. Dong). Address: School of Aeronautic Science and Engineering,

Beihang University, Beijing, 100191, CHINA.

1 of 34 pages
 form expressions show that the effective bulk modulus is invariant to interface

bending resistance parameters, in contrast to the effective shear modulus. We also

put forward for the first time a characteristic interface curvature parameter, near

which the effective shear modulus is affected significantly. Numerical results show

that the effective shear moduli of nano-composites and nano-porous materials can

be greatly improved by an appropriate surface modification. Finally, the derived

effective modulus with the Steigmann–Ogden interface model is provided in the

supplemental MATLAB code, which can be easily executed, and used as a

benchmark for semi-analytical solutions and numerical solutions in future studies.

Keywords: Steigmann–Ogden interface model; effective modulus; Papkovich-Neuber

solution; Eshelby tensors; spherical nano-inhomogeneity

1. Introduction

In recent decades, the interest in modelling surfaces and interfaces is growing with

the wide applications of nano-materials in mechanical, automobile, and aerospace

industries. Various models, such as the free sliding model [1], the linear spring model

[2], the dislocation-like model [3], the interphase model [4], the interface stress model

[5-9] etc., are developed to simulate the mechanical properties of interfaces in nano-

materials. Among these models, the Steigmann–Ogden interface stress model [8, 9]

(hereinafter referred to as the S−O model) has enjoyed an increasing popularity. As both

stretching resistance and the bending resistance are incorporated into the

surface/interface constitutive relation, the S−O model can account for the known

2 of 34 pages
experimental observations and simulation results on the size-dependent surface stresses

of nanowires [10, 11], nanoplates [12] and nanoparticles [13], which cannot be fully

explained by the Gurtin–Murdoch model [6, 7].

The Steigmann–Ogden interface stress model was first put forward by Steigmann

and Ogden[8] in 1997, and has recently attracted increased interest since Eremeyev and

Lebedev[14, 15] derived equilibrium equations and boundary conditions describing an

elastic solid with surface stresses. The interface can be regarded as a negligibly thin

shell attached to the surface/interface of the bulk materials in the S–O model, while in

the G–M model the surface/interface is regarded as a membrane capable only of

stretching (no flexural resistance) leading to the possibility of instability under

compressive surface/interface stresses (e.g. wrinkling) [16]. The S–O model is thus

recognized as a advancement in the field of surface mechanics [14-17], and has been

widely used for mechanical analysis of nano-materials, e.g. nanobeams [18], nanowires

[19], nanoshells [20], polymer brush [21, 22] and half-space material [23, 24].

In contrast to the large number of studies available for materials with the Gurtin-

Murdoch interface model (e.g. [25-39] and many others), the literature on nano-porous

materials and nano-particle reinforced composites, considering the Steigmann–Ogden

model, is rather limited. The only papers we are aware of are those by Dai et al. [40],

Gharahi and Schiavone [16], Han et al. [41] , Zemlyanova and Mogilevskaya [42].

Among these studies, Dai et al. [40] and Zemlyanova and Mogilevskaya [42] presented

analytical/ semi-analytical solutions to the two-dimensional problem of an infinite

isotropic elastic domain containing an isotropic elastic circular inhomogeneity; Gharahi

3 of 34 pages
and Schiavone [16] derived the effective moduli of the micropolar nano-composite

considering plane elasticity.

In our previous study [17], we presented an explicit solution for the problem of a

spherical nano-inhomogeneity embedded in an infinite matrix loaded by uniform far-

field-stresses. It was shown that the existence of interface bending resistance can

significantly change the local stress distributions around the interface. However, due to

the mathematical complexity, studies of the effective properties of 3D nano-composites

with the Steigmann–Ogden interface model have not been reported to the best of our

knowledge. Especially, the explicit expressions of the Eshelby tensor and overall

properties of nano-composites and nano-porous materials considering the Steigmann–

Ogden interface model, which can be quite useful in the designing nanocomposites and

porous materials, are very desirable.

Following our previous study [17], the Eshelby formalism is extended to the problem

of a nano-inclusion with the Steigmann–Ogden interface model for the first time in this

study. Using the obtained stress/strain concentration tensors, we employ the dilute

approximation and the Mori-Tanakas method to derive the effective elastic moduli of

nano-composites following the procedure given in Duan et al. [30]. The derived

formulas can also be used for nano-porous materials by setting the moduli of the

inclusion to be 0. An equivalent interface curvature parameter, and a characteristic

curvature parameter are also defined, and their significances are discussed in detail.

The rest of this paper is organized as follows: In Section 2, the governing equations

for the 3D nano-inhomogeneity with Steigmann–Ogden interface are briefly stated. In

4 of 34 pages
Section 3, explicit analytical solutions and Eshelby tensors for an inclusion subjected

to eigenstrains are given. In Section 4, the explicit solutions and stress/strain

concentration tensors for a spherical inhomogeneity with far-field stresses/strains are

given. In Section 5, the expressions and numerical examples of the effective elastic

moduli of nano-composites, considering the Steigmann–Ogden interface model, are

given. In Section 6, we complete this paper with some concluding remarks.

2. The governing linear elasticity equations

We start by considering an infinite matrix with a nanosized inclusion/inhomogeneity

illustrated in Fig. 1. Solutions for the matrix and the inhomogeneity should satisfy the

equations of stress equilibrium, strain displacement-gradient compatibility, constitutive

relations, as well as far-field boundary conditions.

Fig. 1. A nanosized spherical inhomogeneity embedded in an infinite matrix

The Steigmann–Ogden interface model is employed to consider the interface bending

resistance as well as interface stretching resistance. Derivation of the Steigmann–

5 of 34 pages
 Ogden interface model is detailed in [14, 15], and the governing equations are

summarized below. The displacement vector across the interface is continuous,

u m ui at Γ
= (1)

The stress tensor across the interface has a jump,

n ⋅ ∆σ =∇s ⋅  τ s + ( ∇s ⋅ m s ) n  − ( ∇s ⋅ n ) n ⋅ ( ∇s ⋅ m s ) n at Γ (2)

The constitutive equation of the interface is

τ s 2μ sε s + λ s tr ( ε s ) I s
= (3)

m s 2χ sκ s + ζ s tr ( κ s ) I s
= (4)

where

1
ε s = ∇sus ⋅ I s + I s ⋅ (∇sus )T  (5)
2

1
κ s =− ∇sϑ ⋅ I s + I s ⋅ (∇sϑ )T  (6)
2

∇s ( n ⋅ us ) + B ⋅ us
ϑ= (7)

B = −∇sn (8)

in which us , ε s , τ s , κ s and m s are the interface fields of displacement, strain, stress,

curvature and bending moment, respectively. λ s and μ s are the stiffness parameters

characterizing the interface stretching. χ s and ζ s are the stiffness parameters

characterizing the interface bending. I s is the unit tangent tensor defined on the

interface and I s = eθ ⊗ eθ + eϕ ⊗ eϕ in spherical coordinates. ∇s = (I 3 − nn) ⋅∇ is the

gradient operator defined on the interface where n is the unit outer-normal vector of

the interface Γ .

6 of 34 pages
3. Eshelby tensor for a nano-inclusion

In this section, we consider a spherical inclusion of radius R embedded in an

infinite matrix, along with an interface characterized by the S–O model. The elastic

modulus is the same for both the inclusion and the matrix ( E m = E i ). The eigenstrains

ε* in the inclusion are assumed to be uniform. For this problem, the analytical solution

can be derived easily by employing Papkovich–Neuber solutions [43, 44]. We express

the analytical solution for this problem as being a linear combination of Papkovich–

Neuber potentials, and determine the unknown coefficients by enforcing the far-field

and interface conditions. Details of the derivation can be found in Wang et al. [17].

3.1. Analytical solutions with uniform eigenstrains

First, we consider the case that the eigenstrain tensor ε* has only one non-zero

component ε*xx . The displacement components of u xxj (here we use uαβj to denote the

displacement vector when the eigenstrain has only one non-zero component εαβ
*
) can

be simplified as:

m
u=
rxx
M1xx
r 2
8r
1
( )
+ 4 ( 3M 3 xx + 2M 2 xx r 2 ( −5 + 4v m ) ) 1 + 3cos [ 2θ ] − 6cos [ 2ϕ ] sin [θ ] ;
2

3 ( M 3 xx + 2M 2 xx r 2 (1 − 2v m ) )
cos [θ ] cos [ϕ ] sin [θ ];
2
uθmxx = 4
r
3 ( M 3 xx + 2M 2 xx r 2 (1 − 2v m ) )
m
uϕ xx = − cos [ϕ ] sin [θ ] sin [ϕ ]; (9)
r4
u irxx =
1
( )
C1xx r − r ( C3 xx + 6C2 xx r 2 v m ) 1 + 3cos [ 2θ ] − 6cos [ 2ϕ ] sin [θ ] ;
4
2

uθi xx = 3 ( C3 xx r + C2 xx r 3 ( 7 − 4v m ) ) cos [θ ] cos [ϕ ] sin [θ ];

2

−3 ( C3 xx r + C2 xx r 3 ( 7 − 4v m ) ) cos [ϕ ] sin [θ ] sin [ϕ ];

uϕi xx =

By using the same procedure, the displacement field u yyj with eigenstrain ε*yy can

7 of 34 pages
 be written as:

M1 yy
m
u=
ryy
r 2
+
1
8r 4 ( ( )
3M 3 yy + 2M 2 yy r 2 ( −5 + 4v m ) ) 1 + 3cos [ 2θ ] + 6cos [ 2ϕ ] sin [θ ] ;
2

3 ( M 3 yy + 2M 2 yy r 2 (1 − 2v m ) ) cos [θ ] sin [θ ] sin [ϕ ]

2
m
u θ yy = ;
r4
3 ( M 3 yy + 2M 2 yy r 2 (1 − 2v m ) ) cos [ϕ ] sin [θ ] sin [ϕ ]
uϕmyy = ; (10)
r4
1
(
C1 yy r − r ( C3 yy + 6C2 yy r 2 v m ) 1 + 3cos [ 2θ ] + 6cos [ 2ϕ ] sin [θ ] ;
u iryy =
4
2
)
uθi yy 3 ( C3 yy r + C2 yy r 3 ( 7 − 4v m ) ) cos [θ ] sin [θ ] sin [ϕ ] ;
2
=
uϕi yy =3 ( C3 yy r + C2 yy r 3 ( 7 − 4v m ) ) cos [ϕ ] sin [θ ] sin [ϕ ];

The displacement components of u zzj with eigenstrain ε*zz can be written as:

M1zz ( 3M 3 zz + 2M 2 zz r ( −5 + 4v m ) ) (1 + 3cos [ 2θ ])
2

u mrzz
= − ;
r2 4r 4
3 ( M 3 zz + 2M 2 zz r 2 (1 − 2v m ) ) cos [θ ] sin [θ ]
uθmzz = − ;
r4
uϕmzz = 0; (11)
1
C1zz r + r ( C3 zz + 6C2 zz r 2 v m ) (1 + 3cos [ 2θ ]) ;
u irzz =
2
i
uθ zz =−3 ( C3 zz r + C2 zz r 3 ( 7 − 4v m ) ) cos [θ ] sin [θ ];
uϕi zz = 0;

The displacement field u xyj with eigenstrain ε*xy can be written as:

3 ( 3M 3 xy + 2M 2 xy r 2 ( −5 + 4v m ) ) sin [θ ] sin [ 2ϕ ]
2
m
M1xy
u
= rxy − ;
r2 2r 4
3 ( M 3 xy + 2M 2 xy r 2 (1 − 2v m ) ) sin [ 2θ ] sin [ 2ϕ ]
uθmxy = ;
2r 4
3 ( M 3 xy + 2M 2 xy r 2 (1 − 2v m ) ) cos [ 2ϕ ] sin [θ ]
uϕmxy = ; (12)
r4
(
u irxy = r C1xy + 3 ( C3 xy + 6C2 xy r 2 v m ) sin [θ ] sin [ 2ϕ ] ;
2
)
3
uθi xy = r ( C3 xy + C 2 xy r 2 ( 7 − 4v m ) ) sin [ 2θ ] sin [ 2ϕ ];
2
uϕ xy =3 ( C3 xy r + C2 xy r 3 ( 7 − 4v m ) ) cos [ 2ϕ ] sin [θ ];
i

The displacement field u yzj with eigenstrain ε*yz can be written as:

8 of 34 pages
 m
M1 yz 3 ( 3M 3 yz + 2M 2 yz r 2 ( −5 + 4v m ) ) sin [ 2θ ] sin [ϕ ]
u
= ryz − ;
r2 2r 4
3 ( M 3 yz + 2M 2 yz r 2 (1 − 2v m ) ) cos [ 2θ ] sin [ϕ ]
uθmyz = ;
r4
3 ( M 3 yz + 2M 2 yz r 2 (1 − 2v m ) ) cos [θ ] cos [ϕ ] (13)
uϕmyz = ;
r4
(
u iryz = r C1 yz + 3 ( C3 yz + 6C2 yz r 2 v m ) sin [ 2θ ] sin [ϕ ] ; )
uθi yz =3 ( C3 yz r + C2 yz r 3 ( 7 − 4v m ) ) cos [ 2θ ] sin [ϕ ];
uϕi yz = 3 ( C3 yz r + C2 yz r 3 ( 7 − 4v m ) ) cos [θ ] cos [ϕ ];

The displacement field u zxj with eigenstrain ε*zx can be written as:

M1zx 3 ( 3M 3 zx + 2M 2 zx r ( −5 + 4v m ) ) cos [ϕ ] sin [ 2θ ]

2

u mrzx
= − ;
r2 2r 4
3 ( M 3 zx + 2M 2 zx r 2 (1 − 2v m ) ) cos [ 2θ ] cos [ϕ ]
uθmzx = ;
r4
3 ( M 3 zx + 2M 2 zx r 2 (1 − 2v m ) ) cos [θ ] sin [ϕ ] (14)
uϕmzx = − ;
r4
(
u irzx = r C1zx + 3 ( C3 zx + 6C2 zx r 2 v m ) cos [ϕ ] sin [ 2θ ] ; )
uθi zx =3 ( C3 zx r + C2 zx r 3 ( 7 − 4v m ) ) cos [ 2θ ] cos [ϕ ];
uϕi zx = −3 ( C3 zx r + C2 zx r 3 ( 7 − 4v m ) ) cos [θ ] sin [ϕ ];

where
= M pαβ ( p 1,...,5
= and α , β x, y, z ) and
= 2,3 and α , β x, y, z )
Cqαβ (q 1,= are

constants given in the Appendix.

On the above, we have obtained the basic solutions for a spherical inclusion with six

different eigenstrains. Thus, the analytical solution with an arbitrary uniform

eigenstrain ε* can be expressed as an additive combination of Eqs.(9-14):

u j = u xxj + u yyj + u zzj + u xyj + u yzj + u zxj j = m,i (15)

Using strain displacement-gradient compatibility and the constitutive relations, the

strain/stress fields can be obtained.

9 of 34 pages
 3.2. Eshelby tensors for the matrix and the inclusion

Using the obtained solutions in the previous subsection, we can calculate the Eshelby

tensor. Unlike the classical interior Eshelby tensor, the Eshelby tensor in the case of

the nano-inclusion is no longer uniform due to interface stress effects. Using the

Walpole notation [45], a fourth-order tensor Sk (r ) with radial symmetry can be

expressed as

Sk (r ) = S1k (r )E1 + S2k (r )E2 + S3k (r )E3 + Sk4 (r )E4 + S5k (r )E5 + S6k (r )E6 (16)

where Ei are elementary tensors introduced by Walpole [45]:

1
E1ijkl = bij b kl
2
2
E ijkl = a ij a kl
1
=E 3ijkl
2
( bik b jl +b jk bil − bij bkl ) (17)
1
4
E ijkl
=
2
( bik a jl +bil a jk + b jl a ik + b jk a il )
E 5ijkl = a ij b kl
6
E ijkl = bij a kl

where a ij and bij are components of a= e r ⊗ e r and b = I 2 − e r ⊗ e r , respectively.

In the inclusion, the Eshelby tensor can be written as:

S1i = 2C1 + C3 + 3C2 r 2 ( 7 − 8v m )

Si2 =C1 + 2C3 + 36C2 r 2 v m
Si3 =3 ( C3 + C2 r 2 ( 7 − 4v m ) )
(18)
Si4 =3 ( C3 + C2 r 2 ( 7 + 2v m ) )
S5i = C1 − C3 − 18C2 r 2 v m
Si6 = C1 − C3 + 3C2 r 2 ( −7 + 8v m )

In the matrix, the Eshelby tensor can be written as:

10 of 34 pages
 m
S =
(
2 3M 3 + r 2 ( M1 − 2M 2 (1 + v m ) ) )
1 5
r

Sm2 = −
(
2 −6M 3 + r ( M1 + 2M 2 ( 5 − 4v m ) )
2
)
5
r
3 ( M 3 + 2M 2 r (1 − 2v m ) ) 2

S3m =
r5 (19)
6 ( −2M 3 + M 2 r 2 (1 + v m ) )
Sm4 =
r5
m
S = −
(
2 3M 3 + r 2 ( M1 + M 2 ( −5 + 4v m ) ) )
5 5
r
−6M 3 + r ( M1 + 4M 2 (1 + v m ) )
2

S6m =
r5

where C1 ,C2 ,C3 ,M1 ,M 2 ,M 3 are constants given in the Appendix.

The average Eshelby tensor in the inclusion is thus defined as:

1
Si =
Vi ∫ Vi
S i dV (20)

in which Vi is the volume of the inclusion. After some derivations, it can be shown

that S i is an isotropic tensor:

 63C2 R 2 
S i = 3C1J +  3C3 + K
 5  (21)
1
J =I 2 ⊗ I 2 , K = I 4s − J
3

The solution (Eqs.(16-21)) reveals that the Eshelby tensor is not uniform and is

affected by the interface properties. If the interface stress is neglected, Eqs.(16-21)

degenerate into the classical Eshelby tensor without interface stress effects, and the size

effect will no longer exist. If the surface bending resistance is neglected ( χ s = 0 and

ζ s = 0 ), Eqs.(16-21) degenerate into the Eshelby tensor considering the Gurtin-

Murdoch interface model as given in Duan et al. [30].

11 of 34 pages
4. Stress/strain concentration tensors for a nano-inhomogeneity

In this section, we consider a spherical inhomogeneity of radius R embedded in an

infinite matrix, with the S-O interface, subjected to homogeneous far-field stresses and

strains at infinity. The far-field condition for uniform stresses and strains can be written

as:

σm = Σ at infinity (22)

εm = Ε at infinity (23)

where Σ and Ε are constant far-field strain and stress tensors. It should be pointed

out that the equivalent inclusion method is not applicable to this problem, because the

strain field in the inhomogeneity is not uniform. We will discuss these two far-field

conditions in the following two subsections separately. The detailed derivation of the

displacement fields for this problem is similar to Wang et al. [17], thus we simply list

the results in the present paper.

4.1. Stress concentration tensors

First, we consider the case when the remote loading has only one non-zero stress

component Σ xx . The displacement field u xxj (here we use uαβj to denote the

displacement vector when remote loading has only one non-zero component Σαβ ) can

be simplified as:

12 of 34 pages
 M Σ2xx 1  3M 5Σxx 2M Σ4 xx ( 5 − 4v m ) 
u mrxx= + M Σ
r −  − + 2M Σ
r + 
8 
1xx 3 xx
r2 r4 r2 
(1 + 3cos [2θ ] − 6cos [2ϕ ]sin [θ ] ) ; 2

3
4 (
M 5Σxx + M 3Σxx r 5 + M 4Σxx r 2 ( 2 − 4v m ) ) cos [θ ] cos [ϕ ] sin [θ ];
2
uθmxx =
r
3
uϕmxx =− 4 ( M 5Σxx + M 3Σxx r 5 + M 4Σxx r 2 ( 2 − 4v m ) ) cos [ϕ ] sin [θ ] sin [ϕ ]; (24)
r
1
(
u irxx C1Σxx r − r ( C3Σxx + 6CΣ2 xx r 2 vi ) 1 + 3cos [ 2θ ] − 6cos [ 2ϕ ] sin [θ ] ;
=
4
)
2

uθi xx =3 ( C3Σxx r + CΣ2 xx r 3 ( 7 − 4vi ) ) cos [θ ] cos [ϕ ] sin [θ ];

2

−3 ( C3Σxx r + CΣ2 xx r 3 ( 7 − 4vi ) ) cos [ϕ ] sin [θ ] sin [ϕ ];

uϕi xx =

The displacement field u yyj with the remote tensile stress Σ yy can be written as:

M Σ2 yy 1  3M 5 yy
Σ
2M 4Σ yy ( 5 − 4v m ) 
u mryy= + M1Σyy r −  − 4 + 2M 3Σyy r + 
r2 8  r r2 
(1 + 3cos [2θ ] + 6cos [2ϕ ]sin [θ ] ) ; 2

3
uθmyy =
r4
( M5Σyy + M3Σyy r 5 + M 4Σyy r 2 ( 2 − 4vm ) ) cos [θ ]sin [θ ]sin [ϕ ]2 ;
3
uϕmyy = 4 ( M 5Σyy + M 3Σyy r 5 + M 4Σ yy r 2 ( 2 − 4v m ) ) cos [ϕ ] sin [θ ] sin [ϕ ]; (25)
r
1
(
u iryy C1Σyy r − r ( C3Σyy + 6CΣ2 yy r 2 vi ) 1 + 3cos [ 2θ ] + 6cos [ 2ϕ ] sin [θ ] ;
=
4
)2

uθi yy =3 ( C3Σyy r + CΣ2 yy r 3 ( 7 − 4vi ) ) cos [θ ] sin [θ ] sin [ϕ ] ;

2

uϕi yy =3 ( C3Σyy r + CΣ2 yy r 3 ( 7 − 4vi ) ) cos [ϕ ] sin [θ ] sin [ϕ ];

The displacement field u zzj with the remote tensile stress Σ zz can be written as:

M Σ2 zz 1  3M 5Σzz 2M Σ4 zz ( 5 − 4v m ) 
 (1 + 3cos [ 2θ ]) ;
m Σ Σ
u rzz= + M r +  − + 2M r +
4 
1 zz 3 zz
r2 r4 r2 
3
uθmzz = − 4 ( M 5Σzz + M 3Σzz r 5 + M Σ4 zz r 2 ( 2 − 4v m ) ) cos [θ ] sin [θ ];
r
uϕmzz = 0; (26)
1
i
urzz =C1Σzz r + r ( C3Σzz + 6CΣ2 zz r 2 vi ) (1 + 3cos [ 2θ ]) ;
2
i
−3 ( C3 zz r + C2Σzz r 3 ( 7 − 4vi ) ) cos [θ ] sin [θ ];
uθ zz = Σ

uϕi zz = 0;

The displacement field u xyj with the remote shear stress Σ xy can be written as:

13 of 34 pages
 M Σ2 xy 3  3M 5 xy
Σ
2M 4Σxy ( 5 − 4v m ) 
 sin [θ ] sin [ 2ϕ ];
m Σ 2
u = +  − 4 + 2M 3 xy r +
rxy
r2 2 r r 2

3
4 (
uθmxy= M 5Σxy + M 3Σxy r 5 + M Σ4 xy r 2 ( 2 − 4v m ) ) sin [ 2θ ] sin [ 2ϕ ];
2r
3
uϕmxy = 4 ( M 5Σxy + M 3Σxy r 5 + M Σ4 xy r 2 ( 2 − 4v m ) ) cos [ 2ϕ ] sin [θ ]; (27)
r
(
u irxy = r C1Σxy + 3 ( C3Σxy + 6CΣ2 xy r 2 vi ) sin [θ ] sin [ 2ϕ ] ;
2
)
3
uθi xy = r ( C3Σxy + CΣ2 xy r 2 ( 7 − 4vi ) ) sin [ 2θ ] sin [ 2ϕ ];
2
uϕ xy =3 ( C3Σxy r + C2Σxy r 3 ( 7 − 4vi ) ) cos [ 2ϕ ] sin [θ ];
i

The displacement field u yzj with the remote shear stress Σ yz can be written as:

M Σ2 yz 3  3M 5 yz
Σ
2M Σ4 yz ( 5 − 4v m ) 
 sin [ 2θ ] sin [ϕ ];
m Σ
u = +  − 4 + 2M 3 yz r +
2 
ryz
r2 r r2 
3
uθmyz =
r4
( M5Σyz + M3Σyz r 5 + M 4Σyz r 2 ( 2 − 4vm ) ) cos [ 2θ ]sin [ϕ ];
3
uϕmyz = 4 ( M 5Σyz + M 3Σyz r 5 + M 4Σ yz r 2 ( 2 − 4v m ) ) cos [θ ] cos [ϕ ]; (28)
r
(
u iryz = r C1Σyz + 3 ( C3Σyz + 6CΣ2 yz r 2 vi ) sin [ 2θ ] sin [ϕ ] ; )
uθi yz = 3 ( C3Σyz r + CΣ2 yz r 3 ( 7 − 4vi ) ) cos [ 2θ ] sin [ϕ ];
uϕi yz =3 ( C3Σyz r + CΣ2 yz r 3 ( 7 − 4vi ) ) cos [θ ] cos [ϕ ];

The displacement field u zxj with the remote shear stress Σ xy can be written as:

M Σ2 zx 3  3M 5Σzx 2M Σ4 zx ( 5 − 4v m ) 
 cos [ϕ ] sin [ 2θ ];
m Σ
u =
rzx +  − 4 + 2M 3 zx r +
r2 2 r r2 
3
uθmzx = 4 ( M 5Σzx + M 3Σzx r 5 + M 4Σzx r 2 ( 2 − 4v m ) ) cos [ 2θ ] cos [ϕ ];
r
3
uϕmzx = − 4 ( M 5Σzx + M 3Σzx r 5 + M 4Σzx r 2 ( 2 − 4v m ) ) cos [θ ] sin [ϕ ]; (29)
r
(
u irzx = r C1Σzx + 3 ( C3Σzx + 6CΣ2 zx r 2 vi ) cos [ϕ ] sin [ 2θ ] ; )
uθi zx =3 ( C3Σzx r + CΣ2 zx r 3 ( 7 − 4vi ) ) cos [ 2θ ] cos [ϕ ];
−3 ( C3Σzx r + CΣ2 zx r 3 ( 7 − 4vi ) ) cos [θ ] sin [ϕ ];
uϕi zx =

where
= M Σpαβ ( p 1,...,5
= and α , β x, y, z ) and
= 2,3 and α , β x, y, z )
CΣqαβ (q 1,= are

constants given in the Appendix. Now we have obtained the basic solutions for a

spherical inhomogeneity under six different remote loading cases. Thus, the analytical

14 of 34 pages
 solution under remote loading Σ can be written as:

u j = u xxj + u yyj + u zzj + u xyj + u yzj + u zxj j = m,i (30)

Using strain displacement compatibility and the constitutive relations, the

strain/stress fields can be obtained. Once the stress fields in the inhomogeneity and the

matrix are derived, we can derive the stress concentration tensor for the inhomogeneity

and the matrix considering the S-O model. The stress concentration tensors for the

spherical inhomogeneity can be expressed as:

B k (r ) = B1k (r )E1 + Bk2 (r )E2 + B3k (r )E3 + Bk4 (r )E4 + B5k (r )E5 + B6k (r )E6 (31)

In the inhomogeneity, the stress concentration tensor can be written as:

 Σ 2C1Σ (1 + vi ) 
B= 6C r ( 7 + 6vi ) μ i + 2  C3 −
i Σ 2
 μi ;
−1 + 2vi 
1 2

2CΣ (1 + vi ) μ i
4C3Σ μ i − 12CΣ2 r 2 vi μ i − 1
Bi2 = ;
−1 + 2vi
6C3Σ μ i + 6CΣ2 r 2 ( 7 − 4vi ) μ i ;
Bi3 =
(32)
6C3Σ μ i + 6CΣ2 r 2 ( 7 + 2vi ) μ i ;
Bi4 =
 CΣ (1 + vi ) 
=Bi5 6CΣ2 r 2 vi μ i + 2  −C3Σ + 1  μi ;
 1 − 2vi 
2CΣ (1 + vi ) μ i
−2C3Σ μ i − 1
Bi6 = − 6CΣ2 r 2 ( 7 + 6vi ) μ i ;
−1 + 2vi

In the matrix, the stress concentration tensor can be written as:

15 of 34 pages
 12M 5Σ μ m 4 ( M 2 − 2M 4 + 4M 4 v m ) μ m 2 ( M 3 − 2M 3 v m + 2M1 (1 + v m ) ) μ m
Σ Σ Σ Σ Σ Σ

B1m = + −
r5 r3 −1 + 2vm
24M 5μ m 4 ( M 2 − 2M 4 ( −5 + v m ) ) μ m 2 ( 2M 3 (1 − 2v m ) + M1 (1 + v m ) ) μ m
Bm2 = − −
r5 r3 −1 + 2v m
6M 5Σ μ m 6M 4 ( 2 − 4v m ) μ m
Σ

B3m = 6M 3Σ μ m + +
r5 r3
(33)
24M 5Σ μ m 12M 4 (1 + v m ) μ m
Σ

Bm4 =6M 3Σ μ m − +
r5 r3
12M 5Σ μ m 4 ( M 2 + M 4 ( −5 + v m ) ) μ m 2 ( M1 − M 3 + M1 v m + 2M 3 v m ) μ m
Σ Σ Σ Σ Σ Σ
m
B5 = − − −
r5 r3 −1 + 2v m
12M 5Σ μ m 2 ( M 2 + M 4 ( 4 − 8v m ) ) μ m 2 ( M1 − M 3 + M1 v m + 2M 3 v m ) μ m
Σ Σ Σ Σ Σ Σ
m
B =
−
6 + −
r5 r3 −1 + 2 v m

where C1Σ ,CΣ2 ,C3Σ ,M1Σ ,M Σ2 ,M 3Σ ,M Σ4 ,M 5Σ are constants given in the Appendix.

The average stress concentration tensor in the inhomogeneity is defined as:

1
Bi =
Vi ∫ Vi
Bi dV (34)

Again after some manipulations, it can be shown that B is an isotropic tensor:

Bi = B1i J + Bi2 K
6C1Σ (1 + vi ) μ i
B1i = − (35)
−1 + 2vi
6
Bi2 =
5
( 5C3Σ + 21CΣ2 R 2 ) μ i
1
where =
J I 2 ⊗ I 2 and K
= I 4s − J . I 2 is the second-order identity tensor and I 4s is
3

the fourth-order symmetric identity tensor. As seen from Eq.(35), the average stress

concentration tensor in the inhomogeneity is an isotropic tensor of fourth order.

In order to derive the effective compliance tensor in the next section, we introduce

the interface stress concentration tensors, which is defined as below:

1
∫ (σ − σ i ) ⋅ n ⊗ rdS= B Γ : Σ (36)
m

Vi Γ

where

16 of 34 pages
B Γ = B1Γ J +BΓ2 K
6C1Σ (1 + vi ) μ i 6 ( M1Σ R 3 (1 + v m ) + M Σ2 ( −2 + 4v m ) ) μ m
B =
Γ
− (37)
R 3 ( −1 + 2v m )
1
−1 + 2vi
6 2M Σ4 ( −7 + 5v m ) μ m 
BΓ2 =  −5C3Σ μ i − 21C2Σ R 2μ i + 5M 3Σ μ m + 
5 R3 

In addition to size-dependency, two interesting phenomena are observed from the

closed form Eqs.(24-37). First, displacements and stress concentration tensors depend

on the interface curvature parameter:

ηs = 3ζ s + 5χ s (38)

that is to say, they are not affected individually by ζ s and χ s when ηs is a fixed value.

Second, the stress concentration tensor in the inhomogeneity is transversely isotropic,

while the average stress concentration tensor is an isotropic tensor.

4.2. Strain concentration tensors

The analytical solutions with far-field strain loading have the same structure as the

analytical solutions with far-field stresses, except that the coefficients are different. The

coefficients
= M Εpαβ ( p 1,...,5
= and CΕqαβ (q 1,=
and α , β x, y, z ) = 2,3 and α , β x, y, z ) for

uniform strain far-field condition can be found in the Appendix.

The strain concentration tensors for the spherical inhomogeneity can be expressed as:

A k (r ) = A1k (r )E1 + A 2k (r )E2 + A 3k (r )E3 + A 4k (r )E4 + A 5k (r )E5 + A 6k (r )E6 (39)

In the inhomogeneity, the strain concentration tensor can be written as:

17 of 34 pages
A1i = 2C1Ε + C3Ε + 3CΕ2 r 2 ( 7 − 8vi ) ;
A i2 =C1Ε + 2C3Ε + 36CΕ2 r 2 vi ;
3C3Ε + 3CΕ2 r 2 ( 7 − 4vi ) ;
A i3 =
(40)
3C3Ε + 3C2Ε r 2 ( 7 + 2vi ) ;
A i4 =
A i5 = C1Ε − C3Ε − 18CΕ2 r 2 vi ;
A i6= C1Ε − C3Ε + 3CΕ2 r 2 ( −7 + 8vi ) ;

In the matrix, the strain concentration tensor can be written as:

6M Ε 2M 2 − 4M 4 (1 + v m )
Ε Ε
m Ε Ε
A 1 = 2M + M + 5 5 +
1 3
r r3
12M Ε 2 ( M 2 + 2M 4 ( 5 − 4v m ) )
Ε Ε

A m2 =M1Ε + 2M 3Ε + 5 5 −
r r3
3M Ε 3M 4 ( 2 − 4v m )
Ε

A 3m = 3M 3Ε + 5 5 +
r r3 (41)
12M Ε 6M 4 (1 + v m )
Ε
m
A 4 =3M 3Ε − 5 5 +
r r3
6M 5Ε 2 ( M 2 + M 4 ( −5 + 4v m ) )
Ε Ε

A 5m Ε Ε
= M1 − M 3 − 5 −
r r3
6M Ε M + 4M 4 (1 + v m )
Ε Ε

A 6m = M1Ε − M 3Ε − 5 5 + 2
r r3

where C1Ε ,CΕ2 ,C3Ε ,M1Ε ,M Ε2 ,M 3Ε ,M Ε4 ,M 5Ε are constants given in the Appendix.

The average strain concentration tensor in the inhomogeneity is defined as:

1
Ai =
Vi ∫ Vi
A i dV (42)

in which Vi is the volume of the inhomogeneity. Again, Ai is an isotropic tensor:

A i = A1i J + A i2 K
A1i = 3C1Ε (43)
63CΕ2 R 2
A i2 = 3C3Ε +
5

We also introduce the “interface strain concentration tensors”[31], which is defined

as follows:

1
∫ (σ − σi ) ⋅ n ⊗ r =AΓ : Ε (44)
m

Vi Γ

18 of 34 pages
where

A Γ = A1Γ J +A Γ2 K
6C1Ε (1 + vi ) μ i 12M Ε2 μ m 6M1 (1 + v m ) μ m
Ε

A1Γ = − − (45)
−1 + 2vi R3 −1 + 2v m
Ε126 Ε 2 12M 4 ( −7 + 5v m ) μ m
Ε

A = −6C μ i −
Γ
2 3 C2 R μ i + 6M 3Ε μ m +
5 5R 3

Similar to the solution for far-field stresses, the solutions in Eqs.(39-45) depend on

the interface curvature parameter ηs = 3ζ s + 5χ s . Besides, the strain concentration

tensor in the inhomogeneity is transversely isotropic, while the average strain

concentration tensor is an isotropic tensor.

5. Effective bulk and shear moduli

5.1. The dilute approximation and the Mori-Tanaka method

The effective stiffness tensor Lhom and compliance tensor M hom are respectively

defined as

σ = Lhom : ε
(46)
ε = M hom : σ

where σ , ε denote the average strain and stress tensors. It should be noted that the

definition of the average strain is the same as the traditional one, while the definition

of the average stress should consider the stress-jump across the interface [31]:

ε=
(1 − f ) εm + f εi
f (47)
σ = (1 − f )σ m + f σ i +
Vi ∫
Vi
(σ m − σ i ) ⋅ n ⊗ r dS

where ε j and σ j denote volume averages of the strain and stress in the matrix/

19 of 34 pages
inhomogeneity. f denotes the volume fraction of inhomogeneity. For a two-phase

model, Lhom and M hom are given in Duan et al. [31]:

 Lhom =L m + f ( L i − L m ) : A i + fA Γ
(48)
M hom =M m + f (M i − M m ) : B i − f B Γ

where L j and M j are the stiffness and compliance tensors of the two phases. Ai ,

Bi , A Γ and B Γ are defined as:

ε i = Ai : ε
σ i = Bi : σ
1
∫ (σ m − σ i ) ⋅ n ⊗ r dS =AΓ : ε (49)
Vi Vi

1
Vi ∫Vi
(σ m − σ i ) ⋅ n ⊗ r dS =BΓ : σ

Thus, once the concentration tensors Ai , Bi , A Γ and B Γ are determined, the effective

stiffness tensor Lhom and the effective compliance tensor M hom can be easily

calculated.

The simplest method is the “dilute” approximation [46], in which A i , Bi , A Γ and

BΓ are approximated by considering the case of embedding a single nano-

inhomogeneity in an all-matrix medium. Thus, we have

Α i = Α i
Βi = Βi
(50)
Α Γ = Α Γ
ΒΓ = ΒΓ

where Αi , Βi , Α Γ and Β Γ are given in section 4. Substituting Eq.(50) into Eq.(48), we

can express the effective stiffness tensor Lhom and compliance tensor M hom as:

 Lhom =L m + f ( L i − L m ) : Α i + fΑ Γ
(51)
M hom =M m + f (M i − M m ) : Β i − f Β Γ

Expressing Lhom as 3k hom

DA J + 2μ DA K , we can obtain the effective bulk and shear
hom

modulus:

20 of 34 pages
 1
k hom
DA =
3
( 3k m + A Γ1 f + f ( 3k i − 3k m ) A i1 )
(52)
1
μ hom
DA = ( 2μ m + A Γ2 f + f ( 2μ i − 2μ m ) A i2 )
2

As discussed in[46], the dilute approximation neglects particle interactions and is

therefore valid only for small volume fractions of reinforcements. In order to predict

overall mechanical properties of nano-composites with higher volume fractions, we

employ the Mori-Tanaka method (MTM) [47]. Following the procedure given in [48],

we have the following relations:

ε i = Αi : ε m
σ i = Bi : σ m
1
(53)
Vi ∫Vi
(σ m − σ i ) ⋅ n ⊗ r dS =ΑΓ : ε m

1
Vi ∫
Vi
(σ m − σ i ) ⋅ n ⊗ r dS =BΓ : σ m

Substituting Eq.(53) into Eq.(47), we can express Ai , Bi , A Γ and B Γ in terms of Αi ,

Βi , Α Γ and Β Γ as follows:

Α i = Α i : I 2 + f ( Α i − I 2 ) 
−1

Βi = Βi : I 2 + f ( Βi + Β Γ − I 2 ) 
−1

(54)
A Γ= Α : I 2 + f ( Α i − I 2 ) 
−1
Γ

Β Γ : I 2 + f ( Βi + Β Γ − I 2 ) 
−1
B
= Γ

Substituting Eq.(54) into Eq.(48), Lhom and compliance tensor M hom can be written

as:

L m + f (Li − L m ) : Α i + Α Γ  : I 2 + f ( Α i − I 2 ) 
−1
 Lhom =
(55)
= M m + f (M i − M m ) : Β − M m : Β  : I 2 + f ( Β + Β − I 2 ) 
−1
M hom i Γ i Γ

And we can obtain the effective bulk and shear modulus:

21 of 34 pages
 hom
f ( A Γ1 + 3k i A i1 − 3k m ) + 3k m
k =
3 + 3 f ( −1 + A i1 )
MTM

(56)
f ( A Γ2 + 2A i2 μ i − 2μ m ) + 2μ m
μ hom
MTM =
2 + 2 f ( −1 + A i2 )

Again, in addition to size-dependency, two interesting phenomena are observed from

the closed form expressions of derived effective moduli. First, the effective shear

moduli depend on the interface curvature parameter ηs = 3ζ s + 5χ s , that is to say, it is

not affected individually by ζ s and χ s when ηs is a fixed value. Second, the effective

bulk modulus is invariant to interface bending resistance parameters. The invariability

of effective bulk modulus with respect to interface bending resistance is unexpected

since elastic fields in the composites are affected by ηs as seen from Eqs.(A.33-A.47).

5.2. Numerical results and discussion

In this subsection, we conduct a series of parametric studies to investigate the

influence of surface bending resistance on the effective shear modulus. Material

properties for the inhomogeneity are E i = 410 GPa and vi = 0.14 , while material

properties for the matrix are E m = 71 GPa and v m = 0.35 . The interface parameters are

selected as λ s = 3.4939 N/m and µs = −5.4251 N/m [49]. The radius of the

inhomogeneity is R = 1nm .

Fig.2 and Fig.3 shows the effective shear modulus calculated by different methods

with different interface bending stiffness parameters and with different volume

fractions of the inhomogeneity. Results reveal that the interface bending stiffness

22 of 34 pages
parameters and volume fractions can significantly affect the effective shear modulus.

From Fig.2,it is also observed that the effective modulus calculated by the dilute

approximation agrees with that obtained by the Mori-Tanaka method when the volume

fraction of inhomogeneities is small, but they differ considerably when the volume

fraction is large, because the dilute approximation neglects any interaction between

inhomogeneities. Therefore, we use the effective modulus calculated by the Mori-

Tanaka method hereinafter. Besides, the effective shear modulus computed by Eq.(56)

when ηs = 0 agrees with the effective shear modulus using the G-M interface

model[30], which partially verifies the effective moduli as derived in this paper.

An interesting phenomenon is observed in Fig.3 that the effective modulus is

significantly affected by interface bending when ηs is near the characteristic curvature

parameter η*s ( η*s is given in the Appendix). It should be pointed out that the effective

modulus calculated by the Mori-Tanaka method is expected to have larger errors near

η*s , because it is the singular point of Eq.(56). However, it is interesting to learn that

such a singular point exists near which the effective shear modulus can be significantly

increased by surface modification. We further study the influences of the radius and the

volume fraction of the inhomogeneity on η*s . It can be seen in Fig.4 that η*s is

significantly increased with the increase of the inhomogeneity’s size. In contrast,

volume fraction has little effect on η*s .

23 of 34 pages
 3
without interface stress Mori-Tanaka method
G-M model Dilute approximation
= -60 nN·nm
s
2.5
= -45 nN·nm
s
= -30 nN·nm
s

2 = -0 nN·nm
s

1.5

0.5

0
0 0.1 0.2 0.3 0.4 0.5

Fig.2. Variation of the effective shear modulus with different volume fractions

using two homogenization methods

5
f=0% f=30%
f=10% f=40%
4
f=20% f=50%

-1

-2

-3
-100 -80 -60 -40 -20 0 20 40 60 80 100

Fig.3. Influence of the interface curvature parameter ηs on the effective shear

modulus

24 of 34 pages
 4
-1×10
f=10%
f=20%
f=30%
f=40%

3
-1×10

-1370

-1320

-1270

2
-1×10
-1220
2.95 3 3.05

1 1.5 2 2.5 3 3.5 4 4.5 5

Fig.4. The variation of η*s with different inhomogeneity radius and different volume

fractions

A nano-porous material with the S-O model is also investigated in this study. The

material properties for the matrix are E m = 71 GPa and v m = 0.35 . We choose two sets

of interface elastic constants: (a) λ s = 3.4939 N/m, µs = −5.4251 N/m and (b)

λ s = 6.8511 N/m, µs = −0.376 N/m [49]. The variation of the effective shear modulus with

the interface curvature parameter and volume fractions is illustrated in Fig.5, when the

radius of the pore is R = 1nm . Fig.6 shows the variation of the effective shear modulus

of the porous material with the pore’s radius. Size-dependency is observed. The smaller

the void is, the more significant the interface effects are. Results also reveal that the

effective shear moduli of Nano-porous materials can be greatly improved by an

appropriate surface modification.

25 of 34 pages
 (a)

(b)

Fig.5. The variation of the effective shear modulus with different surface curvature

constants and volume fractions when (a) λ s = 3.4939 N/m, µs = −5.4251 N/m and (b)

λ s = 6.8511 N/m, µs = −0.376 N/m

26 of 34 pages
 1.3

1.2

1.1

0.9

0.8

without interface stress

0.7
=3.4939N/m, =-5.4251N/m, =5 nN·nm
s s s
0.6 =3.4939N/m, =-5.4251N/m, =0 nN·nm
s s s

=6.8511N/m, =-0.376N/m, =5 nN·nm

0.5 s s s

=6.8511N/m, =-0.376N/m, =0 nN·nm

s s s
0.4
2 4 6 8 10 12 14 16 18 20

Fig.6. The influence of the void’s radius on the effective shear modulus

6. Conclusions

In this study, we generalized the fundamental framework of analytical

micromechanics to study nano-composites with both interface stretching and bending

effects.

The first contribution of this paper is that the explicit analytical solutions for

spherical nano-inclusion with an interface defined by the Steigmann–Ogden model is

derived, considering uniform eigenstrains or far-field stress/strain loadings. The

Eshelby tensors and the stress/strain concentration tensors for these problems are also

derived in this paper for the first time.

The second contribution of this paper is that explicit expressions for the effective

bulk/shear moduli for nano-composites with a Steigmann–Ogden(S–O) interface are

provided. In addition to size-dependency, the closed form expressions show that the

effective bulk modulus is invariant to interface bending resistance parameters, in

27 of 34 pages
 contrast to the effective shear modulus.

The third contribution of this paper is that an equivalent interface curvature

parameter ηs and a characteristic value η*s for this curvature parameter are defined.

The concentration tensors and the effective shear modulus are not affected individually

by ζ s and χ s , when ηs is a fixed value. The effective shear modulus is affected

significantly by the interface bending resistance when ηs is near η*s .

The explicit solutions derived in this paper can be used as a benchmark for semi-

analytical solutions and numerical solutions in future studies. The derived explicit

expression of the effective modulus with the Steigmann–Ogden interface model is

provided in the supplemental MATLAB code for the convenience of users.

Acknowledgement

The first three authors thankfully acknowledge the support from the National Key

Research and Development Program of China (No. 2017YFA0207800), the National

Natural Science Foundation of China (grant No. 11872008), and the Fundamental

Research Funds for the Central Universities.

Appendix

For the problem of a nano-inclusion, coefficients in Eqs.(9-21) are:

ηs 3ζ s + 5χ s
= (A.1)

γ s (2η s (8(−4 + 5v m )(−7 + 10v m )λ s + 15R(−1 + v m )(−28 + 39v m )μ m

=
+20(−4 + 5v m )(−7 + 10v m )μ s ) + R 2 (525R 2 (−1 + v m ) 2 μ m 2 + 8(−4 + 5v m
(A.2)
)(−7 + 10v m )μ s (λ s + μ s ) + 10R(−1 + v m )μ m (−56λ s + 79v m λ s − 98μ s +
133v m μ s )));

28 of 34 pages
 R 4 (1+v m ) μ m
M1 = − ; (A.3)
6 ( −1 + 2v m ) λ s + 9R ( −1 + v m ) μ m + 6 ( −1 + 2v m ) μ s

M= (R 4μ m (32(7 − 10v m )ηs + R 2 (24(7 − 10v m )λ s − 175R(−1 + v m )μ m

2
(A.4)
+28(7 − 10v m )μ s ))) / (6γ s );

=M 3 (R 6μ m (16(−2 + v m )(−7 + 10v m )ηs + R 2 (−8(1 + v m )(−7 + 10v m )

(A.5)
λ s − 105R(−1 + v m )μ m + 28μ s + 8(9 − 20v m )v m μ s ))) / (3γ s );

R (1 + v m ) μ m
C1 = − ; (A.6)
6 ( −1 + 2v m ) λ s + 9R ( −1 + v m ) μ m + 6 ( −1 + 2v m ) μ s

C2 = (−4(−4 + 5v m )μ m (2ηs − R 2 (λ s + 2μ s ))) / (3Rγ s ); (A.7)

C= (2R(−4 + 5v m )μ m (4(−7 + 16v m )ηs + R 2 (6(−7 + 8v m )λ s + 7(−1 +

3
(A.8)
v m )(5Rμ m + 8μ s )))) / (3γ s );

*
M1αβ = δαβ M1εαβ (A.9)

*
M 2αβ = M 2 εαβ (A.10)

*
M 3αβ = M 3εαβ (A.11)

*
C1αβ = δαβ C1εαβ (A.12)

*
C2αβ = C2 εαβ (A.13)

*
C3αβ = C3εαβ (A.14)

where α , β = x, y, z and δ is Kronecker delta.

For an inhomogeneity under far-field stresses, coefficients in Eqs.(24-37) are:

γ Σ 2ηs (8(−7 + 10vi )(−4 + 5v m )λ s + R(−49 + 61vi )(−4 + 5v m )μ i + 4R

=
(−7 + 10vi )(−8 + 7v m )μ m + 20(−7 + 10vi )(−4 + 5v m )μ s ) + R 2 (−R 2 (7
μ i + 5vi μ i + 28μ m − 40vi μ m )( −8μ i + 10v m μ i − 7μ m + 5v m μ m ) + 8( −7 +
(A.15)
10vi )( −4 + 5v m )μ s (λ s + μ s ) + 2 R(( −35 + 47vi )( −4 + 5vm )λ s μ i + 4(−7 +
10vi )(−5 + 4v m )λ s μ m + 7(7(−1 + vi )(−4 + 5v m )μ i + 6(−7 + 10vi )(−1 +
v m )μ m )μ s ));

1 − 2v m
M1Σ = ; (A.16)
6μ m + 6v m μ m

29 of 34 pages
 M Σ2 (R 3 (2(−1 + 2vi )(−1 + 2v m )λ s − R (1 + vi )(−1 + 2v m )μ i + R(−1 +
=
2vi )(1 + v m )μ m + 2(−1 + 2vi )(−1 + 2v m )μ s )) / (6(1 + v m )μ m ((−2 + 4vi ) (A.17)
λ s − R(μ i + vi μ i + 2μ m − 4vi μ m ) + 2(−1 + 2vi )μ s ));

1
M 3Σ = ; (A.18)
6μ m

−5R 3 (R 4 (μ i − μ m )((7 + 5vi )μ i + 4(7 − 10vi )μ m ) + Rηs ((49 − 61

M Σ4 =
vi )μ i + 4(−7 + 10vi )μ m ) − 4R 2 (−7 + 10vi )μ s (λ s + μ s ) − 4(−7 + 10vi )ηs
(A.19)
(2λ s + 5μ s ) + R 3 ((35 − 47vi )λ s μ i + 4(−7 + 10vi )λ s μ m − 49(−1 + vi )μ i μ s
)) / (12μ m γ Σ ));

6 Σ 2 −2R 6
M 5Σ
= M4 R + (−7 + 10vi )(−1 + v m )(−6ζ s + R 2 λ s + 2R 2μ s −
5 γΣ (A.20)
10χ s );

R ( −1 + 2vi )( −1 + v m )
C1Σ = ; (A.21)
2 (1 + v m ) ( ( 2 − 4vi ) λ s + R ( μ i + vi μ i + 2μ m − 4vi μ m ) + 2μ s − 4vi μ s )

CΣ2 = (5(−1 + v m )(2ηs − R 2 (λs + 2μ s ))) / (− Rγ Σ ); (A.22)

C3Σ =−5R(−1 + v m )((28 − 64vi )ηs + R 2 (6(7 − 8vi )λ s + R(7μ i + 5vi μ i

(A.23)
+28μ m − 40vi μ m ) − 56( −1 + vi )μ s )) / (2γ Σ ));

M1Σαβ δαβ M1Σ Σαβ

= (A.24)

M
= Σ
2αβ δαβ M Σ2 Σαβ (A.25)

M 3Σαβ
= M 3Σ Σαβ (A.26)

M Σ4αβ
= M Σ4 Σαβ (A.27)

M 5Σαβ
= M 5Σ Σαβ (A.28)

C1Σαβ δαβ C1Σ Σαβ

= (A.29)

CΣ2αβ= CΣ2 Σαβ (A.30)

C3Σαβ= C3Σ Σαβ (A.31)

30 of 34 pages
=η*s (R 2 (R 2 ((7 + 5vi )μ i + 4(7 − 10vi )μ m )(2(−1 + f )(−4 + 5v m )μ i + (7
+8 f − 5(1 + 2 f )v m )μ m ) − 8(−1 + f )(−7 + 10vi )(−4 + 5v m )μ s (λ s + μ s )
+ R(−2(−1 + f )(−35 + 47vi )(−4 + 5v m )λ s μ i + 8(−7 + 10vi )(−5 + 4v m +
f (−4 + 5v m ))λ s μ m − 98(−1 + f )(−1 + vi )(−4 + 5v m )μ i μ s + 84(−7 + 10 (A.32)
vi )(−1 + v m )μ m μ s ))) / (2(8(−1 + f )(−7 + 10vi )(−4 + 5v m )λ s + (−1 + f )
R(−49 + 61vi )(−4 + 5v m )μ i − 4R(−7 + 10vi )(−8 − 4 f + 7v m + 5 fv m )μ m
+20(−1 + f )(−7 + 10vi )(−4 + 5v m )μ s ))

where α , β = x, y, z and δ is Kronecker delta.

For an inhomogeneity with far-field strains, coefficients in Eqs.(39-45) are:

1
M
=1
Ε
M
= Ε
3 ; (A.33)
3

2 (1 + v m ) μ m
M Ε2 = − M Σ2 ; (A.34)
−1 + 2v m

M Ε4 = 2μ m M Σ4 ; (A.35)

M 5Ε = 2μ m M 5Σ ; (A.36)

2 (1 + v m ) μ m
C1Ε = − C1Σ ; (A.37)
−1 + 2v m

CΕ2 = 2μ m CΣ2 ; (A.38)

C3Ε = 2μ m C3Σ ; (A.39)

M1Εαβ δαβ M1Ε Εαβ

= (A.40)

M
= Ε
2αβ δαβ M Ε2 Εαβ (A.41)

M 3Εαβ
= M 3Ε Εαβ (A.42)

M Ε4αβ
= M Ε4 Εαβ (A.43)

M 5Εαβ
= M 5Ε Εαβ (A.44)

C1Εαβ δαβ C1Ε Εαβ

= (A.45)

CΕ2αβ= CΕ2 Εαβ (A.46)

C3Εαβ= C3Ε Εαβ (A.47)

31 of 34 pages
Reference
[1] Ghahremani F. Effect of grain boundary sliding on anelasticity of polycrystals.
International Journal of Solids & Structures. 1980;16(9):825-45.
[2] Aboudi J. Damage in composites—Modeling of imperfect bonding. Composites
Science and Technology. 1987;28(2):103-28.
[3] Yu HY, Wei YN, Chiang FP. Load transfer at imperfect interfaces––dislocation-like
model. International Journal of Engineering Science. 2002;40(14):1647-62.
[4] Cherkaoui M, Sabar H, Berveiller M. Elastic composites with coated reinforcements:
A micromechanical approach for nonhomothetic topology. International Journal of
Engineering Science. 1995;33(6):829-43.
[5] Gibbs JW. The scientific papers of J. Willard Gibbs: Longmans, Green; 1906.
[6] Gurtin ME, Murdoch AI. A continuum theory of elastic material surfaces. Archive
for Rational Mechanics & Analysis. 1975;57(4):291-323.
[7] Gurtin ME, Murdoch AI. Surface stress in solids. International Journal of Solids &
Structures. 1978;14(6):431-40.
[8] Ogden RW, Steigmann DJ, Haughton DM. The Effect of Elastic Surface Coating on
the Finite Deformation and Bifurcation of a Pressurized Circular Annulus. J Elast.
1997;47(2):121-45.
[9] Steigmann DJ, Ogden RW. Elastic surface—substrate interactions. Proceedings of
the Royal Society of London Series A: Mathematical, Physical and Engineering
Sciences. 1999;455(1982):437-74.
[10] McDowell MT, Leach AM, Gall K. Bending and tensile deformation of metallic
nanowires. Modelling and Simulation in Materials Science and engineering.
2008;16(4):045003.
[11] Yun G, Park HS. Surface stress effects on the bending properties of fcc metal
nanowires. Physical Review B. 2009;79(19):195421.
[12] Miller RE, Shenoy VB. Size-dependent elastic properties of nanosized structural
elements. Nanotechnology. 2000;11(3):139.
[13] Medasani B, Park YH, Vasiliev I. Theoretical study of the surface energy, stress,
and lattice contraction of silver nanoparticles. Physical Review B. 2007;75(23).
[14] Eremeyev VA. On dynamic boundary conditions within the linear steigmann-
ogden model of surface elasticity and strain gradient elasticity. Advanced Structured
Materials: Springer Verlag; 2019. p. 195-207.
[15] Eremeyev VA, Lebedev LP. Mathematical study of boundary-value problems
within the framework of Steigmann–Ogden model of surface elasticity. Continuum
Mechanics and Thermodynamics. 2016;28(1-2):407-22.
[16] Gharahi A, Schiavone P. Effective elastic properties of plane micropolar nano-
composites with interface flexural effects. International Journal of Mechanical Sciences.
2018;149:84-92.
[17] Wang J, Yan P, Dong L, Atluri SN. Spherical nano-inhomogeneity with the
Steigmann–Ogden interface model under general uniform far-field stress loading.

32 of 34 pages
International Journal of Solids and Structures. 2019:In Press.
[18] Chhapadia P, Mohammadi P, Sharma P. Curvature-dependent surface energy and
implications for nanostructures. Journal of the Mechanics and Physics of Solids.
2011;59(10):2103-15.
[19] Zhao T, Luo J, Xiao Z. Buckling analysis of a nanowire lying on Winkler-Pasternak
elastic foundation. Mech Adv Mater Struct. 2015;22(5):394-401.
[20] Davini C. Material Symmetry of Elastic Shells. J Elast. 2019.
[21] Gerasimov RA, Petrova TO, Eremeyev VA, Maximov AV, Maximova OG. On the
equations of the surface elasticity model based on the theory of polymeric brushes.
Advanced Structured Materials: Springer Verlag; 2019. p. 153-61.
[22] Manav M, Anilkumar P, Phani AS. Mechanics of polymer brush based soft active
materials– theory and experiments. Journal of the Mechanics and Physics of Solids.
2018;121:296-312.
[23] Li X, Mi C. Nanoindentation hardness of a Steigmann–Ogden surface bounding
an elastic half-space. Mathematics and Mechanics of Solids. 2018.
[24] Mi C. Elastic behavior of a half-space with a Steigmann–Ogden boundary under
nanoscale frictionless patch loads. International Journal of Engineering Science.
2018;129:129-44.
[25] Altenbach H, Eremeyev VA, Lebedev LP. On the existence of solution in the linear
elasticity with surface stresses. ZAMM Zeitschrift fur Angewandte Mathematik und
Mechanik. 2010;90(3):231-40.
[26] Altenbach H, Eremeyev VA, Lebedev LP. On the spectrum and stiffness of an
elastic body with surface stresses. ZAMM Zeitschrift fur Angewandte Mathematik und
Mechanik. 2011;91(9):699-710.
[27] Altenbach H, Eremeyev VA, Morozov NF. Surface viscoelasticity and effective
properties of thin-walled structures at the nanoscale. International Journal of
Engineering Science. 2012;59:83-9.
[28] Altenbach H, Eremeyev VA, Morozov NF. Mechanical properties of materials
considering surface effects. IUTAM Bookseries2013. p. 105-15.
[29] Dong L, Wang J, Yan P, Guo Z. A Trefftz collocation method for multiple
interacting spherical nano-inclusions considering the interface stress effect.
Engineering Analysis with Boundary Elements. 2018;94:172-83.
[30] Duan HL, Wang J, Huang ZP, Karihaloo BL. Eshelby formalism for nano-
inhomogeneities. Proceedings of the Royal Society A: Mathematical, Physical and
Engineering Sciences. 2005;461(2062):3335-53.
[31] Duan HL, Wang J, Huang ZP, Karihaloo BL. Size-dependent effective elastic
constants of solids containing nano-inhomogeneities with interface stress. Journal of
the Mechanics & Physics of Solids. 2005;53(7):1574-96.
[32] Duan HL, Wang J, Huang ZP, Luo ZY. Stress concentration tensors of
inhomogeneities with interface effects. Mechanics of Materials. 2005;37(7):723-36.
[33] Duan HL, Wang J, Karihaloo BL. Theory of Elasticity at the Nanoscale. Advances
in Applied Mechanics. 2009;42(08):1-68.
[34] Lim CW, Li ZR, He LH. Size dependent, non-uniform elastic field inside a nano-
scale spherical inclusion due to interface stress. International Journal of Solids and

33 of 34 pages
Structures. 2006;43(17):5055-65.
[35] Mi C, Kouris D. On the significance of coherent interface effects for embedded
nanoparticles. Mathematics and Mechanics of Solids. 2014;19(4):350-68.
[36] Mohammadi P, Liu LP, Sharma P, Kukta RV. Surface energy, elasticity and the
homogenization of rough surfaces. Journal of the Mechanics and Physics of Solids.
2013;61(2):325-40.
[37] Sharma P, Ganti S. Size-dependent eshelbys tensor for embedded nano -inclusions
incorporating surface/interface energies. J Appl Mech Trans ASME. 2004;71(5):663-
71.
[38] Sharma P, Ganti S, Bhate N. Effect of surfaces on the size-dependent elastic state
of nano-inhomogeneities. Applied Physics Letters. 2003;82(4):535-7.
[39] Sharma P, Wheeler LT. Size-dependent elastic state of ellipsoidal nano-inclusions
incorporating surface/interface tension. J Appl Mech Trans ASME. 2007;74(3):447-54.
[40] Dai M, Gharahi A, Schiavone P. Analytic solution for a circular nano-
inhomogeneity with interface stretching and bending resistance in plane strain
deformations. Applied Mathematical Modelling. 2018;55:160-70.
[41] Han Z, Mogilevskaya SG, Schillinger D. Local fields and overall transverse
properties of unidirectional composite materials with multiple nanofibers and
Steigmann–Ogden interfaces. International Journal of Solids and Structures.
2018;147:166-82.
[42] Zemlyanova AY, Mogilevskaya SG. Circular inhomogeneity with Steigmann–
Ogden interface: Local fields, neutrality, and Maxwells type approximation formula.
International Journal of Solids and Structures. 2018;135:85-98.
[43] Neuber Hv. Ein neuer ansatz zur lösung räumlicher probleme der elastizitätstheorie.
der hohlkegel unter einzellast als beispiel. ZAMM‐Journal of Applied Mathematics and
Mechanics/Zeitschrift für Angewandte Mathematik und Mechanik. 1934;14(4):203-12.
[44] Papkovich P. Solution générale des équations differentielles fondamentales
d’élasticité exprimée par trois fonctions harmoniques. CR Acad Sci Paris.
1932;195(3):513-5.
[45] Walpole LJ. Elastic Behavior of Composite Materials: Theoretical Foundations.
Advances in Applied Mechanics Volume 211981. p. 169-242.
[46] Nemat-Nasser S, Hori M. Micromechanics: overall properties of heterogeneous
materials: Elsevier; 2013.
[47] Mori T, Tanaka K. Average stress in matrix and average elastic energy of materials
with misfitting inclusions. Acta metallurgica. 1973;21(5):571-4.
[48] Benveniste Y. A new approach to the application of Mori-Tanakas theory in
composite materials. Mechanics of materials. 1987;6(2):147-57.
[49] Tian L, Rajapakse RKND. Finite element modelling of nanoscale inhomogeneities
in an elastic matrix. Computational Materials Science. 2007;41(1):44-53.

34 of 34 pages