International Journal of Mechanical Sciences 135 (2018) 158–167

Contents lists available at ScienceDirect

International Journal of Mechanical Sciences

journal homepage: www.elsevier.com/locate/ijmecsci

Plane strain cylindrical indentation of functionally graded half-plane with

exponentially varying shear modulus in the presence of residual surface
tension
Thamarai Selvan Vasu, Tanmay K. Bhandakkar∗
Department of Mechanical Engineering, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India

a r t i c l e i n f o a b s t r a c t

Keywords: A functionally graded half-plane with shear modulus varying exponentially along the direction normal to the
Functionally graded material surface and surface effects accounted through Gurtin–Murdoch model, indented by long rigid smooth cylindrical
Residual surface tension indenter is solved to understand the effect of material inhomogeneity and surface effects on indentation response.
Contact mechanics
The Green’s function relating surface load to surface displacement under plane strain condition is obtained semi-
Cylindrical indenter
analytically through the combination of Airy stress function approach and Fourier transforms and utilized to
solve the contact problem. The solution is used to study the effect of inhomogeneity through grading parameter
and surface effects through residual surface tension based intrinsic length scale on the contact pressure, contact
size and in-plane normal stress on the surface responsible for cracks during indentation.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction tance as compared to homogeneous material which suffer from Hertzian

Presence of material inhomogeneity in many naturally occurring ma-
terial systems is claimed as the reason for their superior mechanical
response [1]. Man-made efforts to exploit inhomogeneity with an aim
to better mechanical performance as compared to their homogeneous
counterparts has led to development of systems such as composites,
layer-substrate (LS) systems, functionally graded (FG) materials. FG ma-
terial, the focus of present study is a material system, where the mate-
rial properties vary continuously and the variation is usually described
through a mathematical function of position co-ordinates [2,3]. FG ma-
terials have garnered attention as the current material synthesis and
processing capabilities can allow controlled gradations of materials over
nanometer to microscopic length scales and based on it’s functionality,
it has found applications in areas as varied as bio-medical field, ther-
mal barrier coating, sensors and ene.g. [4]. FG materials with varying
co-efficient of thermal expansion are designed to withstand very high
thermal gradient and used in space plane body, rocket engine compo-
nent [5] and energy conversion devices like gas turbine engine as pro-
tective coating on turbine blades [6]. Artificial FG materials are used
to replace damaged bones and teeth in human beings, which consti-
tute examples of naturally found FG materials [7–9]. Tailor-made FG
materials which resist against penetration and crack propagation have
been proposed as the material for bullet-proof vest [10]. In the context
of indentation problem, FG materials have shown better fracture resis-
cone cracking [11–13]. The indentation problem with its many varia-
tion have also been explored simultaneously through modeling for FG
material vis-a-vis homogeneous material and the following text attempts
to summarize the most important works.
Bakirtas [14] was one of the first authors to solve the two-
dimensional normal contact problem of rigid cylindrical and flat punch
indenting a FG isotropic elastic half-plane with shear modulus varying
exponentially along depth and constant Poisson’s ratio using Airy stress
function approach and Fourier integral transform and studied the ef-
fect of inhomogeneity parameter and Poisson’s ratio on the stress dis-
tribution in the half-plane during indentation and in particular stress
intensity factor in the flat punch case. Giannakopoulos and Suresh
[15,16] carried out analytical and finite element (FE) studies of point
load and axisymmetric indentation problem on a FG half-space with
shear modulus varying as power law and exponential function of depth.
It was shown that the contact response offered by the power law and
exponential FG material varied significantly with operating conditions
for instance, Poisson’s ratio or rate of change of material stiffness along
the depth direction. Suresh et al. [17], performed spherical indentation
experiments on exponentially graded material and observed cracking
(no cracking) on the surface depends upon the transition of stiffness as
a function of depth from soft (stiff) to stiff (soft). The observation was
rationalized through FE studies [17] which showed that the stiffness
variation from soft (stiff) to stiff (soft) leads to tensile (compressive)

∗
Corresponding author.
E-mail addresses: thamaraivasu@iitb.ac.in (T.S. Vasu), tbhanda2@iitb.ac.in (T.K. Bhandakkar).

https://doi.org/10.1016/j.ijmecsci.2017.11.009
Received 3 October 2017; Received in revised form 28 October 2017; Accepted 7 November 2017
Available online 13 November 2017
0020-7403/© 2017 Elsevier Ltd. All rights reserved.
T.S. Vasu, T.K. Bhandakkar
principal stresses on the surface. Giannakopoulos and Pallot [18] pro-
posed use of FG material with power law variation in shear modulus
along depth for wear resistance sliding surfaces based on their study
which revealed lesser principal tensile and shear stresses on surfaces
under contact load in turn lessening the likelihood of fracture, plastic
shake-down, grooving.
Ke and Wang [19,20] obtained the point load solution for multiple
layers with linearly varying shear modulus firmly attached to a homoge-
neous substrate. The solution was used to approximate power law and
exponential variation of shear modulus in a layer and further used to
solve indentation problem with different indenter shapes. Ke and Wang
extended their multi-layer approach to fretting contact [21], thermoe-
lastic contact involving frictional heating [22] with an exponential vari-
ation in the shear modulus of layer attached to a homogeneous half-
plane. Guler and Erdogan [23] solved the frictionless contact of rect-
angular and triangular rigid punch indenting an exponentially graded
layer attached to a homogeneous half-plane. The work was extended
to present a series of benchmarking solutions for the case of contact of
two graded cylinders [24] and frictional sliding contact of parabolic and
cylindrical rigid punch [25–27]. Chidlow et al. [28], solved the adhesive
and non-adhesive contact problem of FG layer with exponential grading
bonded to homogeneous half-plane using Airy stress function approach
and Fourier series and noted that FG material also exhibit both JKR-like
and DMT-like contact response. Unlike previous works, where the ma-
terial properties varied in the direction of depth, Chen et al. [29,30],
considered an arbitrarily oriented gradient half-plane with exponen-
tial grading and solved frictional contact problem using Fourier integral
transform method and looked at the effect of gradient orientation angle
apart from surface friction coefficient and non-homogeneity parameter
on the contact pressure and in-plane surface stress. Comez [31] studied
the contact problem for a rigid cylindrical punch moving steadily with a
constant subsonic velocity on the surface of a functionally graded layer
with constant Poisson’s ratio and exponentially varying elastic modulus
and mass density in the depth direction, bonded to a rigid substrate us-
ing Fourier integral transform technique and studied the effect of punch
velocity on contact width and stress distribution in layer.
A lot many experiments have shown that the mechanical response
of a variety of systems at sub-micron or nano level differs greatly from
the systems at macro level because of the influence of surface effects on
the system due to high surface to volume ratio [32]. One of the success-
ful model to deal with surface effects is given by Gurtin and Murdoch
[33,34], where the surface is treated as a mathematical layer of zero
thickness and perfectly bonded with the bulk material [35]. The two pa-
rameters “residual surface tension” 𝜏 s and “surface elasticity constant”
𝜅 s governs the Gurtin-Murdoch (GM) model. The homogeneous half-
plane or half-space considering surface effects, subjected to specified
surface loading is studied [36,37], and the contact problem is solved
[38–43] to obtain contact pressure and contact size. The results show
that the loading length or contact size approaches the ratio of residual
surface tension to elastic modulus known as surface intrinsic length scale
parameter, the distribution of stresses, displacements profile, contact
pressure and contact size shows significant difference in the presence
and absence of surface effect. Inclusion of adhesion in the contact prob-
lem of homogeneous half-plane with residual surface tension have been
carried out [44–47]. All these works consistently show that presence of
surface tension along with adhesion leads to reduction in contact di-
mension and indentation depth compared to the classical JKR-limit and
that the adhesion with surface effects acts a linking model connecting
the limits of no surface tension (JKR-limit) and very high surface ten-
sion (Young-Laplace equation). Vasu and Bhandakkar [48] studied the
cylindrical contact problem of LS system in the presence of surface ef-
fects and showed that the deviation of result from classical elasticity is
more if the layer is stiff as compared to the substrate. Additionally a
map defined by Hertzian contact size and layer to substrate shear mod-
ulus ratio was presented which defined regions where the response can
be approximated by LS system in the presence and absence of surface
159
International Journal of Mechanical Sciences 135 (2018) 158–167
effects and the layer as a homogeneous half-plane with or without con-
sidering surface effects. Finite element method has also been applied to
solve the contact problem coupled with surface elasticity [42,49,50].
Attia and Mahmoud [49] carried out finite element simulations of the
indentation of triangular, parabolic and flat punch on a FG layer with
shear modulus varying as a power law function of depth bonded to ho-
mogeneous substrate considering GM surface elasticity model with sec-
ond order displacement gradient contributions along the free surface
and layer-substrate interface. Consistent with the studies on homoge-
neous materials, it is shown that for a given depth of penetration, the
contact force required is more if surface elasticity is accounted in FG
materials. Influence of surface elasticity on the contact pressure is more
with increase in the gradation parameter leading to higher indentation
force for the same indentation depth.
Thus it appears that barring the finite element based work of Attia
and Mahmoud [49], no study is available which explores the combined
effect of surface and functional grading on the indentation response in a
semi-analytical framework, a problem the present work aims to address.
Such study will clearly help to bring out the size effect of indenter on the
indentation response in FG half-plane, an aspect which has been missing
in the past and the recent works discussed above. Interestingly none of
the works dealing with indentation on FG material quoted above explic-
itly provide contact size, an information equally valuable along with the
contact pressure. The plan of the paper is as follows: Section 2 provides
the description of the problem, necessary equations, boundary condi-
tions in accordance with the GM model and the Green’s function re-
lating derivative of surface displacement to contact pressure obtained
through Airy stress function approach and Fourier transform. The con-
tact problem in terms of the unknown contact pressure and contact size
is set up and the numerical procedure for it’s solution is summarized in
Appendix A. In Section 3, results are presented to analyze the impact of
inhomogeneity and surface effects on the contact pressure, contact size
and in-plane normal stress on the surface. Finally in Section 4, conclu-
sions are presented.
2. Problem formulation
Fig. 1 shows the schematic of a Functionally Graded (FG) half-plane
described by 𝑥 − 𝑦 co-ordinates and indented quasi-statically by a long
rigid and smooth cylindrical indenter of radius R symmetrical about
x-axis. The resulting deformation of the half plane is assumed to be in-
finitesimal and obeying plane strain condition. The cylinder is subjected
to a line load P leading to a contact over edge of length ’2a’ with ’a’ as
the semi-contact length. The FG half-plane is modeled as a linear elastic
isotropic material with shear modulus 𝜇 = 𝜇(𝑥) varying in the direction
of indentation and a constant Poisson’s ratio 𝜈. The indenter and half-
plane are assumed to be non-interacting so that the indentation is as-
sumed to adhesion-free. The component of displacement along x and y
direction in the half-plane are denoted by ux and uy respectively while
the normal and shear component of stresses (strains) are represented as
𝜎 aa (𝜖 aa ) and 𝜏 ab (𝛾 ab ) respectively where a and b can denote x, y or z
co-ordinates and a ≠ b. The in-plane stress-strain relations under plane
strain conditions are given by [51],
𝜏𝑥𝑦
1−𝜈 1−𝜈
𝜀𝑥𝑥 = (𝜎 − 𝜈 ∗ 𝜎𝑦𝑦 ), 𝜀𝑦𝑦 = (𝜎 − 𝜈 ∗ 𝜎𝑥𝑥 ), 𝛾𝑥𝑦 = , (1)
2𝜇(𝑥) 𝑥𝑥 2𝜇(𝑥) 𝑦𝑦 𝜇(𝑥)
where 𝜈 ∗ = 𝜈∕(1 − 𝜈) and the compatibility condition binding the strains
is [51],
𝜕 2 𝜀𝑥𝑥 𝜕 2 𝜀𝑦𝑦 𝜕 2 𝛾𝑥𝑦
+ = . (2)
𝜕 𝑦2 𝜕 𝑥2 𝜕 𝑥𝜕 𝑦
In the absence of body forces, the in-plane stresses in FG half-plane rep-
resented in terms of Airy stress function 𝜑(x, y) as,
𝜕2 𝜑 𝜕2 𝜑 𝜕2 𝜑
𝜎𝑥𝑥 = , 𝜎𝑦𝑦 = , 𝜏𝑥𝑦 = − , (3)
𝜕𝑦 2 𝜕𝑥 2 𝜕 𝑥𝜕 𝑦
T.S. Vasu, T.K. Bhandakkar International Journal of Mechanical Sciences 135 (2018) 158–167

satisfies the equilibrium equation trivially [51]. Combining Eqs. (1) to and displacement along x and y direction can be obtained from
(3), leads to the sole equation governing the stresses and displacements Eq. (1) and strain-displacement relation [54],
in a FG half-plane with depth dependent shear modulus 𝜇(x) in terms of 1 [ ]
Airy stress function 𝜑 as, 𝑢𝑥 = (1 − 𝜈)𝜎𝑥𝑥 − 𝜈𝜎𝑦𝑦 𝑑𝑦 + 𝐶1 (𝑥), (14)
∫ 2𝜇(𝑥)
( 4 ) ( )
𝜕 𝜑 𝜕4 𝜑 𝜕4 𝜑 𝑑𝜇(𝑥) 𝜕 3 𝜑 𝜕3 𝜑
+2 + (𝜇(𝑥))2 − 2𝜇(𝑥) + 1 [ ]
𝜕 𝑥4 𝜕 𝑥2 𝜕 𝑦2 𝜕 𝑦4 𝑑𝑥 𝜕 𝑥3 𝜕 𝑥𝜕 𝑦2 𝑢𝑦 = (1 − 𝜈)𝜎𝑦𝑦 − 𝜈𝜎𝑥𝑥 𝑑𝑦 + 𝐶2 (𝑥), (15)
∫ 2𝜇(𝑥)
( )2 ( 2 ) ( 2 )
𝑑𝜇(𝑥) 𝜕 𝜑 ∗𝜕 𝜑 2 𝑑 𝜇(𝑥) 𝜕 𝜑 ∗ 𝜕 2 𝜑
2
where C1 (x) and C2 (x) are the constants of integration which ensure that
+2 −𝜈 −𝜇(𝑥) −𝜈 = 0. (4)
𝑑𝑥 𝜕 𝑥2 𝜕 𝑦2 𝑑 𝑥2 𝜕 𝑥2 𝜕 𝑦2 the compatibility Eq. (2) is satisfied. The governing Eq. (4) subjected to
In case of a homogeneous material (𝜇 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡), Eq. (4) reduces to the the boundary conditions given by Eq. (9) can be solved using Fourier
well-known bi-harmonic equation in 𝜑 [51]. The distinctive effect of transform method [54]. Application of Fourier transform along the y-
surface (𝑥 = 0) is modeled using the surface elasticity model proposed direction to Eq. (4) leads to [14],
by Gurtin and Murdoch [33,34] hereafter referred as GM model, which ( 4 ) ( )
𝑑 𝐺 𝑑2𝐺 𝑑𝜇(𝑥) 𝑑 3 𝐺 𝑑𝐺
assumes the surface as a membrane of zero thickness adhered perfectly − 2𝜉 2 + 𝜉 4 𝐺 (𝜇(𝑥))2 − 2𝜇(𝑥) − 𝜉2
𝑑 𝑥4 𝑑 𝑥2 𝑑𝑥 𝑑 𝑥3 𝑑𝑥
to the solid beneath it. Henceforth the solid and any physical quantity [ ( ]
)2
associated with it is referred as bulk to distinguish it from surface. The 𝑑𝜇(𝑥) 𝑑 𝜇(𝑥) 𝑑 𝐺
2 2
𝑠 in terms of sur- + 2 − 𝜇(𝑥)
surface constitutive equation relating surface stress 𝜎𝛽𝛾 𝑑𝑥 𝑑 𝑥2 𝑑 𝑥2
face displacement and surface parameters are given by [33,34], [ ( ) ]
2
𝑑 2 𝜇(𝑥) 𝑑𝜇(𝑥)
𝑠
𝜎𝛽𝛾 = [𝜏 𝑠 + (𝜆𝑠 + 𝜏 𝑠 )𝑢𝜌,𝜌 ]𝛿𝛽𝛾 + 𝜇 𝑠 (𝑢𝛽,𝛾 + 𝑢𝛾,𝛽 ) − 𝜏 𝑠 𝑢𝛾,𝛽 , (5) − 𝜉 2 𝜇(𝑥) −2 𝜈 ∗ 𝐺 = 0, (16)
𝑑 𝑥2 𝑑𝑥

𝜎3𝑠𝛾 = 𝜏 𝑠 𝑢3,𝛾 .
where the superscript ‘s’ indicates that the material property or field
variable like displacement or stress corresponds to surface and indices
𝛽, 𝛾 and 𝜌 can denote y or z co-ordinate. The GM model is governed
by three surface material parameters 𝜇s , 𝜆s and 𝜏 s , where the first two
are referred as surface Lame constants and the last one is the resid-
ual surface tension under unconstrained conditions. The displacement
components of surface and bulk are same at 𝑥 = 0 to satisfy the compat-
ibility of displacement and hence the surface displacement components
in Eqs. (5) and (6) are represented without the superscript “s”. 𝜎𝛽𝛾 𝑠 is a
(6) a fourth-order linear ordinary differential equation in x co-ordinate. The
present work considers an exponential variation of shear modulus along
the depth direction in the half-plane as given by [14,19–25,27–31],
𝜇(𝑥) = 𝜇0 𝑒𝑘𝑥 ; 𝑘 ≥ 0, (17)
where 𝜇 0 is the shear modulus on the surface (𝑥 = 0) and k is the grading
parameter emblematic of the inhomogeneity with it’s inverse (1/k) in-
dicative of the length over which shear modulus varies in the half-plane.
Substituting the exponential variation of shear modulus from Eq. (17) in
Eq. (16) leads to [14],

two dimensional surface stress tensor with in-plane (𝑦 − 𝑧 plane) normal 𝑑4𝐺 𝑑2𝐺 𝑑3𝐺 𝑑𝐺 𝑑2𝐺
− 2𝜉 2 + 𝜉 4 𝐺 − 2𝑘 + 2𝜉 2 𝑘 + 𝑘2 + 𝜉 2 𝑘2 𝜈 ∗ 𝐺 = 0.
and shear components. u𝛽, 𝛾 implied derivative of 𝛽 component of sur- 𝑑 𝑥4 𝑑 𝑥2 𝑑 𝑥3 𝑑𝑥 𝑑 𝑥2
face displacement with respect to 𝛾 component of position co-ordinate. (18)
The surface equilibrium conditions relating the surface and bulk stresses
The solution to the above ODE gives the Airy stress function G in Fourier
with the applied loading (fx (y), fy (y), 0) are as follows [33,34,40],
domain as,
𝜎𝑥𝑥 + 𝜏 𝑠 𝑢𝑥,𝑦𝑦 + 𝑓𝑥 (𝑦) = 0, −∞ < 𝑦 < ∞ 𝐺(𝑥, 𝜉) = 𝐴(𝜉)𝑒𝑛1 𝑥 + 𝐵(𝜉)𝑒𝑛2 𝑥 + 𝐶(𝜉)𝑒𝑛3 𝑥 + 𝐷(𝜉)𝑒𝑛4 𝑥 , (19)
(7)
𝜏𝑥𝑦 + 𝜏,𝑦𝑠 + 𝜅 𝑠 𝑢𝑦,𝑦𝑦 + 𝑓𝑦 (𝑦) = 0, −∞ < 𝑦 < ∞
where, A(𝜉), B(𝜉), C(𝜉) and D(𝜉) are the integration constants. n1 , n2 , n3
where corresponding to frictionless indentation in Fig. 1,
and n4 are roots of the characteristic equation of ODE Eq. (18), which
𝑓𝑥 (𝑦) = 𝑝(𝑦); |𝑦| ≤ 𝑎, 𝑓𝑥 (𝑦) = 0; |𝑦| > 𝑎, 𝑓𝑦 (𝑦) = 0; −∞ < 𝑦 < ∞. (8) are given by [14],
𝜅𝑠 = 2𝜇 𝑠
+ 𝜆𝑠
and p(y) is the contact pressure beneath indenter within 𝑛1 = 𝑘(1 − 𝑚)∕2 + 𝑖𝑙, 𝑛2 = 𝑘(1 − 𝑚)∕2 − 𝑖𝑙,
(20)
the contact region (|y| ≤ a). Assuming 𝜏 s to be constant, the term 𝜏,𝑦𝑠 in 𝑛3 = 𝑘(1 + 𝑚)∕2 + 𝑖𝑙, 𝑛4 = 𝑘(1 + 𝑚)∕2 − 𝑖𝑙,
Eq. (7), will vanish [52]. It has been seen that the parameter 𝜅 s hardly √
influences the friction less normal contact response with surface elas- where 𝑖 = −1,
√ √
ticity [36,48,53] and hence in the present work, the influence of 𝜅 s is √ √
√( )⎡ √ ( √ )2 ⎤
ignored. Therefore Eq. (7) becomes, √ √
√ 1 2𝜉 2 ⎢ √ 𝜈 ∗ 𝑘𝜉 ⎥
𝑚=√ √ + ⎢1 + 1 + ⎥, (21)
𝜎𝑥𝑥 + 𝜏 𝑠 𝑢𝑥,𝑦𝑦 + 𝑝(𝑦) = 0, |𝑦| ≤ 𝑎, 𝜎𝑥𝑥 + 𝜏 𝑠 𝑢𝑥,𝑦𝑦 = 0, |𝑦| > 𝑎, 2 𝑘2 ⎢ 𝜉 2 + 𝑘2 ∕4 ⎥
(9) ⎣ ⎦
𝜏𝑥𝑦 = 0, −∞ < 𝑦 < ∞.
The Fourier transform of Airy stress function 𝜑 and it’s inverse transform and
√
are expressed in terms of each other as, 𝑙 = 𝜉 𝜈 ∗ ∕𝑚. (22)
∞ ∞
−𝑖𝜉𝑦 1
𝐺= 𝜑𝑒 𝑑 𝑦; 𝜑 = 𝐺𝑒−𝑖𝜉𝑦 𝑑 𝜉, (10) It can be seen from Eq. (21) that for any value of 𝜉, m ≥ 1, since is a 𝜈∗
∫−∞ 2𝜋 ∫−∞
positive real quantity due to the constraint imposed on Poisson’s ratio 𝜈
where G(x, 𝜉) is the Fourier transform of Airy stress function 𝜑. Invoking to lie between 0 to 0.5 due to practicality and thermodynamics. In this
Eq. (3), the in-plane stresses in terms of G are given by [54], work, the FG half-plane is assumed to become stiffer with depth (k ≥ 0)
∞ and hence from Eq. (20), the real part of n1 and n2 are negative while
−1
𝜎𝑥𝑥 = 𝜉 2 𝐺𝑒−𝑖𝜉𝑦 𝑑𝜉, (11)
2𝜋 ∫−∞ the real part of n3 and n4 are positive. It should be noted that even if a
∞
gradually softening FG half-plane (k < 0) is assumed, owing to m > 1, n1
1 𝑑 2 𝐺 −𝑖𝜉𝑦 and n2 will have positive real part while n3 and n4 will have negative
𝜎𝑦𝑦 = 𝑒 𝑑𝜉, (12)
2𝜋 ∫−∞ 𝑑 𝑥2 real part. Thus in any case, two roots will have positive real part while
∞ remaining two roots will have negative real part. Based on the physical
1 𝑑𝐺 −𝑖𝜉𝑦
𝜏𝑥𝑦 = 𝜉 𝑖𝑒 𝑑𝜉, (13) requirement, that the stresses have to vanish far away from the loading
2𝜋 ∫−∞ 𝑑𝑥 √
region ( 𝑥2 + 𝑦2 → ∞) and the fact that the real part of roots n3 and n4

160
T.S. Vasu, T.K. Bhandakkar International Journal of Mechanical Sciences 135 (2018) 158–167

are positive, leads to the constants C(𝜉) and D(𝜉) to be zero. Substituting Given the applied load P and shape of the indenter, integral Eqs. (29) and
the stresses and displacements expressed in terms of Airy stress function (31) can be solved simultaneously to obtain the semi-contact size a
in Fourier domain from the Eqs. (11) to (15) in Eq. (9) and solving for and pressure profile p(y). Numerical procedure for the solution of
the unknown constants A(𝜉) and B(𝜉) leads to, Eqs. (29) and (31) is summarized in Appendix A. In the next section,
( ( )( ( ( ) ))) results are presented to analyze the effect of residual surface tension
𝐴(𝜉) = 𝑝̄(𝜉)𝑘1 𝑛2 ∕ 𝜉 2 𝑛1 − 𝑛2 𝑘1 + 𝑆 (1 − 𝜈)𝜉 2 𝑛1 + 𝑛2 − 𝑘 + 𝑛1 𝑛2 𝑘𝜈 , and inhomogeneity on the indentation response.
(23)
3. Results and discussion

𝐵(𝜉) = −𝐴(𝜉)𝑛1 ∕𝑛2 , (24) aH is defined as the Hertzian semi-contact size obtained due to
where, 𝑝̄(𝜉) is the Fourier transform of the contact pressure p(y) in the smooth plane strain indentation of cylindrical rigid indenter of radius
contact region |y| ≤ a and invoking Eq. (10), is defined, R subjected to load P on a homogeneous half-plane with shear modulus
𝑎 𝜇 0 and Poisson’s ratio 𝜈. It is given as [55],
𝑝̄(𝜉) = 𝑝(𝑦) 𝑒−𝑖𝜉𝑦 𝑑𝑦, (25) √
∫−𝑎 2(1 − 𝜈)𝑅𝑃
( )( ) 𝑎𝐻 = . (32)
and 𝑘1 = 2(1 − 𝜈) 𝑘 − 𝑛1 𝑘 − 𝑛2 , 𝑆 = 𝜏 𝑠 (1 − 𝜈)∕𝜇0 . The parameter S has 𝜋𝜇0
unit of length and its magnitude signifies the length scale over which In this section, the results are presented to study the effect of inhomo-
the impact of residual surface tension is considerable. Thus with known geneity (k ≠ 0) and residual surface tension (S ≠ 0) on contact pressure
p(y) and hence 𝑝̄(𝜉), Eqs. (23) and (24) can be coupled with Eqs. (11) to and contact size. As shown through Eqs. (A.3) and (A.5) in Appendix A,
(15) to find stresses and displacement anywhere in the half-plane. In the input quantities R and P with regards to the indentation problem in
particular, the derivative of the normal displacement along the surface Fig. 1 can be combined and presented as a single input aH . Hence all the
(𝑥 = 0) in terms of contact pressure p(y) is given as, results in this work are provided with aH as the input rather than pre-
scribing R or P individually. Unless otherwise mentioned, the numerical
(1 − 𝜈) 𝑎 ∞ 𝑒−𝑖𝜉(𝑦−𝑦 ) 𝑘𝑘2 𝑝(𝑦′ )
′
𝜕
𝑢 (0, 𝑦) = ( ) 𝑑 𝜉𝑑 𝑦 ,
′ values relevant for the results shown in this section are taken as 𝜈 = 0.4,
𝜕𝑦 𝑥 𝜋𝜇0 ∫−𝑎 ∫−∞ 𝜉 𝑆𝑘𝑘2 + 2(1 − 𝜈)𝑘2 𝑚2 (𝑚 + 1)2 + 4𝜈 ∗ 𝜉 2
𝜇0 = 1 MPa and 𝜏𝑠 = 0.1 N/m [38,48,52].
(26) Fig. 2 compares the indentation response in the absence of residual
where 𝑘2 = 𝑘2 𝜈𝑚2 (𝑚 −
1) + 2
4 𝜉 2 (1 −
− 𝜈)(𝜈 ∗ 𝑚3 ).
Coming back to the surface tension (𝑆 = 0) calculated using the “Full-Solution” approach
cylindrical indentation problem of Fig. 1, rather than the contact pres- shown as line and the “Approx-Solution” shown through markers for
sure p(y), it is the displacement ux (0, y), that is prescribed in the contact 𝑘𝑎 = 0.5, 1, 2. The solution approaches viz. “Full-Solution” and “Approx-
region (|y| ≤ a) as [55], Solution” are discussed in Appendix A. The semi-contact size ratio
a/aH obtained through the calculation is noted beside the pressure pro-
√ 𝑦2
𝑢𝑥 (0, 𝑦) = 𝛿 − (𝑅 − 𝑅2 − 𝑦2 ) ≈ 𝛿 −
, |𝑦| ≤ 𝑎, (27) file. It is clear from the semi-contact size ratio and pressure profile in
2𝑅 Fig. 2 that the computationally efficient “Approx-Solution” is compa-
where 𝛿 is the vertical displacement of the indenter under the action of rable to “Full-Solution” till ka ≤ 1. Hence the limit of applicability of
force P and the term ‘y2 /2R’ is the parabolic approximation of the circu- “Approx-Solution” is ka ≤ 1. Since the results that follow involve sur-
lar boundary applicable when a, 𝛿 ≪ R [55]. In case of any other inden- face effects (S ≠ 0) and grading parameter (ka > 1), the “Full-Solution”
ter, the parabolic term will be replaced with the appropriate expression approach is used to calculate the indentation solution shown in further
representative of it’s shape. As 𝛿 is indeterminate in two-dimensional figures.
contact problems [55], derivative of Eq. (27) with respect to y is taken, Figs. 3 and 4 shows the effect of the inhomogeneity of the half-plane
leading to the modified boundary condition, through the variation of the normalized grading parameter kaH on the
𝜕 𝑦 indentation response of cylindrical indenter indenting a FG half-plane
𝑢 (0, 𝑦) = − , |𝑦| ≤ 𝑎. (28) as shown in Fig. 1 in the absence (𝑆 = 0) and presence ((S ≠ 0)) of resid-
𝜕𝑦 𝑥 𝑅
ual surface tension respectively. Throughout the results of Figs. 3 and
Combining Eqs. (26) and (28) gives,
4, the Hertzian semi-contact size aH defined in Eq. (32) is maintained
𝑎 ∞ 𝑦 𝜇0
1 equal to 0.618 μm and in case of Fig. 4, S is set to 60 nm. The value of
𝐻(𝑦, 𝑦′ , 𝜉)𝑑𝜉𝑝(𝑦′ )𝑑𝑦′ = − , (29)
𝜋 ∫−𝑎 ∫0 𝑅(1 − 𝜈) semi-contact size ratio a/aH is obtained as a part of the solution and in-
where H(y, y′, 𝜉) is the Green’s function relating derivative of surface dicated by the arrows for each curve in the plot. Irrespective of whether
displacement to contact pressure, surface effects are considered or not, with increasing kaH , the bulk of
the half-plane becomes relatively stiffer compared to the surface and
sin(𝜉(𝑦 − 𝑦′ ))𝑘𝑘2
𝐻(𝑦, 𝑦′ , 𝜉) = ( ). (30) for the same indentation conditions, the contact pressure magnitude in-
𝜉 𝑆𝑘𝑘2 + 2(1 − 𝜈)𝑘2 𝑚2 (𝑚 + 1)2 + 4𝜈 ∗ 𝜉 2 creases and the contact size reduces as seen in Figs. 3 and 4, a conclu-
The parameter S occurring in Eq. (30) can also be interpreted as a switch sion also observed in the finite element study of Attia and Mahmoud
to turn ON/OFF the effect of residual surface tension. A non-zero value [49]. The results can be contrasted with the indentation response of
of S implies involvement of residual surface tension in the model while layer-substrate (LS) system in literature. For instance, in the case of soft
𝑆 = 0 means classical Hertzian indentation. In-fact for 𝑆 = 0, the Green’s layer on a stiff substrate in the absence and presence of residual surface
function H(y, y′, 𝜉) in Eq. (30) reduces to the Green’s function obtained tension, Gupta and Walowit [56] and Vasu and Bhandakkar [48] re-
in the work of [14]. In case of a homogeneous half-plane (𝑘 = 0) with spectively have shown that with decrease in shear modulus ratio of film
S ≠ 0, the Green’s function in Eq. (30) reduces to the one obtained in the to substrate, under similar indentation conditions, the contact pressure
work of Long et al. [38]. Note that the integration range in Eq. (29) is magnitude increases and contact size reduces. Thus irrespective of the
changed to 0 to ∞ due to even nature of the integrand and symmetry nature of inhomogeneity; discrete as in LS system or continuous as in
of the pressure profile owing to symmetry of the indenter and loading. FG material, the indentation response is qualitatively same with and
The contact pressure p(y) is related to the applied load P through the without the inclusion of surface effects. In both the Figs. 3 and 4, as the
force balance of the indenter in the vertical direction as, grading parameter k is reduced the result approached the homogeneous
𝑎 (kaH =0) half-plane result given by Johnson [55] and Long et al. [38],
𝑝(𝑦′ )𝑑𝑦′ = 𝑃 . (31) respectively. The results in Fig. 3 is qualitatively similar to the result
∫−𝑎
in the work of Bakirtas and Guller and Erdogan [14,23] but distinguish

161
T.S. Vasu, T.K. Bhandakkar
Fig. 1. Schematic of functionally graded (FG) half-plane indented by a long rigid cylindrical indenter of radius R with normal load P leading to a contact zone of length 2a, where a is
the semi-contact length. The FG half-plane is modeled as a linear elastic isotropic material with a constant Poisson’s ratio 𝜈 and shear modulus varying exponentially along the direction
of indentation as 𝜇(𝑥) = 𝜇0 𝑒𝑘𝑥 , k > 0.
themselves through the normalization of the pressure. Especially in the
work of Guller and Erdogan [23], the inhomogeneity is induced through
different ka value and the result is presented in terms of 2ap(y)/P. The
issue with such a normalization is that for different ka, the semi-contact
size a will also differ and directly comparing the pressures for different
grading parameters from the figure is not correct. Hence in the present
work, since aH is maintained same, variation of kaH is equivalent to vari-
ation in k and pressure p(y) and the semi contact size a can be readily
compared. None of the previous works concerning FG material inden-
tation [14,18–20,22,23,25–27,29–31] report contact size as done in the
present work. In Fig. 3, following the classical indentation, the pressure
is zero at the contact edge while in the presence of surface effects as
seen in Fig. 4 and consistent with past study on half-plane [38] and LS
system [48], the contact pressure at the edge (p( ± a)) is non-zero and
whose magnitude increases with increasing inhomogeneity. In contrast,
the finite element based solution of contact problem in the presence of
surface elasticity predicts zero pressure at the edge of the contact region
[42,49,50].
In order to examine the role of the smallness of the indenter di-
mension R in the presence of surface effects and inhomogeneity, Fig. 5
presents the indentation response of FG half-plane (Fig. 1) for two values
of Hertzian semi-contact 𝑎𝐻 = {0.618, 1.95} μm . For an applied load of P
= 1 N/m and elastic constants 𝜇 0 = 1MPa and 𝜈= 0.4, 𝑎𝐻 = {0.618, 1.95}
μm corresponds to an indenter radius R equal to {1, 10}μm . The grad-
ing parameter k is maintained equal to 2∕𝑎∗𝐻 where 𝑎∗𝐻 is chosen equal
to 0.618 μm and in cases where surface effects are invoked, parameter
S is set to 60 nm. Compared to previous figures, the contact pressure in
162
International Journal of Mechanical Sciences 135 (2018) 158–167
Fig. 5 is normalized as 2𝑎∗𝐻 𝑝(𝑦)∕𝑃 in order to facilitate ready comparison
of pressure for different aH values. In order to bring out the impact of
residual surface tension, for each aH value, results for surface parameter
𝑆 = 60𝑛𝑚 are compared with the classical indentation in the absence of
any surface effects (𝑆 = 0). For the case of 𝑎𝐻 = 1.95 μm (𝑅 = 10 μm for
values of P, 𝜇 0 , 𝜈 stated above), the radius of indenter is too big com-
pared to the residual surface tension length scale given by the parame-
ter S to demonstrate any significant difference between the indentation
response of FG half-plane with and without inclusion of residual sur-
face tension. But in case of 𝑎𝐻 = 0.618 μm (𝑅 = 1 μm for values of P,
𝜇 0 , 𝜈 stated above) the radius of the indenter is small enough to bring
out the surface effects into action, leading to FG half-plane with resid-
ual surface tension exhibiting higher stiffness as compared to FG half-
plane without surface effects. Similar conclusion can also be obtained if
Figs. 3 and 4 are compared simultaneously. The increased stiffness ex-
hibited by half-plane in the presence of surface effects for a sub-micron
scale indenter is observed even when the material is homogeneous [38].
It should be noted from Eq. (32) and allied discussion in Appendix A,
that the same results can be arrived in Fig. 5 by varying P and keeping
R constant.
It has been reported in literature that FG half-plane with shear mod-
ulus increasing exponentially away from the free surface (k > 0) is pe-
culiar compared to other gradings in the sense that upon indentation
it exhibits a reasonably high magnitude of in-plane tensile stress in
the vicinity of contact zone which in turn can lead to surface cracking
[3,15,16,25]. So it would be worthwhile to see the influence residual
surface tension has on the in-plane tensile stress 𝜎 yy (0, y) during cylin-
T.S. Vasu, T.K. Bhandakkar International Journal of Mechanical Sciences 135 (2018) 158–167

1.4
a/aH=0.447
a/aH=0.562
1.2
Normalized Pressure 2ap(y)/P a/a =0.6865
H
a/a =0.447
H
1
a/aH=0.812
0.8

0.6
ka=2
0.4 ka=2
ka=1
0.2 ka=1
ka=0.5
ka=0.5
0
0 0.5 1 1.5 2
Normalized distance y/aH
Fig. 2. Comparison of the normalized contact pressure 2ap(y)/P and semi-contact size ratio a/aH in the absence of residual surface tension (𝑆 = 0) corresponding to the cylindrical
indentation (Fig. 1) obtained using “Full-Solution” shown as line and ”Approx-Solution” shown using marker respectively for different ka values. aH is the Hertzian semi-contact size
given by Eq. (32). P is the load applied to the cylinder leading to contact over edge of length 2a. Please refer Appendix A for “Full-solution” and ”Approx-solution” approaches. On
account of symmetry of the problem, result is shown only for y ≥ 0.

2.5
a/a =0.5877
H
a/aH=0.65
2
Normalized Pressure 2aHp(y)/P

a/a =0.75
H

a/aH=0.812
1.5

1
ka =0, Johnson [54]
H
kaH=0.5 a/aH=1
0.5 kaH=1
kaH=2
kaH=3
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Distance y/a
Fig. 3. Effect of inhomogeneity expressed through the variation in the normalized grading parameter kaH on the contact pressure p(y) and semi-contact size a of FG half-plane subjected
to cylindrical indentation (Fig. 1) in the absence of residual surface tension (𝑆 = 0). The relevant input are Hertzian semi-contact size (Eq. (32)) aH = 0.618 μm and 𝜈 = 0.4. P is the load
applied on the cylinder and a is the semi-contact size. On account of symmetry of the problem, result is shown only for y ≥ 0.

163
T.S. Vasu, T.K. Bhandakkar International Journal of Mechanical Sciences 135 (2018) 158–167

2.5
a/aH=0.52287
a/aH=0.5764
Normalized Pressure 2a p(y)/P
2
a/aH=0.66326
H

a/a =0.7367
H
1.5

ka =0, Long et al. [38]

H
kaH=0.5
1
ka =1
H
kaH=2 a/a =0.8774
H

kaH=3
0.5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized Distance y/a
Fig. 4. Effect of inhomogeneity expressed through the variation in the normalized grading parameter kaH on the contact pressure p(y) and semi-contact size a of FG half-plane subjected
to cylindrical indentation (Fig. 1) in the presence of residual surface tension (S ≠ 0). The relevant input are Hertzian semi-contact size (Eq. (32)) aH = 0.618 μm and 𝜈 = 0.4, S = 60 nm.
P is the load applied on the cylinder and a is the semi-contact size. On account of symmetry of the problem, result is shown only for y ≥ 0.

Fig. 5. Effect of residual surface tension expressed through the variation of hertzian semi-contact size aH on the contact pressure p(y) and semi-contact size a of FG half-plane subjected
to cylindrical indentation (Fig. 1). 𝑎∗𝐻 is introduced for uniform normalization and the normalized grading parameter 𝑘𝑎∗𝐻 =2, where 𝑎∗𝐻 = 0.618 μm . On account of symmetry of the
problem, result is shown only for y ≥ 0.

164
T.S. Vasu, T.K. Bhandakkar
Fig. 6. Effect of inhomogeneity and residual surface tension expressed through normalized grading parameter kaH and residual surface tension length scale S respectively on the in-plane
normal stress component 𝜎 yy along the surface x=0, during cylindrical indentation of FG half-plane (Fig. 1). The relevant input are Hertzian semi-contact size (Eq. (32)) aH = 0.618 μm
and 𝜈 = 0.4. P is the load applied on the cylinder and a is the semi-contact size. On account of symmetry of the problem, result is shown only for y ≥ 0.
drical indentation of FG half-plane dealt in the present work (Fig. 1).
Fig. 6 shows the variation of in-plane normal stress 𝜎 yy (0, y) along
the surface from the axis of the indenter to its right to reveal the in-
dividual and combined effect of material inhomogeneity (𝑘𝑎𝐻 = 0, 1)
and residual surface tension (𝑆 = 0, 60𝑛𝑚). Throughout the four cases
discussed in Fig. 6, the Hertzian contact size signifying applied load
P and indenter radius R is maintained at 𝑎𝐻 = 0.618 μm. The effect of
exponential inhomogeneity on the nature of 𝜎 yy (0, y) is evident if the
two curves referred by (𝑘𝑎𝐻 = 1, 𝑆 = 0𝑛𝑚) and (𝑘𝑎𝐻 = 0, 𝑆 = 0𝑛𝑚) cor-
responding to classical indentation on half-plane with and without inho-
mogeneity respectively are compared. The former case shows transition
of in-plane normal stress from compressive to tensile as the contact edge
is approached (𝑦 → 𝑎− ), a result consistent with literature [25]. Intro-
duction of surface effects through residual surface tension (𝑆 = 60𝑛𝑚)
leads to smoother transition of the in-plane normal stress 𝜎 yy at the
contact edge from the contact (|y| ≤ a) to non-contact region (|y| > a)
for both homogeneous and exponentially graded material as seen from
cases labeled as (𝑘𝑎𝐻 = 0, 𝑆 = 60𝑛𝑚) and (𝑘𝑎𝐻 = 1, 𝑆 = 60𝑛𝑚) in Fig. 6.
The effect of residual surface tension in a homogeneous medium (𝑘 = 0)
is to turn the in-plane normal stress 𝜎 yy on the surface compressive
even outside the contact zone (|y| > a) as seen from comparing curves
corresponding to (𝑘𝑎𝐻 = 0, 𝑆 = 0𝑛𝑚) with (𝑘𝑎𝐻 = 0, 𝑆 = 60𝑛𝑚) labels in
Fig. 6. Similar effect of residual surface tension persists for exponen-
tially graded half-plane. In the result allied to (𝑘𝑎𝐻 = 1, 𝑆 = 60𝑛𝑚), the
in-plane normal stress 𝜎 yy on the surface continues to remain tensile
outside the contact zone analogous to the case of (𝑘𝑎𝐻 = 1, 𝑆 = 0𝑛𝑚),
but the magnitude of the stress is smaller in the former as compared
to the latter case. Thus in situations when surface effects are perti-
nent, for instance in sub-micron level indentation, based on the result
of Fig. 6 it is conjectured that likelihood of crack formation will be
lessened.
165
International Journal of Mechanical Sciences 135 (2018) 158–167
4. Conclusion
With a motive to study the effect of material inhomogeneity and sur-
face effects on indentation, the present work deals with the contact prob-
lem of FG half-plane indented by long rigid cylindrical indenter with the
shear modulus varying exponentially along the direction of indentation
and surface effects modeled using GM model. The effect of inhomogene-
ity and surface effects is represented through the grading parameter k
which denotes the severity of the variation of shear modulus along the
depth direction and residual surface tension intrinsic length scale S re-
spectively. The solution procedure enlists Airy stress function approach
and Fourier transforms to obtain Green’s function relating derivative of
surface displacement to contact pressure of FG half-plane in the presence
of surface effects and the same is utilized to obtain contact pressure and
contact size. The obtained solution is thoroughly benchmarked to re-
cover the available solution in literature when the surface effects are ab-
sent or the half-plane is homogeneous. In the absence of surface effects
(𝑆 = 0), as a computationally efficient alternative to the full numerical
evaluation, an approximate but closed form expression for Green’s func-
tion is presented which is shown to provide satisfactory results for the
indentation problem up to normalized grading parameter ka ≤ 1, where
a is the semi-contact length. With regards to the effect of inhomogeneity
and surface effects to the indentation response, the observations from
present work are summarized as follows:
• The qualitative nature of indentation response is unaffected by the
nature of inhomogeneity i.e. continuous as in functionally graded
half-plane dealt here or discrete as in a layer-substrate system.
• Irrespective of the system being homogeneous or not, as the inden-
ter dimension approaches the residual surface tension based intrinsic
length scale, the indentation response based on residual surface ten-
sion deviates significantly from the classical one.
T.S. Vasu, T.K. Bhandakkar International Journal of Mechanical Sciences 135 (2018) 158–167

• In cases, where the indenter size is small enough to invoke surface The normalized version of HApprox (y, y′, 𝜉) is given as,
effects, the indentation response of half-plane with and without in-
𝑐1 𝜈 ∗ 𝜂𝐾 𝑐3
homogeneity based on residual surface tension is stiffer compared to 𝐻̂ 𝐴𝑝𝑝𝑟𝑜𝑥 (𝑌𝑗 , 𝑌𝑗′ , 𝜂) = 1 − −( )2 +( )2 , (A.7)
𝜂 𝜂 + 𝑐1 + 𝑐2 𝜂 + 𝑐1 + 𝑐2
the classical response. Thus for the same load and same indenter size,
residual surface tension based indentation will predict a higher peak where 𝑐1 = 𝐾∕2, 𝑐2 = 𝑐1 2 𝜈 ∗ , 𝑐3 = 𝑐1 2 (4𝜈 − 1)∕(1 − 𝜈). The integral in
contact pressure and smaller contact size compared to the classical Eq. (A.3) with kernel H(Y, Y′, 𝜂) replaced with HApprox (Y, Y′, 𝜂) can be
counterpart. integrated in closed form and the solution is expressible in terms of Sine
• As a result of smoother transition of in-plane normal stress along Si(x) and Cosine Ci(x) integral function given in Abramowitz et al. [58].
the surface at the contact edge from contact to non-contact region It is too long to be provided here and is given in the form of MAPLE© file
and it’s lesser tensile magnitude in the vicinity of contact edge ob- “GF-Approx-Sln” as a supplementary file. In case of absence of residual
served in the residual surface tension based solution, it is speculated surface tension (𝑆 = 0), if H(Y, Y′, 𝜂) is replaced with HApprox (Y, Y′, 𝜂),
that chances of crack formation are lesser during sub-micron level numerical evaluation of integral in Eq. (A.3) is side-stepped and matrix
indentation as compared to micro or macro indentation in an ex- B can be readily calculated using the expression given in the supple-
ponentially graded half-plane with increasing shear modulus away mentary file “GF-Approx-Sln”. The corresponding solution which is a
from the surface. computationally efficient alternative is mentioned as Approx-Solution in
• Based on the prior observation in the layer-substrate system [48], it the main text. The limits of applicability of the Approx-Solution as a re-
is expected that the surface elasticity constant 𝜅 s will have negligible placement for Full-Solution is discussed in the main text.
influence on the results and observation presented in this work based Once semi-contact size a and the contact pressure p(y) is known, in-
solely on the residual surface tension constant. plane normal stress on the surface 𝜎 yy (0, y) shown in Fig. 6 is calculated
combining Eqs. (12), (19) and (23) to (25) resulting in a double integral
Appendix A. Normalizing and numerical scheme for the similar to Eq. (26) or (A.3) and given by,
evaluation of contact pressure
𝑎 { ∞ ( ) ( )}
𝜎𝑦𝑦 = 𝐻𝐹 𝐺𝑀 𝜉, 𝑦, 𝑦′ − 𝐻𝐻𝑜𝑚𝑜𝑔𝑒𝑛𝑒𝑜𝑢𝑠 𝜉, 𝑦, 𝑦′
The normalization of the variables for numerical convenience in the ∫−𝑎 ∫0
present analysis is as follows, [ ( )]
× cos 𝜉 𝑦 − 𝑦′ 𝑝̄(𝜉)𝑑𝜉𝑑𝑦′
𝑌 = 𝑦∕𝑎, 𝑌 ′ = 𝑦′ ∕𝑎, 𝜂 = 𝜉𝑎, 𝐾 = 𝑘𝑎, 𝐾2 = 𝑘2 𝑎2 , (A.1) 𝑎 ∞ ( ) [ ( )]
+ 𝐻𝐻𝑜𝑚𝑜𝑔𝑒𝑛𝑒𝑜𝑢𝑠 𝜉, 𝑦, 𝑦′ 𝑝̄(𝜉) cos 𝜉 𝑦 − 𝑦′ 𝑑𝜉𝑑𝑦′ , (A.8)
∫−𝑎 ∫0
and the pressure p(y) is rewritten as,
2𝑎𝑝(𝑦) 𝑔(𝑌 ′ ) where,
𝑝̂(𝑌 ′ ) = = √ . (A.2) ( )( )
𝑃 ( ) − 𝑘3 − 4𝜈𝜉 2 𝑘4 − 4𝜈𝜉 2
1 − (𝑌 ′ )2 𝐻𝐹 𝐺𝑀 𝜉, 𝑦, 𝑦′ = ( ⟨ ( ) ⟩) , (A.9)
2(𝜈 − 1) 𝑘4 − 4𝜈𝜉 2 + 𝑆𝑘
The normalized form of Eqs. (29) to (31) are, 2𝑚 𝜉
2 2 { [ ]}
𝑘3 𝜈 + 4𝜉 2 𝑚3 (𝜈 − 1)2 − 𝜈 2
( )2
1
1 ∞
𝑔(𝑌 ′ ) 4𝑌 𝑎 ( ) −1
𝐻̂ (𝑌 , 𝑌 ′ , 𝜂)𝑑𝜂 √ 𝑑𝑌 ′ = − , (A.3) 𝐻𝐻𝑜𝑚𝑜𝑔𝑒𝑛𝑒𝑜𝑢𝑠 𝜉, 𝑦, 𝑦′ = , (A.10)
𝜋 ∫−1 ∫0 𝜋 𝑎𝐻 𝜋(𝑆𝜉 + 1)
1 − (𝑌 ) ′ 2

and 𝑘3 = 𝑘2 𝑚2 (𝜈 − 1)(𝑚 − 1)2 , 𝑘4 = 𝑘2 𝑚2 (𝜈 − 1)(𝑚 + 1)2 . The y′ integra-

where,
tion is carried out using Gauss-Chebyshev quadrature [57], while the
sin(𝜂(𝑌 − 𝑌 ′ ))𝐾 𝐾2 Fourier domain integration (𝜉) is carried out in MAPLE© . The 𝜎 yy com-
𝐻̂ (𝑌 , 𝑌 ′ , 𝜂) = ( ). (A.4)
𝜂 𝑆𝐾 𝐾2 + 2(1 − 𝜈)𝐾 2 𝑚2 (𝑚 + 1)2 + 4𝜈 ∗ 𝜂 2 ponent of stress is calculated in two parts as given in Eq. (A.8), to cir-
cumvent the difficulty in numerical integration of HFGM (𝜉, y, y′) alone
1
1
𝑔(𝑌 ′ ) 2 with respect to 𝜉.
√ 𝑑𝑌 ′ = . (A.5)
𝜋 ∫−1 𝜋 The numerical procedure mentioned above has been benchmarked
1 − (𝑌 ) ′ 2
through reproduction of the numerical results in the works of Bakirtas,
It is evident from Eqs. (A.3) and (A.5) that the applied load P and the ra- Guler and Erdogan [14,23] when residual surface tension is absent (𝜏 𝑠 =
dius of the indenter R which are input to the problem posed in Fig. 1 can 0) and Long et al. [38], when the half-plane is homogeneous (k = 0).
be combined through the Hertzian semi-contact size aH to represent a
single input. Thus the solution to Eqs. (A.3) to (A.5) will be affected Supplementary material
only if any change in P or R alters aH . The results in Figs. 2 to 6 are
hence presented with aH as the input. Using Gauss-Chebyshev quadra- Supplementary material associated with this article can be found, in
ture and following the numerical procedure given by Erdogan and Gupta the online version, at 10.1016/j.ijmecsci.2017.11.009.
[57] Eqs. (A.3) and (A.5) can be transformed to system of algebraic equa-
References
tions as given in the work of Vasu and Bhandakkar [48]. On account of
the non-linear nature of m(𝜉), it is not possible to integrate the expres- [1] Henriques B. Inhomogeneous materials perform better: functionally graded ma-
sion H(y, y′, 𝜉) with respect to 𝜉 in closed form. Hence it is unavoidable terials for biomedical applications. J Powder Metall Mining 2013;2(3):e112.
to resort to numerical quadrature for the 𝜉 variable integration. In the doi:10.4172/2168-9806.1000e112.
[2] Suresh S. Graded materials for resistance to contact deformation and damage. Sci-
present work, Eq. (A.3) is calculated numerically using the “integral” ence 2001;292(5526):2447–51.
command in MAPLE© and the corresponding result is referred as ”Full- [3] Giannakopoulos A. Indentation of graded substrates. Thin Solid Films
Solution” in the main text. However for the special case of 𝑆 = 0, in the 1998;332(1):172–9.
[4] Mahamood R.M., Akinlabi E.T., Shukla M., Pityana S.. Functionally graded material.
limit of 𝜉 → ∞, m reduces to 2𝜉/k and H(y, y′, 𝜉) simplifies to, An overview.
[5] Niino M, Hirai T, Watanabe R. The functionally gradient materials. J Jap Soc Comp
𝑘𝜈 ∗ 𝜉𝑘
𝐻𝐴𝑝𝑝𝑟𝑜𝑥 (𝑦, 𝑦′ , 𝜉) = 1 − − Mater 1987;13(1):257.
2𝜉 (𝜉 + 𝑘∕2)2 + 𝑘2 𝜈 ∗ ∕4 [6] Müller E, Drašar v, Schilz J, Kaysser W. Functionally graded materials for sensor and
energy applications. Mater Sci Eng 2003;A 362(1):17–39.
𝑘2 (4𝜈 − 1) [7] Zhang Y. Overview: damage resistance of graded ceramic restorative materials. J
+ [ ]. (A.6)
European Ceramic Soc 2012;32(11):2623–32.
4(1 − 𝜈) (𝜉 + 𝑘∕2)2 + 𝑘2 𝜈 ∗ ∕4

166
T.S. Vasu, T.K. Bhandakkar
[8] Watari F, Yokoyama A, Omori M, Hirai T, Kondo H, Uo M, et al. Biocompatibility of
materials and development to functionally graded implant for bio-medical applica-
tion. Compos Sci Technol 2004;64(6):893–908.
[9] Pompe W, Worch H, Epple M, Friess W, Gelinsky M, Greil P, et al. Function-
ally graded materials for biomedical applications. Mater Sci Eng 2003;A 362(1):
40–60.
[10] Lu L, Chekroun M, Abraham O, Maupin V, Villain G. Mechanical properties esti-
mation of functionally graded materials using surface waves recorded with a laser
interferometer. NDT & E International 2011;44(2):169–77.
[11] Pender D, Padture N, Giannakopoulos A, Suresh S. Gradients in elastic modulus
for improved contact-damage resistance. part i: the silicon nitride–oxynitride glass
system. Acta Materialia 2001;49(16):3255–62.
[12] Jørgensen O, Giannakopoulos A, Suresh S. Spherical indentation of compos-
ite laminates with controlled gradients in elastic anisotropy. Int J Solids Struct
1998;35(36):5097–113.
[13] Jitcharoen J, Padture NP, Giannakopoulos AE, Suresh S. Hertzian-crack sup-
pression in ceramics with elastic-modulus-graded surfaces. J Am Ceramic Soc
1998;81(9):2301–8.
[14] Bakirtas I. The problem of a rigid punch on a non-homogeneous elastic half space.
Int J Eng Sci 1980;18(4):597–610.
[15] Giannakopoulos A, Suresh S. Indentation of solids with gradients in elastic proper-
ties: part i. point force. Int J Solids Struct 1997;34(19):2357–92.
[16] Giannakopoulos A, Suresh S. Indentation of solids with gradients in elastic proper-
ties: part ii. axisymmetric indentors. Int J Solids Struct 1997;34(19):2393–428.
[17] Suresh S, Giannakopoulos A, Alcala J. Spherical indentation of compositionally
graded materials: theory and experiments. Acta Materialia 1997;45(4):1307–21.
[18] Giannakopoulos A, Pallot P. Two-dimensional contact analysis of elastic graded ma-
terials. J Mech Phys Solids 2000;48(8):1597–631.
[19] Ke L-L, Wang Y-S. Two-dimensional contact mechanics of functionally graded ma-
terials with arbitrary spatial variations of material properties. Int J Solids Struct
2006;43(18):5779–98.
[20] Ke L-L, Wang Y-S. Two-dimensional sliding frictional contact of functionally graded
materials. Euro J Mech-A/Solids 2007;26(1):171–88.
[21] Ke L-L, Wang Y-S. Fretting contact of two dissimilar elastic bodies with functionally
graded coatings. Mech Adv Mater Struct 2010;17(6):433–47.
[22] Liu J, Ke L-L, Wang Y-S. Two-dimensional thermoelastic contact problem of
functionally graded materials involving frictional heating. Int J Solids Struct
2011;48(18):2536–48.
[23] Guler M, Erdogan F. Contact mechanics of graded coatings. Int J Solids Struct
2004;41(14):3865–89.
[24] Guler M, Erdogan F. Contact mechanics of two deformable elastic solids with graded
coatings. Mech Mater 2006;38(7):633–47.
[25] Guler M, Erdogan F. The frictional sliding contact problems of rigid parabolic and
cylindrical stamps on graded coatings. Int J Mech Sci 2007;49(2):161–82.
[26] Alinia Y, Beheshti A, Guler MA, El-Borgi S, Polycarpou AA. Sliding contact analysis
of functionally graded coating/substrate system. Mech Mater 2016;94:142–55.
[27] Comez I, Guler M. The contact problem of a rigid punch sliding over a functionally
graded bilayer. Acta Mechanica 2017;228(6):2237–49.
[28] Chidlow S, Chong W, Teodorescu M. On the two-dimensional solution of both ad-
hesive and non-adhesive contact problems involving functionally graded materials.
Eur J Mechanics-A/Solids 2013;39:86–103.
[29] Chen P, Chen S, Peng J. Frictional contact of a rigid punch on an arbitrarily oriented
gradient half-plane. Acta Mechanica 2015;226(12):4207–21.
[30] Peijian C, Shaohua C, Juan P. Sliding contact between a cylindrical punch
and a graded half-plane with an arbitrary gradient direction. J Appl Mech
2015;82(4):041008.
[31] Çömez I. Contact problem for a functionally graded layer indented by a moving
punch. Int J Mech Sci 2015;100:339–44.
167
International Journal of Mechanical Sciences 135 (2018) 158–167
[32] Zhang GP, Sun KH, Zhang B, Gong J, Sun C, Wang ZG. Tensile and fatigue strength
of ultrathin copper films. Mater Sci Eng: A 2008;483–484:387–90.
[33] Gurtin ME, Murdoch AI. A continuum theory of elastic material surfaces. Arch Ra-
tional Mech Anal 1975;57(4):291–323.
[34] Gurtin ME, Murdoch AI. Surface stresses in solids. Int J Solids Struct
1978;14(6):431–40.
[35] Miller RE, Shenoy VB. Size-dependent elastic properties of nanosized structural ele-
ments. Nanotechnology 2000;11(3):139–47.
[36] Gao X, Hao F, Fang D, Huang Z. Boussinesq problem with the surface effect
and its application to contact mechanics at the nanoscale. Int J Solids Struct
2013;50(16–17):2620–30.
[37] Zhao XJ, Rajapakse RKND. Elastic field of a nano-film subjected to tangential surface
load: asymmetric problem. Eur J Mechan A/Solids 2013;39:69–75.
[38] Long JM, Wang GF, Feng XQ, Yu SW. Two-dimensional hertzian contact problem
with surface tension. Int J Solids Struct 2012;49(13):1588–94.
[39] Long JM, Wang GF. Effects of surface tension on axisymmetric hertzian contact prob-
lem. Mech Mater 2013;56:65–70.
[40] Pinyochotiwong Y, Rungamornrat J, Senjuntichai T. Rigid frictionless indentation
on elastic half space with influence of surface stresses. Int J Eng Sci 2013;71:15–35.
[41] Ou ZY, Pang SD. Fundamental solutions to hertzian contact problems at nanoscale.
Acta Mechanica 2013;224(1):109–21.
[42] Gad A, Mahmoud F, Alshorbagy AE, Ali-Eldin S. Finite element modeling for elas-
tic nano-indentation problems incorporating surface energy effect. Int J Mech Sci
2014;84:158–70.
[43] Long J, Chen W. Effects of surface tension on the nanoindentation with a conical
indenter. Acta Mechanica 2017:1–10.
[44] Xu X, Jagota A, Hui C-Y. Effects of surface tension on the adhesive contact of a rigid
sphere to a compliant substrate. Soft Matter 2014;10(26):4625–32.
[45] Gao X, Hao F, Huang Z, Fang D. Mechanics of adhesive contact at the nanoscale: the
effect of surface stress. Int J Solids Struct 2014;51(3):566–74.
[46] Hui C-Y, Liu T, Salez T, Raphael E, Jagota A. Indentation of a rigid sphere
into an elastic substrate with surface tension and adhesion. Proc R Soc A
2015;471:20140727. The Royal Society.
[47] Long J, Wang G, Feng X-Q, Yu S. Effects of surface tension on the adhesive contact
between a hard sphere and a soft substrate. Int J Solids Struct 2016;84:133–8.
[48] Vasu TS, Bhandakkar TK. A study of the contact of an elastic layer–substrate sys-
tem indented by a long rigid cylinder incorporating surface effects. J Appl Mech
2016;83(6):061009.
[49] Attia MA, Mahmoud FF. Analysis of nanoindentation of functionally graded layered
bodies with surface elasticity. Int J Mech Sci 2015;94:36–48.
[50] Rahman AAA, Mahmoud FF. Analysis of nanocontact problems of layered viscoelas-
tic solids with surface energy effects under different loading patterns. Acta Mechan-
ica 2016;227(2):527.
[51] Timoshenko SP, Goodier JN. Theory of Elasticity. New York: McGraw-Hill; 1987.
[52] Zhao XJ, Rajapakse RKND. Analytical solutions for a surface-loaded isotropic elastic
layer with surface energy effects. Int J Eng Sci 2009;47(11–12):1433–44.
[53] Zhou S, Gao XL. Solutions of half-space and half-plane contact problems
based on surface elasticity. Zeitschrift fr angewandte Mathematik und Physik
2013;64(1):145–66.
[54] Sneddon IN. Fourier Transforms. New York: Dover Publications; 1995.
[55] Johnson KL. Contact Mechanics. UK: Cambridge University Press; 1985.
[56] Gupta PK, Walowit J. Contact stresses between an elastic cylinder and a layered
elastic solid. J Tribol 1974;96(2):250–7.
[57] Erdogan F, Gupta G. On the numerical solution of singular integral equa-
tions(singular integral equations numerical solution from Gauss–Chebyshev formu-
las for mixed boundary value problems). Q Appl Math 1972;29:525–34.
[58] Abramowitz M, Stegun IA, et al. Handbook of mathematical functions with formulas,
graphs, and mathematical tables, Vol. 9. New York: Dover; 1972.