Composite Structures 92 (2010) 254–264

Contents lists available at ScienceDirect

Composite Structures
journal homepage: www.elsevier.com/locate/compstruct

Deformation in viscoelastic sandwich composites subject to moisture diffusion

Nikhil Joshi, Anastasia Muliana *
Department of Mechanical Engineering, Texas A and M University, College Station, TX 77843-3123, USA

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

Article history: This study analyzes the effect of moisture diffusion on the deformation of viscoelastic sandwich compos-
Available online 23 July 2009 ites, which are composed of orthotropic fiber-reinforced laminated skins and viscoelastic polymeric foam
core. It is assumed that the elastic and time-dependent (transient) moduli at any particular location in the
Keywords: foam core depend on the moisture concentration at that location. Sequentially coupled analyses of mois-
Viscoelastic ture diffusion and deformation are performed to predict overall performance of the studied viscoelastic
Moisture diffusion sandwich systems. Time and moisture dependent constitutive model is used for the polymer foam core,
Sandwich composite
while skins are assumed linear elastic. The overall time-dependent responses of the sandwich composites
subject to moisture diffusion are analyzed using finite element (FE) method. Experimental data available
in the literature and analytical solutions are used to support convergence studies in the FE analyses. Con-
tributions of moisture dependent elastic and the time-dependent moduli to the overall stress, strain and
displacement field are studied. FE analyses of the delamination between skins and core in sandwich com-
posite under combined moisture diffusion and mechanical loading are also performed.
Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction Koeing [3] showed that the mechanical properties of glass rein-
forced epoxy composites are irreversibly degraded when exposed
Polymer foam cored sandwich composites are widely used in to moist environment. Allred [2] studied the effect of moisture
load-bearing components in buildings and naval structures be- on flexural elastic response of kevlar/epoxy laminate with 0°, 90°
cause they can provide high strength and stiffness without signif- fiber orientations under three-point bending test. The experimen-
icantly adding weight and the cost. To improve reliability in using tal results show a reduction in the maximum load carrying capac-
sandwich composites, it is essential to understand their mechani- ity and the deflections with the increasing moisture concentration.
cal responses in terms of stresses and deformations under external Gibson et al. [15] and Botelho et al. [6,7] investigated the effects of
mechanical and environmental stimuli. One of the prominent char- moisture on the dynamic properties of glass fiber reinforced com-
acteristics of polymers is their viscoelasticity. The viscoelastic posites subject to mechanical vibration. Gibson et al. [15] observed
behavior of polymers is even more aggravated under high stress that the moisture induced changes in damping were fully revers-
levels and extreme environmental conditions. Extreme tempera- ible and no permanent structural changes occurred. Woldesenbet
ture changes and humid environmental conditions can signifi- et al. [37] analyzed the effect of high strain rate (500–1000 s1)
cantly degrade the stiffness and strength of the polymer foam on the mechanical properties of polymer matrix composite im-
core, which intern degrades the performance of the entire sand- mersed in water. In contrast to the quasi-static loadings, the
wich structure. However, the effects of combined time–stress– mechanical properties of the composite fully saturated with mois-
temperature–moisture in the degradation of polymer foam on ture had increased under the high strain rates. Bouadi and Sun [8]
the performance of sandwich structures have not been fully combined the theoretical and experimental work concerned with
understood. predicting the hygrothermal effects on the stress field of some lam-
External temperature and moisture play important roles in inated composite plates under hygothermal loading. It was ob-
determining the long-term behavior of polymer matrix composite. served that hygrothermal loading induced very high stresses in
Shen and Springer [30,31] and Springer et al. [32] conducted exper- cross ply laminates. Weitsman [35] reviewed several investiga-
iments to reveal the significant influence of moisture on tensile tions regarding anomalous fluid sorption in polymeric composites
strength, tensile modulus, shear strength, shear modulus, and com- to discuss two versions of coupled diffusion/damage model which
pression behaviors of graphite–epoxy and glass–epoxy fiber rein- provides a reasonable explanation of the experimental findings.
forced composites with different fiber orientations. Antoon and Observations made using Scanning Electron Microscope (SEM) pro-
vided consistent evidence of fluid-induced damage occurred in the
* Corresponding author. Tel.: +1 979 458 3579. form of interfacial debondings at the fiber/matrix interfaces.
E-mail address: [email protected] (A. Muliana). Weitsman and Guo [36] have discussed the detailed mechanism

0263-8223/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compstruct.2009.07.021
 N. Joshi, A. Muliana / Composite Structures 92 (2010) 254–264 255

of the coupled diffusion/damage model, where in damage in the developed incorporates temperature dependent elastic properties
form of distributed fiber/matrix interfacial microcracks is caused of the constituents. Pradeep et al. [27] compared the vibration
by residual hygrothermal and mechanically induced stresses. and thermal buckling responses of sandwich composites having a
Several experiments have been performed to investigate the ef- viscoelastic foam core using the beam and plate finite elements.
fect of moisture on the mechanical properties of polymeric foams The temperature effects were incorporated through coefficient of
used as cores in many sandwich composites. Patel et al. [26] thermal expansions (CTE). Different CTEs of skins and core create
observed that moisture degrades foamed siloxane composite poly- thermal stresses, which can be significant with increasing temper-
mers both chemically and physically. Presence of moisture resulted ature changes.
in an accelerated stress relaxation in the polymer. It is assumed This study analyzes the effect of moisture diffusion on the
that moisture aids chain rearrangement through plasticization. deformation of viscoelastic sandwich composites composed of lin-
Ashby and Gibson [4] showed that in case of the open cell poly- ear elastic skins and viscoelastic foam core. Sequentially coupled
meric foams the presence of the pore fluid produces strong strain analyses of moisture diffusion and deformation are performed to
rate dependence on the strength. This is because of the fluid viscos- predict overall time-dependent performance of the sandwich
ity which necessitates work in forcing the fluid through the inter- beams. As suggested in existing experimental studies, moisture
connected porosity of the foam. Ionita and Weitsman [20] concentration significantly reduces mechanical properties and life-
proposed a model to conclude that the ingress of sea water into time of composites. Instead of incorporating the effect of moisture
polymeric foams resulted into substantial weight gain ratios and in the deformation of sandwich beam through the coefficient of
noticeable damage to cellular foam walls. The water ingress phe- moisture expansion, which is often presented in the literature, this
nomenon was assumed to be governed by the cell wall breakage study considers degradation of the viscoelastic material properties
(sequential time-dependent buckling), rather than by the diffusion of the foam core with moisture concentration. The main purpose is
process. Numerical simulations presented using the proposed to investigate the effect of degradation in the viscoelastic foam
model indicates that the fluid ingress into the foam depends on core on the overall deformation of sandwich beam during transient
the material properties of the cell walls and the structural config- moisture diffusion process. We neglect the deformation due to the
uration at the cellular level. coefficient of moisture expansion on the overall deformation in the
Limited work has been done until date to understand the effect viscoelastic sandwich beams in order to isolate the moisture-
of environmental condition on the performance of sandwich com- dependent viscoelastic responses from other hygro-mechanical
posites made up of fiber-reinforced laminated skins and polymeric coupling effects. The overall non-linear time-dependent responses
core. Granville [16] conducted experiment to carry out a compara- of sandwich composites are analyzed using finite element (FE). For
tive study on the effects of moisture and temperature on the problems with complex geometry, loading histories, and boundary
performance of sandwich composites composed of polymeth- conditions it is not always possible to obtain an exact closed form
acrylimide (PMI) foams and honeycomb foams glued to glass/ analytical solution and it is for these reasons we present FE analy-
epoxy facesheets with intent to understand the effects of moisture ses. Convergence studies are first performed to assure accuracy in
on structural, dimensional stability, weight gain and peel strength the FE analyses. Analytical solution and available experimental
of sandwich composites. It was observed that moisture absorbed data on moisture diffusion and creep deformation in sandwich
by the PMI foam core sandwich structure was significantly higher composites are used to validate the FE analyses. Parametric studies
than the honeycomb core panels. However, the torque required for are conducted on understanding the effects of degradation in the
peeling did not suffer much due to the migration of moisture. Ish- time-dependent moduli due to moisture diffusion on the overall
iaku et al. [19] conducted experiments to correlate the effect of mi- responses of sandwich composites. An example of delamination
cro-structural and environmental conditions on the tensile and between skins and foam core during moisture diffusion is exam-
fracture properties of hybrid composites of Polyamide 6. The ined as part of FE analyses.
experiments were conducted on a set of samples exposed to ambi-
ent conditions of temperature and humidity (conditioned samples)
and another set of unconditioned samples. Significant decrease in 2. Moisture diffusion through the sandwich composite
the tensile modulus, tensile strength, and fracture toughness was
observed in the conditioned samples as compared to uncondi- To investigate the effect of moisture diffusion on the degrada-
tioned samples. The conditioned sampled exhibited a higher elon- tion of the composite materials and its subsequent response,
gation at break as compared to unconditioned samples. simultaneous study of the problem of moisture diffusion (which
Several analytical and numerical solutions have been presented is assumed to be governed by unsteady Fick’s equation) and defor-
on the deformation of elastic sandwich composites under various mation of viscoelastic materials is carried out. The equations for
environmental conditions. Doxsee et al. [13] evaluated in-plane the motion of the fluid and the motion of the composite are cou-
elastic deformations and curvature changes in the sandwich panels pled as the material moduli depend on the concentration of the
composed of honeycomb core and fiber-reinforced laminated skin, fluid (moisture). Diffusion in polymer systems is generally non-Fic-
due to non-uniform temperature and moisture distributions using kian accompanied by damage formation when subjected to exter-
the classical lamination theory. Morganti et al. [24] also analyzed nal stress. The non-Fickian diffusion can adequately capture the
the effects of moisture on the dimensional stability of sandwich changes to the polymer system induced by the absorbed penetrant
composites. It was concluded that moisture and temperature af- such as chemical degradation, swelling and/or changes induced
fects the physical behavior of the composite directly by modifying due to microcracks/damage [9,11] and Weitsman [35]. This is gen-
its structural characteristics such as degrading the matrix elastic erally the case for glassy polymers, which show anomalous fluid
properties and inducing microcracks between fibers and matrix. sorption behaviors. The anomalous behaviors are considered due
The geometric stability of the composites is affected by the ‘‘swell- to relaxation characteristics of the polymer macromolecules and
ing” phenomenon attributed to the moisture absorption and can internal stresses formed during the sorption/desorption. Despite
induce stress effects. Frostig and Thomsen [14] performed analyt- the diffusion process being generally non-Fickian, in absence of
ical and numerical studies to evaluate the non-linear response of damage or negligible damage, the diffusion of moisture through
unidirectional sandwich panels under localized thermally induced the composite solids can be assumed to be Fickian. The use of
deformation and the interaction of mechanical and thermal stimuli the Fickian diffusion model tends to overestimate the moisture
on the response sandwich composites. The mathematical model absorption for a short diffusion time. This study assumes that the
256 N. Joshi, A. Muliana / Composite Structures 92 (2010) 254–264

moisture diffusion through the sandwich composite is governed by considered for the purpose of the convergence study. Moisture con-
the unsteady Fick’s law of diffusion. The concentration of the fluid centration is applied at the surface of the layer 1 as shown in Fig. 1. It
at any point (x, y, z) in the sandwich structure at time t is denoted is assumed that the moisture diffuses from the surface of the layer 1,
by c = c (x, y, z, t). The rate of the diffusing fluid related to the con- through the thickness of the composite into the layer 2. No other
centration gradient through the three-dimensional form of Fick’s mode of diffusion such as moisture diffusion from the sides of the
law of diffusion is written as: composite is considered.
The solution of moisture diffusion through the two-layered
@c
¼ r  ½Diff rc ð1Þ composite is derived using the unsteady Fick’s law of diffusion as
@t
stated in Eq. (2) by prescribing proper initial and boundary value
Here Diff is a second order diffusivity tensor and r denotes the gra- conditions, which are written as:
dient operator. In the case when the diffusing fluid is more pro-
cð1Þ ðl; tÞ ¼ C 0 ; tP0
nounced in one direction (x-direction) and the diffusivity of the
material is constant, the three-dimensional form of Fick’s law of dif- cð2Þ ða; tÞ ¼ 0; tP0
ð3Þ
fusion can be reduced to one-dimensional form which is stated as: cð1Þ ðx; 0Þ ¼ 0; l < x 6 0
@c cð2Þ ðx; 0Þ ¼ 0; 06x6a
¼ Diff r2 c ð2Þ
@t The conditions of moisture concentration and flux of the fluid at
Here the diffusivity Diff is a scalar. Thus the concentration of the the interface being continuous yield the following interfacial
fluid is now a function of only one spatial coordinate and time conditions:
written as c = c (x, t). Weitsman [34] incorporated the effect of cð1Þ ð0; tÞ ¼ cð2Þ ð0; tÞ; t>0
stress-assisted diffusion in homogeneous materials by allowing ð1Þ ð2Þ ð4Þ
ð1Þ @c ð2Þ @c
the diffusivity to depend on the stress. Diff ð0; tÞ ¼ Diff ð0; tÞ; t>0
@x @x
This study examines responses of viscoelastic sandwich com-
posites during moisture diffusion. The elastic material properties Superscripts (1) and (2) denote the layer number as shown in Fig. 1.
of the studied sandwich beam change with time, as in viscoelastic c(i) denotes the moisture concentration at any point in the ith layer
ðiÞ
materials, and with moisture concentrations. We monitor time- and Diff denotes the moisture diffusivity of the ith layer, where i = 1
dependent deformation, strain, and stress fields. Our intention is or 2.
not to solve a particular problem of moisture diffusions. Thus, we The solution to Eq. (2) for a two-layered composite using the
take the simplest form of moisture diffusion problems by imposing boundary and the initial condition as in Eqs. (3) and (4), respec-
continuous concentrations1 at the interfaces, between the skins and tively, is written as:
core, and one-dimensional through the thickness diffusion process,
cð1Þ " #
which are acceptable assumption elsewhere. We ignore the potential ð1Þ ð2Þ
aDiff  xDiff
formation of microcrack, existence of contact resistant and imperfect ¼ C0 ð1Þ ð2Þ
bonding, and the actual three-dimensional diffusion. A FE approach aDiff þ lDiff
is used to obtain the solution to the coupled moisture diffusion
"      ð1Þ 2 #
P
1 cosðbm xÞsin bmR a  Rsinðbm xÞ cos bmR a eDiff bm t
and deformation in the viscoelastic sandwich composite. Conver- 2C 0  bm a bm a  a 
m¼1 bm sinðbm lÞ sin R ða þ lÞ  cosðbm lÞcos R þ Rl
gence studies in terms of mesh size, tolerance, and time-increments R

are performed for the FE model and the responses of the moisture
concentration fields are compared with the ones obtained from ana- cð2Þ " #
lytical solution. In order to reduce the complexity in developing the ða  xÞDiff
ð1Þ

analytical solution to the moisture diffusion problem of the ¼ C0 ð1Þ ð2Þ

sandwich (three-layered) composite, a two-layered composite is aDiff þ lDiff
2   ð1Þ 2 3
P1 sin bm ðaxÞ
R
eDiff bm t
1
Weistman [35] discussed that the jump discontinuity in the moisture concen- 2C 0  4  bm a bm aa 5
m¼1 bm sinðbm lÞsin R ða þ lÞ  cosðbm lÞcos R R
þ Rl
tration at the interface of the composite layers arouse due to damage in the form of
microcracks during the fluid sorption. These microcracks draw a portion of the
diffusing fluid, like in a capillary process, and enhance the capacity of fluid vffiffiffiffiffiffiffiffi
concentration, which results in jump discontinuity in the moisture concentration at u ð2Þ
uD
the interface. In order to describe this anomalous behavior, it is assumed that the where R ¼ t iff
ð1Þ
ð5Þ
chemical potential at the interface of the composite layers should be continuous. Diff
Upon imposing the continuity condition of the chemical potential, the concentration
condition at the interface of the composite layers takes the following form: Here bm are the roots of Eq. (6), which is written as:
c1 ðxin ; tÞ ¼ a c2 ðxin ; tÞ þ b, which involves two constants a and b. Variables c1 and c2

are concentration in materials 1 and 2, respectively, and xin is the position of the bm a b a
interface. The above condition is only one type of concentration conditions at cosðbm lÞ sin þ R sinðbm lÞ cos m ¼0 ð6Þ
R R
the interfaces that can be used and it is not the only concentration condition at the
interfaces that must be satisfied. This condition gives sensible response predictions in
describing anomalous fluid sorption/desorption in glassy polymer composite. In
practice, the rate of diffusion between two surfaces is proportional to the difference in
the concentrations of the two surfaces. Thus, the concentration condition at the
interface should have the following form: D1 @c @x ¼ Hðc 1 ðxin ; tÞ  c 2 ðxin ; tÞÞ, as
1

discussed in Crank [11], Carslaw and Jaeger [10], Mikhailov and Ozisik [23]. D1 is
the diffusivity of material 1, H is the surface diffusivity/conductivity, and x is the
normal direction to the surfaces. It has also been discussed in many articles and books
on moisture diffusion and heat conduction problems in heterogeneous materials,
when the contact resistance at the interfaces is negligible, damage formation during
diffusion is ignored, and the interfaces are perfectly bonded, the fluid concentrations
at the interfaces can be assumed continuous ([10,11,23]). This is the simplest form of
moisture diffusion conditions by setting the parameters a and b in the former
equation to be 1 and 0, respectively. Fig. 1. A schematic of the two-layered composite.
 N. Joshi, A. Muliana / Composite Structures 92 (2010) 254–264 257

Fig. 2 shows the moisture diffusion profile through the thick- Table 3
ness of the two-layered composite at various times beginning from Effect of different tolerance on the convergence rate.

time t = 0 s to steady state time of t = 1828 s for an applied concen- Tolerance Steady state time (h) Moisture concentration at the furthest distance
tration of C0 = 0.5606 gm/mm3. Table 1 gives the dimensional de- 1.00E05 12.22 0.717
tails and the material properties. A finite element (FE) analysis is 1.00E07 49.72 0.997
done on the same problem using 2D solid continuum elements 1.00E08 68.33 0.9997
(DC2D4 elements form ABAQUS element library). The moisture 1.00E09 86.94 0.999972
1.00E11 117.12 1.0
profiles obtained from the FE analyses are in good agreement with
the analytical solution in Eq. (5).
A similar method can also be used to obtain the analytical solu-
tion to the problem of moisture diffusion for the sandwich com- into insignificant resolution at the cost of large computational
posite, which can be found in [23]. With the proper boundary time. It is concluded that an initial time increment can be chosen
conditions similar to Eqs. (3) and (4) for a three layered (sandwich) in the range: 10–50 s in order to obtain an accurate time required
structure and an applied concentration of 1 gm/mm3 on the sur- to reach the steady state to the current moisture diffusion problem.
face of the top skin, the steady state is said to have reached when Another parameter determining the steady state is the tolerance on
the concentration throughout the sandwich structure reaches a va- the moisture diffusion. The value of this parameter indicates that if
lue of 1 gm/mm3. The chosen time increment in the FE analyses all the nodal moisture concentrations are changing at less than a
can significantly affect the accuracy of the numerical results. Table certain rate (tolerance) then the FE analysis considers to have
2 shows the effect of different time increment size on the time re- reached the steady state. Table 3 shows the effect of different tol-
quired to reach the steady state. The larger time increments tend to erances on the time required to reach the steady state. A tolerance
predict a higher time to reach the steady state. It is observed that can be chosen in the range: 1e7–1e11 in order to represent
on reducing the time increment convergence in terms of steady moisture concentration profile throughout the studied sandwich
state time is achieved. Further reduction of time increments results composite.

3. Deformation in the viscoelastic sandwich composite

The sandwich composites are generally made up of metallic or

fiber-reinforced laminated skins and polymeric cores. These poly-
meric cores can exhibit significant viscoelastic responses. In this
study, a generalized single integral constitutive relation is used
to model the linear viscoelastic behavior of the polymeric core of
the sandwich structure. The general single integral constitutive
relation for a linear viscoelastic material under uniaxial loading
(non-aging) incorporating dependence of the material moduli on
the moisture concentration is written as:
Z t
drs
et  eðtÞ ¼ g 0 ðct ÞD0 rt þ g 1 ðct Þ DDðt  sÞ ds ð7Þ
0 ds
Here D0 and DD are the uniaxial instantaneous and transient
creep compliances. Where the uniaxial transient compliance in
terms of the Prony series coefficients is expressed as:
Fig. 2. Evolution of the moisture diffusion profile through the two-layered
composite. X
N
DDt ¼ Dn ð1  exp½kn tÞ ð8Þ
n¼1

Table 1
Here N is the number of terms, Dn is the nth coefficient of the Prony
Material properties and dimensional details of the two-layered composite. series, and kn is the nth reciprocal of creep time. The parameters g0,
g1 measure the increase or decrease in the instantaneous and the
Properties and dimensions Value
transient compliance with the change in moisture concentration,
Moisture diffusivity of the first layer,
ð1Þ
Diff 2
(mm /s) 0.1144 respectively. The parameter ct represents the current moisture con-
ð2Þ
Moisture diffusivity of the second layer, Diff (mm2/s) 0.2 centration. The material properties at any given location and time
Thickness of the first layer, l (mm) 1 are assumed to depend on the moisture at that location at that gi-
Thickness of the second layer, a (mm) 25.4
ven instant of time. Thus, the parameter g1(ct) is out of the integral
as we do not incorporate the history of moisture diffusion process
on the deformation of sandwich beams. When it is necessary to in-
clude effect of history of moisture diffusion on the viscoelastic
Table 2 material response, then the parameter g1(ct) is placed inside the
Effect of different initial time increment size.
integral and two integrands with different time variables of integra-
Initial time increment (s) Steady state time (h) tions are required. It is also possible to include the effect of coeffi-
4000 52.22 cient of hygro expansion to the time-dependent strain in Eq. (7),
2000 50.56 as presented in Muliana and Khan [25].
1000 49.72 The uniaxial relation in Eq. (7) is generalized for multi-axial
500 49.17
200 48.83
(3D) constitutive relations of isotropic materials [17] by separating
100 48.72 the deviatoric and volumetric stress–strain relations as:
50 48.67 Converged
10 48.66
258 N. Joshi, A. Muliana / Composite Structures 92 (2010) 254–264

1 Table 5
etij ¼ etij þ etkk dtij Elastic properties of aluminum facings.
3
Z t s
1 1 dSij Elastic properties Value
etij ¼ g 0 ðct ÞJ 0 Stij þ g 1 ðct Þ DJðt  sÞ ds ð9Þ
2 2 0 ds E (MPa) 70,000
Z t m 0.29
t 1 1 drs
e kk ¼ g 0 ðct ÞB0 rtkk þ g 1 ðct Þ DBðt  sÞ kk ds
3 3 0 ds
The parameters J0 and B0 are the instantaneous elastic shear and
bulk compliances, respectively. The transient shear and bulk com-
pliances are represented by DJ and DB, respectively. The corre-
sponding linear elastic Poisson’s ratio, t, for viscoelastic materials
is assumed to be time independent, as a result of which the shear
and bulk compliances are expressed as:

J 0 ¼ 2ð1 þ tÞD0 B0 ¼ 3ð1  2tÞD0

t t
ð10Þ
DJ ¼ 2ð1 þ tÞDD DBt ¼ 3ð1  2tÞDDt
Eq. (9) is solved using a recursive method and it is designed to be
compatible with a general displacement based FE framework.
The solution to the deformation of a viscoelastic sandwich
structure under a three-point bending test can be obtained using
the elasticity or to some extent strength of material approach. Here
the strength of material approach has been used. For a sandwich
beam, under three-point bending, with the span length L, width Fig. 3. Deformation at the midspan of the viscoelastic sandwich beam.
b, face and core thickness of f and c, respectively, and Young’s mod-
uli of the face and core to be Ef and Ec, respectively, the total deflec-
tion at the midspan of the beam is written as: thick polyurethane foam core is subjected to three-point bending
3 test with point load of 250 N applied at the half span of the beam.
PL PL
d¼ þ ð11Þ The creep properties of the polyurethane core are taken from
48ðEIÞeq 4ðAGc Þ Huang and Gibson [18]. The calibrated time-dependent parameters
Here A is the cross sectional area and Gc is the shear modulus of the (Prony series) representing the creep compliance up to 1200 h and
core. P is the applied load at the midspan. EIeq is the equivalent flex- elastic properties for the polyurethane foam core are given in Table
ural rigidity written as: 4. The elastic properties of aluminum are given in Table 5. Fig. 3
shows the deformation at the midspan of the viscoelastic sandwich
2
Ef bfc beam for two different mesh densities. The two dimensional con-
ðEIÞeq ¼ ð12Þ
2 tinuum plane strain elements (CPE4) in ABAQUS are used. A uni-
form mesh (relatively coarser) is used in case of FEA model 1,
Since the sandwich under consideration consists of a linear vis-
whereas a finer mesh in the region near the point of application
coelastic foam core and linear elastic facings, considering the time-
of the load and the constraints is used in FEA model 2. FE model
dependent behavior of the foam core, shear component of the total
with finer mesh can capture better localized effects of the point
deflection increases with time due to the decrease in the shear
load and point supports. For sandwich beam with the viscoelastic
modulus. Since the decreasing value of the material modulus of
core, the response to the applied static (creep) load is time-depen-
the viscoelastic core has negligible effect on the equivalent flexural
dent and hence the size of the initial time increment significantly
rigidity the bending component is assumed to remain constant
affects the instantaneous (elastic) material response. It is con-
with time. Thus, the creep deflection of the beam as shown by
cluded that an initial time increment can be chosen in the range:
Huang and Gibson [18] is written as:
0.001–0.01 s in order to represent an instantaneous response of
PL3 PLJc ðtÞ the sandwich beam.
dðtÞ ¼ þ ð13Þ Responses from FE analyses of three-point bending sandwich
48ðEIÞeq 4ðAÞ
beam are also compared with limited experimental data available
Here Jc(t) is the shear creep compliance of the foam core. The in the literature. A sandwich beam composed of carbon–epoxy fac-
analytical solution in Eq. (13) is used to validate the FE results. ings and Divinycell H-250 core is subjected to a three-point bend-
To accomplish this, a 254 mm long and 25.4 mm wide sandwich ing test with a load of 2018 N at the center of the beam. Table 6
beam composed of 1 mm thick aluminum facings and 25.4 mm
Table 6
Elastic properties of carbon–epoxy laminate and Divinycell H250.
Table 4
Elastic properties and Prony series coefficients for polyurethane foam core. Material Properties Values

n kn (s1) Dn  103 (MPa1) Core: Divinycell H250 Youngs modulus, E (MPa) 227.5
Poisson’s ratio, m 0.19
1 1 1.5
Skin: carbon–epoxy composite Young’s modulus, E11 (MPa) 147,000
2 101 2
(AS4/3501-6) Young’s modulus, E22 (MPa) 10,000
3 102 2.5
Young’s modulus, E33 (MPa) 10,000
4 103 4.25
Shear modulus, G12 (MPa) 7600
5 104 8.4
Shear modulus, G13 (MPa) 7600
6 105 10.68
Shear modulus, G23 (MPa) 3355
7 106 11.45
Poisson’s ratio, m12 0.27
8 107 24
Poisson’s ratio, m13 0.27
E = 31 MPa m = 0.35 Poisson’s ratio, m23 0.49
 N. Joshi, A. Muliana / Composite Structures 92 (2010) 254–264 259

the viscoelastic sandwich composite with moisture dependent

material properties are determined. It must also be noticed that
the analysis does not address a fully coupled problem. While the
deformation depends on the moisture concentration, the moisture
concentration is obtained without the knowledge of stress/strain.
The moisture diffusion process is not stress or strain assisted.
The instantaneous modulus (E0) of the polyurethane foam core is
assumed to degrade linearly with moisture concentration:
1 c
E0 ¼ ¼ Eref  a ð14Þ
D0 C0
where C0 (gm/mm3) is the applied moisture concentration, a (MPa)
is the slope of the graph of instantaneous modulus E0 (MPa) versus
non-dimensionalized moisture concentration c/C0. Eref is the instan-
taneous modulus at a reference condition. Therefore, the non-
Fig. 4. Deflection at the mid span of the sandwich beam.
dimensional compliance parameter g0 which accounts for the deg-
radation in Eq. (9) can be written as:
shows the elastic properties of carbon–epoxy faces and the Diviny-
Eref
cell H-250 core. It is assumed that the facings and the core are lin- g0 ¼ ð15Þ
Eref  a Cc0
ear elastic materials. The total deflection at the midspan of the
beam obtained is shown in Fig. 4. The results from FE code are in The specific choice of the type of degradation is made to merely
good agreement with the experimental data ensuring the adequacy illustrate that the strain induced due to the degradation in the
of the number of elements used in the mesh. material stiffness can be pronounced. For more severe degradation
than the linear form, the magnitude of strains could be much high-
4. Coupled moisture diffusion and deformation in the er. The degradation behavior should be experimentally character-
viscoelastic sandwich beam ized. In this case the value of a is chosen to be 12 MPa, which
assumes 50% degradation in elastic modulus of the polymer foam
The sandwich beam composed of orthotropic linear elastic car- core at a fully saturated condition. With the moisture diffusivities
bon–epoxy laminated skin and isotropic viscoelastic polyurethane of the carbon–epoxy facings is assumed to be 1.06E03 mm2/s and
foam core is subjected to simultaneous moisture prescribed at the that of the polyurethane core equal to 1.06E02 mm2/s [12] and an
surface of the top skin and a point load at the midspan of the beam
in a three-point bending test as shown in Fig. 5. Deformation in vis-
coelastic sandwich beam at fixed environmental condition contin-
ues to grow with time until an asymptotic value is reached. In this
study, we investigate the effect of degradation in viscoelastic foam
core with moisture diffusion on the overall performance of the
sandwich beam. For this purpose, the effect of moisture is incorpo-
rated only in the moisture-dependent viscoelastic model for the
foam core. The effects of free expansion due to moisture concentra-
tion, which is carried by the coefficient of hygro expansion, and
degradation in the carbon/epoxy skins with moisture, are ne-
glected. If all the above effects are incorporated in analyzing defor-
mation of the sandwich beam, stronger non-linear time-dependent
responses should be expected.
The solution to the coupled displacement-transport equations is
obtained in two sequential steps. In the first step, the moisture dif-
fusion profile through the sandwich beam is obtained. Once the
moisture distribution throughout the sandwich beam during the Fig. 6. Evolution of concentration of diffusing moisture.
transient analysis is known, the time-dependent deformations of

Fig. 5. Schematic of sandwich composite under simultaneous moisture and

mechanical stimuli. Fig. 7. Comparison of transverse displacement using 2D and 3D elements at 1200 h.
260 N. Joshi, A. Muliana / Composite Structures 92 (2010) 254–264

applied moisture concentration of 1 gm/mm3 at the surface of the

top skin, the evolution of the concentration of moisture is plotted
in Fig. 6. The steady state condition is reached at 432 h.
The degradation of the instantaneous modulus (E0) due to the
moisture concentration is incorporated in the constitutive material
model in Eq. (9) and is solved numerically and implemented in the
material subroutine (UMAT) of the ABAQUS FE code. Both 2D
(CPE4) and 3D (C3D8) solid continuum elements are used to deter-
mine the deformation in the viscoelastic sandwich beam under
three-point bending. The results obtained using 3D elements are
in good agreement with those obtained using 2D elements, as illus-
trated in Fig. 7. In order to reduce the computational time required
the further analysis is performed using 2D elements. The analysis is
performed up to 1200 h. Parametric studies are performed to
investigate the effects of viscoelastic foam core and non-linear
Fig. 8. Comparison of the transverse displacement field for case (i), case (ii) and
case (iii).
material parameters on the time-dependent performance of sand-
wich beams.

Fig. 9. Effect of moisture on stress and strain fields.

 N. Joshi, A. Muliana / Composite Structures 92 (2010) 254–264 261

4.1. Effect of moisture degradation on the overall response of the approximately 30% in the top skin at the interface for the case
sandwich beam (iii) is observed as compared to the case (i). Whereas the transverse
strain in the degrading core at the interface (x2 = 26.4 mm) for the
To highlight effects of moisture degradation on the time-depen- case (iii) is as high as 52% when compared to the case (i). Fig. 9e
dent behaviors of sandwich beam, the following cases are shows the effect of moisture on the shear strain (e12) at a distance
considered: of 70 mm from the midspan of the beam. Shear strains in the core
(through the thickness) are almost constant in all the three cases at
 Case (i): The sandwich beam composed of orthotropic elastic a particular time and increases with increase in time. This study
carbon–epoxy laminated skin and linear elastic polyurethane shows that the effect of moisture degradation on the behavior of
foam core whose material modulus degrades with the moisture viscoelastic materials is much more pronounced as compared to
concentration as governed by Eq. (14). the elastic materials.
 Case (ii): The sandwich beam composed of orthotropic elastic Fig. 10 illustrates full-field multi-axial stress contours of the vis-
carbon–epoxy laminated skin and isotropic linear viscoelastic coelastic sandwich beams at 1200 h of deformation. In order to
polyurethane foam core without any effect of moisture capture localized effects near the point of applied concentrated
concentration. load and supports finer meshes are generated at these locations.
 Case (iii): The sandwich beam composed of orthotropic elastic Fig. 10b shows that the FE model can capture the localized normal
carbon–epoxy laminated skin and isotropic linear viscoelastic stresses at these critical locations. In the above discussion, we re-
polyurethane foam core with the instantaneous material modu- port stress, strain, and displacement away from the critical loca-
lus degrading with the moisture concentration as governed by tions since the deformation and stresses determined from the
Eq. (14). analytical solution, i.e. strength of material approach, are valid only
at the distance away from the singular point. At these critical loca-
The effect of degradation in the instantaneous modulus of the tions, the effects of moisture degradation on the field variables are
foam core on the displacement in the transverse direction (U2, more pronounced which is expected. It is also seen that stress-dis-
direction of loading) at the midspan of the beam is as shown in continuities are observed at the skin-core interphases.
Fig. 8. At a distance of half the thickness of the sandwich beam
(x2 = 13.5 mm) an increase of about 23% in the transverse displace-
4.2. Moisture degradation in the instantaneous and transient moduli
ment is observed in the case (ii) as compared to case (i) at time
t = 1200 h. The effect of material viscoelasticity without moisture
For the viscoelastic materials the time-dependent modulus can
degradation on the long-term deformation of sandwich beam is
also degrade with moisture concentration. Assuming the time-
more significant than having elastic properties of the foam core de-
dependent modulus to degrade linearly with moisture concentra-
grade with moisture concentration. It is noted that severe elastic
tion which can be accounted through the non-dimensionalized
degradation is assumed by taking modulus of the foam core at
parameter g1 in Eq. (9) as:
the saturated condition to be 50% of the initial elastic modulus.
The viscoelastic properties for the foam core are taken from Huang c
g1 ¼ 1 þ b  ð16Þ
and Gibson [18]. There is an increase of about 46% in the transverse C0
displacement in the case (iii) as compared to the case (i). As ex-
Here C0 is the applied moisture concentration of 1 gm/mm3 and
pected, the effect of moisture degradation is found to be more pro-
b is a constant equal to 0.62. The overall response of the sandwich
nounced in case of the viscoelastic materials.
structure under the three-point bending test is once again
The effect of moisture degradation on the longitudinal (bend-
ing) stress (r11) at a distance of 70 mm from the midspan of the
beam is shown in Fig. 9a. At the midspan (point of application of
load) highly localized stress concentrations are observed. Thus, in
order to avoid the effect of localized stress concentrations the
stresses are measured at a distance of 70 mm from the midspan
of the beam. In all the three cases, the maximum bending stresses
(at the top and bottom faces) at time t = 1200 h are higher than the
stresses at initial time t = 0 h as expected. The magnitude of the
bending stresses during the moisture diffusion process is maxi-
mum in case (iii) as compared to the case (ii) and case (i). A jump
in the bending stresses at the top and the bottom interfaces in the
sandwich beam are observed. The magnitude of the jump in case
(iii) is the highest since the mismatch between the moduli in the
skins and core increase with time, which is due to severe relaxation
in the foam core. This jump can act as a critical parameter in deter-
mining the delamination at the interface. Different stress–strain
fields are observed in the tension and compression regions. It is
worth observing that even though the increase in the jump at
the bottom interface after 1200 h is higher but the magnitude of
the jump is higher at the top interface.
Fig. 9b shows the effect of moisture on the shear stress (r12) at a
distance of 70 mm from the midspan of the beam. The effect of
degradation on the bending (longitudinal) strain (e11) and trans-
verse strain (e22) at the midspan of the beam is shown in Fig. 9c
and d. A jump in the transverse strain is observed at both the top
and the bottom interfaces of the sandwich beam. The jump in
strain at the interface also increases with time. An increase of Fig. 10. Multi-axial stress fields in the sandwich beams at 1200 h.
262 N. Joshi, A. Muliana / Composite Structures 92 (2010) 254–264

analyzed. Eq. (14) for a = 12 MPa governs the degradation of the pared to case (i). At any time greater than 233 h, which is the stea-
instantaneous part (E0), while the transient part degrades as per dy state time, a nearly constant increase of around 8.5% is observed
Eq. (16). The moisture diffusivity ratio of 0.1 (skin to core) is con- in case (ii) when compared to case (i). The instantaneous and the
sidered. The following cases are studied to understand the impact transient modulus for any time greater t = 233 h do not change
of degradation in the transient compliance due to diffusion of with moisture concentration as the steady state is already attained.
moisture: However, the deformation in the two cases is still progressing with
time with the same rates. If the characteristic of creep time in Eq.
 Case (i): Only the instantaneous modulus of the viscoelastic (8) is made to be moisture dependent, the deformation in the two
polyurethane foam core degrades with the moisture cases would be progressing with different rates even after the stea-
concentration. dy state condition is reached.
 Case (ii): Both the instantaneous and transient moduli of the vis- The variation in the bending stress and shear stress with time
coelastic polyurethane foam core degrade with the moisture are negligible. A jump in the bending stress (r11) is observed at
concentration. both the top and the bottom interfaces in both the cases (i) and
(ii). It is worth observing that the magnitude of jump at the top
At the early times, the degradation of the transient modulus interface decreases with the increase in time for both the cases.
with the moisture concentration is less significant. As a result, Whereas the magnitude of jump in the bottom interface for both
not much difference is observed in the field variables for the case cases increases up to time t = 116 h and then decreases for the later
(ii) when compared to case (i). But with the increase in the time times. It is also observed that the magnitude of the maximum
the moisture diffusion through the sandwich beam progresses stresses (in the facings) increase with the increase in time in both
and the creep compliance for the case (i) is significantly different the cases.
than the case (ii). Once the steady state for the moisture diffusion
process is reached, for all the times ahead nearly a constant differ- 5. FE analysis of delamination at the skin-core interface of the
ence in the field variables is observed in the two cases. sandwich composite
As shown in Fig. 11 at time t = 3.64 h an increase of around 2.5%
in the transverse displacement for the case (ii) as compared to case Adhesive polymers are widely used for joining structural com-
(i) is observed. With the increase in time this difference increases. ponents made of similar and dissimilar materials. Since the sand-
At time t = 116 h the increase is about 9.5% in the case (ii) as com- wich composites may be exposed to severe temperature and

Fig. 11. Transverse displacement field at the midspan of the sandwich beam.
 N. Joshi, A. Muliana / Composite Structures 92 (2010) 254–264 263

moisture conditions depending upon the application, it becomes 1 gm/mm3 is reached at around 1150 h. Mode I failure analysis is
essential to study the delamination behavior at the skin-core inter- then performed on the DCB specimen using the 2D plain strain
face under such hostile environmental conditions. Debonding be- (CPE4) elements as shown in Fig. 12. Cohesive elements are applied
tween core and face sheets in sandwich composites have been along the predetermined failure paths, which are located at (a) the
analyzed by several researchers [5,28]. Efforts have also been made interface between the top skin and the core, i.e. between the skin
to analyze the skin-core interfacial fracture properties [29,33] of and adhesive and between core and adhesive, and (b) within the
the dry and wet sandwich composite specimen. In majority of adhesive itself. The crack opening displacement (COD) failure crite-
the work done till date, the delamination at the skin-core interface rion, available in the ABAQUS FE code [1], is first used to evaluate
is analyzed under static loading conditions for a short span of time. crack initiation and propagation:
This study analyzes the transient effect of moisture diffusion on
the delamination at the interface under the creep loading for about
d
f ¼ ð18Þ
1200 h (greater than the steady state time of 1150 h for the mois- dc
ture diffusion process). A sandwich beam composed of orthotropic Here d is the current (measured) normal displacement across the
S2-glass epoxy laminate and linear viscoelastic polyurethane foam cohesive surface and dc is the critical opening displacement. The
core under the combined moisture and mechanical stimuli is stud- failure criterion in Eq. (18) is monitored at the distance of 3 mm
ied. Table 7 gives the material properties of the skin. The instanta- ahead the current crack tip. This distance is determined by varying
neous part (E0) of the polyurethane foam core is assumed to its values until the overall crack propagation responses do not
degrade linearly with the moisture concentration as per Eq. (14) change with this value. When the failure criterion reaches a value
with a = 12 MPa. The facings and the core material are glued to- of one, crack is initiated and the fracture energy is released to create
gether by the FM73 M adhesive. The calibrated time-dependent two traction free surfaces.
parameters (Prony series) representing the creep compliance and The analysis is performed for two different constant load levels
the elastic properties for the FM73 M [21] are given in Table 8. of 110 and 120 N, which correspond to 46% and 50% of the failure
The instantaneous part (E0) of the FM73 M adhesive is assumed load of the dry specimen, respectively, with critical displacement
to degrade linearly with moisture concentration as per Eq. (17), as 0.34 mm [28]. Figs. 13 and 14 show the graph of cumulative
where Eref is the instantaneous modulus at a reference condition. crack length versus time along the predetermined failures paths
between the skin-core interface and within the adhesive layers,
1 c
E0 ¼ ¼ Eref  1335  ð17Þ respectively. From the FE simulation it is observed that for the
D0 C0
same sandwich beam in a dry condition under the applied load
The analysis is performed in the FE framework. The purpose of 120 N the crack does not propagate until time t = 1200 h indicating
the numerical study is to simulate the crack initiation time and the the detrimental effect of moisture concentration on crack propaga-
cumulative crack length under the effect of moisture degradation. tion. Significantly different crack behaviors are observed when fail-
The analysis is performed using the Double Cantilever Beam (DCB) ures are assumed between the skin-core interfaces or within the
test on the sandwich beam (254 mm  32.6 mm  38.1 mm)
Majumdar et al. [22] with an initial crack length of 50 mm. To
simulate the moisture diffusion process using FE code, (DC2D4)
2D diffusivity elements are used. With the moisture diffusivity of
the skin and the core assumed to be to 1.06E3 mm2/s and
1.06E2 mm2/s, respectively, the steady state of the moisture
diffusion process for the applied moisture concentration of (C0)

Table 7
Elastic properties of S2-glass epoxy laminate.

Material Properties Values

Skin: S2-glass epoxy Young’s modulus, E11 (MPa) 76419
Young’s modulus, E22 (MPa) 20184
Young’s modulus, E33 (MPa) 20184
Shear modulus, G12 (MPa) 7395.8
Shear modulus, G13 (MPa) 7395.8
Fig. 12. A schematic of the double cantilever beam.
Shear modulus, G23 (MPa) 4031.9
Poisson’s ratio, m12 0.265
Poisson’s ratio, m13 0.265
Poisson’s ratio, m23 0.335

Table 8
Elastic properties and Prony series coefficients for FM73 M.

n kn (s1) Dn  106 (MPa1)

1 1 3.70
2 101 0.296
3 102 40.7
4 103 33.3
5 104 55.6
6 105 259
7 106 1000
8 107 3810
9 108 7040
E0 = 2710 MPa m = 0.35 Fig. 13. Cumulative crack length versus time for a DCB in mode I under creep
loading with the crack at the skin-core interface governed by the COD criterion.
264 N. Joshi, A. Muliana / Composite Structures 92 (2010) 254–264

[3] Antoon MK, Koening JL. Irreversible effects of moisture on the epoxy matrix in
glass-reinforced composites. J Polym Sci 1981;19(2):197–212.
[4] Ashby MF, Gibson LJ. Cellular solids structure and properties. Cambridge:
Cambridge University Press; 1997.
[5] Berggreen C, Simonsen BC, Borum KK. Experimental and numerical study of
interface crack propagation in foam-cored sandwich beams. J Compos Mater
2007;41(4):493–520.
[6] Botelho EC, Costa ML, Pardini LC, Rezende MC. Processing and hygrothermal
effects on viscoelastic behavior of glass fiber/epoxy composites. J Mater Sci
2005;40(14):3615–23.
[7] Botelho EC, Costa ML, Rezende MC. Hygrothermal effects on damping behavior of
metal/glass fiber/epoxy hybrid composites. Mater Sci Eng 2005;399(1-2):190–8.
[8] Bouadi H, Sun CT. Hygrothermal effects on the stress field of laminated
composites. J Reinf Plast Compos 1989;8:40–55.
[9] Carter HG, Kibler KG. Langmuir-type model for anomalous moisture diffusion
in composite resins. J. Compos Mater 1978;12(4):118–31.
[10] Carslaw HS, Jaeger JC. Conduction of heat in solids. Oxford University Press
1959.
[11] Crank J. The mathematics of diffusion. 2nd ed. Oxford University Press; 1975.
Fig. 14. Cumulative crack length versus time for a DCB in mode I under creep
[12] Diab Inc., Diffusivity of Divinycell H-250; 2002 <http://fiberglasswarehouse.
loading with the crack within the glue governed by the COD criterion.
com/tds/H_divinycell.pdf>.
[13] Doxsee LE, Lee W, Springer GS, Chang SS. Temperature and moisture induced
deformations in composite sandwich panels. J Reinf Plast Compos
adhesive layers. In practice, both failure paths are possible and it is 1985;4(4):326–53.
[14] Frostig Y, Thomsen OT. Non-linear thermal response of sandwich panels with a
often a combination of the two failure paths.
flexible core and temperature dependent mechanical properties. Compos Part
B: Eng 2007;39:165–84.
6. Conclusions [15] Gibson RF, Yau A, Mende EW, Osborn WE, Riegner DA. The influence of
environmental conditions on the vibration characteristics of chopped-fiber-
reinforced composite materials. J Reinf Plast Compos 1982;1(3):225–41.
The FE analysis is performed to evaluate the effect of transient [16] Granville DM. Moisture effects of polymethacrylimide foam and honeycomb
moisture diffusion on the deformations of the viscoelastic sand- core in sandwich/skin structures. Annual Forum Proceedings – American
Helicopter Society 1987;2(1):907–14.
wich composite consisting of polymeric foam core and fiber-rein- [17] Haj-Ali RM, Muliana AH. Numerical finite element formulation of the schapery
forced laminated skins. It is assumed that the elastic and time- non-linear viscoelastic material model. Int J Numer Meth Eng 2004;59:25–45.
dependent moduli at any particular location in the polymer foam [18] Huang JS, Gibson LJ. Creep of sandwich beams with polymer foam cores. J
Mater Civil Eng 1990;2(3):171–82.
core depend on the moisture concentration at that location at
[19] Ishiaku US, Hamada H, Mizoguchi M. The effect of ambient moisture and
any instant of time. Since the responses of the polymer foam core temperature conditions on the mechanical properties of glass fiber/carbon
depend on the history of stress–strain, it is essential to concur- fiber/nylon 6 sandwich hybrid composites consisting of skin-core
morphologies. Polym Compos 2005;26:52–9.
rently solve transient moisture diffusion and deformation in the
[20] Ionita A, Weitsman YJ. A model for fluid ingress in closed cell polymeric foams.
viscoelastic sandwich beams. Parametric studies have been per- Mech Mater 2007;39:434–44.
formed to investigate the effects of viscoelastic foam core and [21] Jurf RA, Winson JR. Effect of moisture on the static and viscoelastic shear
non-linear material parameters on the overall time-dependent properties of epoxy adhesives. J Mater Sci 1985;20:2979–89.
[22] Majumdar P, Srinivasagupta D, Mahfuz H, Joseph B, Thomas MM, Christensen
performance of sandwich beams. It is shown that the viscoelastic S. Effect of processing conditions and material properties on the debond
foam core without moisture degradation significantly increases fracture toughness of foam-core sandwich composites: experimental
the overall long-term deformation of sandwich beam compared optimization. Compos Part A: Appl Sci Manufact 2003;34:1097–104.
[23] Mikhailov MD, Ozisik MN. Unified analysis and solutions of heat and mass
to when the elastic degradation in the foam core is considered. diffusion. Dover Publication; 1984.
The increase in the overall deformation enhances stress-disconti- [24] Morganti F, Marchett M, Reibedli G. Effects of moisture and thermal ageing on
nuities at the skin-core interphases, which potentially lead to structural stability of sandwich panels. Acta Astronaut 1983;11(7–
8):489–508.
delamination. With the instantaneous (elastic) and the transient [25] Muliana AH, Khan KA. A time-integration algorithm for thermo-rheologically
moduli of the foam core degrading with moisture concentration, complex polymers. Comput Mater Sci 2008;41(4):576–88.
the deformation of the sandwich beam is even more aggravated. [26] Patel M, Morrell P, Cunningham J, Khan N, Maxwell RS, Chinn SC. Complexities
associated with moisture in foamed polysiloxane composites. Polym Degrad
It is found that at an early stage of diffusion process; the effect of
Stabil 2008;93(2):513–9.
degradation in the transient modulus on the overall deformation [27] Pradeep V, Ganesan N, Bhaskar K. Vibration and thermal buckling of composite
of the sandwich beam is less significant compared to its effect at sandwich beams with viscoelastic core. Compos Struct 2007;81:60–9.
[28] Saha MC, Kabir E, Jeelain S. Study of debond fracture toughness of sandwich
later times. Once the steady state for the moisture diffusion pro-
composites with nanophase core. Mater Lett 2008;62:567–70.
cess is reached, the deformations in the sandwich beams with elas- [29] Scudamore RJ, Cantwell WJ. The effect of moisture and loading rate on the
tic degradation and viscoelastic degradation foam cores progress interfacial fracture properties of sandwich structures. Polym Compos
with the same rates. This is due to constant characteristics of creep 2002;23(3):406–17.
[30] Shen CH, Springer GS. Effect of moisture and temperature on the tensile
time given in Table 4. If the characteristic of creep time in Eq. (8) is strength of composite materials. J Compos Mater 1977;11(1):2–16.
made to be moisture dependent, the deformation in the two sand- [31] Shen CH, Springer GS. Environmental effects in the elastic moduli of composite
wich beams would be progressing with different rates even after materials. J Compos Mater 1977;11(3):250–64.
[32] Springer GS, Sanders BA, Tung RW. Environmental effects on glass fiber
the steady state condition is reached. reinforced polyester and vinylester composites. Environ Effects Compos Mater
1981;1(1):126–44.
Acknowledgement [33] Veazie DR, Robinson KR, Shivkumar K. Effects of the marine environment on
the interfacial fracture toughness of PVC core sandwich composites. Compos
Part B: Eng 2004;35:461–6.
This research is sponsored by the National Science Foundation [34] Weitsman Y. Stress assisted diffusion in elastic and viscoelastic materials. J
under Grant No. 0546528. Mech Phys Solid 1987;35(1):73–93.
[35] Weitsman YJ. Anomalous fluid sorption in polymeric composites and its
relation to fluid-induced damage. Compos Part A: Appl Sci Manufact
References 2006;37:617–23.
[36] Weitsman YJ, Guo Ya-Jun. A correlation between fluid-induced damage and
[1] ABAQUS, user’s manual, Version 6.5, Hibbitt, Karlsson and Sorensen Inc.; 2005. anomalous fluid sorption in polymeric composites. Compos Sci Technol
[2] Allred RE. The effect of temperature and moisture content on the flexural 2002;62:889–908.
response of kevlar/epoxy laminates. Part I. [0/90] Filament orientation. J [37] Woldesenbet E, Gupta N, Vinson JR. Determination of the moisture effects on
Compos Mater 1981;15(2):100–16. the impact properties of composite materials. J Mater Sci 2002;37(13):2693–8.