polymers

Article
Moisture Absorption in Polymer Composites
Reinforced with Vegetable Fiber:
A Three-Dimensional Investigation via
Langmuir Model
Mirenia Kalina Teixeira de Brito 1 , Wanessa Raphaella Gomes dos Santos 1, * ,
Balbina Raquel de Brito Correia 1 , Robson Araújo de Queiroz 1 ,
Francisca Valdeiza de Souza Tavares 2 , Guilherme Luiz de Oliveira Neto 3
and Antonio Gilson Barbosa de Lima 1
1 Postgraduate Program in Process Engineering, Federal University of Campina Grande, Campina Grande,
Paraiba 58429-900, Brazil; mireniabrito@icloud.com (M.K.T.d.B.); balbinacorreia@hotmail.com (B.R.d.B.C.);
robson_frr@hotmail.com (R.A.d.Q.); antonio.gilson@ufcg.edu.br (A.G.B.d.L.)
2 Department of Production Engineering, Faculdade Paraíso – CE, Juazeiro do Norte 63010-220, Brazil;
valdeiza.tavares@hotmail.com
3 Mathematics Department, Federal Institute of Education, Science and Technology of Piauí (IFPI), Floriano,
Piauí 64808-475, Brazil; guilherme@ifpi.edu.br
* Correspondence: wanessa.raphaella@yahoo.com.br; Tel.: +55-83-99699-7210




Received: 3 October 2019; Accepted: 7 November 2019; Published: 9 November 2019


Abstract: This work aims to study numerically the moisture absorption in polymer composite
reinforced with vegetable fibers using the Langmuir model which considers the existence of free
and entrapped water molecules inside the material. A three-dimensional and transient modeling for
describing the water absorption process inside the composite and its numerical solution via finite
volume method were presented and discussed. Application has been made for polymer composites
reinforced with sisal fiber. Emphasis was given to the effect of the layer thickness of fluid close to the
wall of the composite in the progress of water migration. Results of the free and entrapped solute
(water) concentration, local moisture content and average moisture content, at different times of
process, and inside the composite were presented and analyzed. It was verified that concentration
gradients of the molecules (free and entrapped) are higher in the material surface, at any time of
the process, and concentration of free solute is greater than the concentration of entrapped solute.
It was verified that the water layer thickness surrounding the composite strongly affects the moisture
absorption rate.

Keywords: polymer composites; vegetal fibers; Langmuir model; numerical solution

1. Introduction
Composite is a material consisting of two or more insoluble materials, which are combined to
create a useful engineering material having some properties not obtained by the constituents taken
separately [1].
The main purpose to manufacture composite materials is to obtain a new material with increased
properties, mainly from the mechanical point of view (mechanical resistance) and low weight [2].
Among the different types of composite materials, we can cite those which are reinforced with
particles, reinforced with fibers (high length/diameter ratio) and those structural (combination of
composites and homogeneous materials whose properties depend on the geometric designs of the
structural elements).

Polymers 2019, 11, 1847; doi:10.3390/polym11111847 www.mdpi.com/journal/polymers

Polymers 2019, 11, 1847 2 of 30

Fibers are efficient reinforcements whose main function is to support the loads transmitted by
the matrix. The functions of the matrix are to join and guide the fibers, to transmit and distribute the
mechanical stress, and to protect them from mechanical damage, such as abrasion and environmental
degradation [3].
Concerning the types of fibrous reinforcements, which may be used in the composites, the vegetable
fiber stands out.
The vegetable fibers show some advantages, such as they are renewable and biodegradable
materials, having good mechanical properties and lower cost compared with synthetic fibers, high
availability, and great diversity in nature [4–6].
The main disadvantages of vegetable fibers as reinforcement in polymer composites are related to
their susceptibility to the influence of environmental agents, with regard to high moisture absorption,
nonuniformity, and decrease in the mechanical properties [7,8]. The high moisture absorption provokes
an increase in the weight of the material, decline in the mechanical properties, and swelling of the fibers
causing changes in its shape, which impairs its application in projects that require greater dimensional
accuracy [9]. Thereby, the knowledge of hygrothermal behavior and the transport phenomena becomes
essential for the best uses of these composite types as engineering materials.
The moisture absorption causes several problems in the composite materials, such as changes in
the thermophysical, mechanical, and chemical characteristics of the polymer matrix by plastification
and hydrolysis. In these materials, the moisture diffusion process is very complex once it depends on
the mass diffusivity of the individual constituents, volume fraction, and the constitutive arrangement
and morphology of the reinforcement [10]. Furthermore, hygrothermal exposure can reduce glass
transition temperature of the polymers and provoke plastification of the polymer matrix that results in
decreasing of the dominant resistance properties of the matrix. The hygrothermal effect increase the
internal voids, promotes polymer chain expansion, and creates microcracks in the matrix [11].
It is worth noting that the effects caused by a long period of moisture exposure can be irreversible
due to water affinity with specific functional groups of the reinforcements. The destructive changes
occur, generally, due to degradation of probable physical–chemical interactions between resin and fiber,
and, consequently, there is a displacement in the fiber [12,13], causing delamination of the composite
and reduction in the properties of the material. Thus, the prediction of the moisture content along
the time is particularly relevant, because, from the results obtained, it is possible to predict which
areas are most susceptible to cracks and deformations capable of decreasing the quality of the product,
especially in warm and humid environments [14,15].
The migration of water inside the polymer composites occurs by diffusion. Hence, to predict the
phenomenon of moisture migration an adequate mathematical model is necessary, in conjunction with
its analytical and/or numerical solution. Therefore, many diffusion models have been proposed to
predict moisture transport in polymer composites. Fick’s model is one of the most used models due to
its simplicity.
Fick’s laws of diffusion are the simplest and were inspired in the analogy between diffusion and
heat or electricity conduction, as proposed by Fourier and Ohm [16]. The preference by Fick’s model is
generally related with the facility for fitting the model equation to the experimental data of sorption [17].
However, for longer times of process, this model does not satisfy the moisture absorption phenomenon
with precision [17], which causes the need to develop new models to predict this process adequately.
Some works attribute anomalies related to the classical Fick’s law of diffusion to chemical reactions
that occur during some absorption processes [18–21]. Experimental studies reveal, for example, the
existence of coupling between humid and oxidation aging, characterized by the presence of several
different physical mechanisms [20]. Other studies show increased water absorption in the presence of
molar hydroxyl groups. In these schemes, the water absorption process is governed by two phases, one
physical and one chemical, although there is no hierarchical relationship between both phases [22,23].
In this sense, although Fick’s model is simpler, more complete models must be used to describe
water diffusion behavior, such as the Jacob’s–Jones model [24], model of time-dependent diffusion [25],
Polymers 2019, 11, 1847 3 of 30

and Langmuir model [26]. The Langmuir model has the main advantage of considering the interaction
between the entrapped and free water molecules inside the composite.
In Langmuir model, one quantity of the water molecules is free to move, while the complementary
quantity is connected to the polymer chains due to reversible chemical reactions, such as that occurring
when moisture-induced plastification occurs. A free water molecule can become an entrapped molecule
and, mutually, an entrapped water molecule can become free. These effects are computed in the model
by using two additional parameters: λ (probability of a free water molecule being entrapped) and µ
(probability of an entrapped molecule becoming free) [27,28]. So, this model includes water–polymer
molecular interactions by assuming that water molecules can exist in two states (free and linked) and
that there is a dynamic exchange process between the two stages. Therefore, Langmuir model is an
improved version of the Fick’s diffusion model, thence it is known as anomalous model or non-Fickian.
Currently, there are, fundamentally, three kind of analysis to obtain results related to the dynamics
of moisture absorption in porous solids: analytical, numerical, and experimental. The disadvantage of
performing experiments of moisture absorption in the laboratory consists of the fact that the results are
dependents of the predefined operational conditions. In general, the water absorption process is very
slow, thus, the material requires a long time to reach the hygroscopic saturation condition. Analytical
methods are restricted to problems whose hypotheses are simplified and applied a simple geometry,
which limits the use. Numerical methods, on the other hand, are practically unrestricted, can be used
in more complex problems, with boundary conditions defined in arbitrary geometries, and show faster
and more economical results than other methods [25,29–31].
Theoretical studies of water absorption in polymeric composites reinforced with fibers using
Langmuir model have been reported in the literature [16,27,28,32–36]. Hereafter, some studies
are discussed.
Joliff et al. [16] conducted a comparative study, in a one-dimensional approach, between
experimental, analytical, and numerical data of the water absorption in composites reinforced with
unidirectional glass-fiber, in order to show the influence of the distribution of fibers and interfacial
properties in the diffusion kinetics of water using Fickian and non-Fickian models. The authors
concluded that the progress of water absorption is best described by Langmuir model than by Fick’s
model, however, in short times, both models correctly predict the experimental behavior of the water
inside the composite.
Peret et al. [27] used finite element method in a one-dimensional approach to solve problems related
to Fickian and non-Fickian water diffusion mechanics in carbon fiber and glass-reinforced polymer
composites. The simulations were performed at a macroscopic level to capture the heterogeneity
of composite materials and study comparatively the Fick’s and Langmuir’s models. The authors
concluded that there are significant discrepancies between the two models in moisture diffusion process
in the transient state.
Melo et al. [32] have reported important information about water absorption process in polymer
composites. Emphasis was addressed to Langmuir–type model application in this process. In
this work, the authors have cited analytical and numerical solutions of the Langmuir-type model,
in one-dimensional (1D) approach, and applied the study for polymer composite reinforced with Caroá
vegetable fiber.
Santos et al. [36] studied, theoretically, the anomalous behavior of transient and one-dimensional
moisture absorption in polymer composite reinforced with vegetable fibers using the Langmuir model.
In this research, the authors developed analytical and numerical solutions for the study, and they
verified that water absorption is faster in the initial stages, tending to decrease for long periods until
equilibrium. The predicted average moisture content data obtained from the numerical solution
developed showed a good agreement with the experimental data reported in the literature.
Despite the importance, these works are directed towards a one-dimensional approach, and there
are scarce studies that use a three-dimensional approach, especially that involved the effect of the water
layer near the composite wall.
Polymers 2019, 11, 1847 4 of 30
Polymers 2019, 11, 1847 4 of 29

Therefore,the
Therefore, theobjective
objectiveofofthis
thisresearch
researchisistotopredict
predictmass
masstransfer
transferduring
duringthe
theprocess
processofofwater
water
absorption in polymer composites (parallelepipedic shape) reinforced with vegetable
absorption in polymer composites (parallelepipedic shape) reinforced with vegetable fibers, placed fibers, placed
randomlyinto
randomly intothe
thematrix
matrix byby using
using thethe Langmuir
Langmuir model,
model, including
including thethe effect
effect of water
of water layer
layer thickness
thickness at
at the surface of the solid.
the surface of the solid.

2.2.Methodology
Methodology

2.1.
2.1.The
ThePhysical
PhysicalProblem
Problemand
andGeometry
Geometry
The
Thephysical
physicalproblem
problemto tobebestudied
studiedisisthe
the absorption
absorption of of water into a porous solid material,
material, with
witha
ageometry
geometry ofof a parallelepiped with
with dimensions
dimensions2R 2Rxx×× 2R
2Ryy ×× 2R
2Rzz, immersed in in aafluid
fluidmedium,
medium,distant
distant
from
fromthe
thecontainer
containerwall
wallby
bylxlx(in
(inx-direction),
x-direction),1y1y(in
(iny-direction),
y-direction),and andlzlz(in
(inz-direction),
z-direction),asasshown
showninin
Figure
Figure1.1.

(a)

(b)
Figure1.1.Parallelepipedic
Figure Parallelepipedicsolid
solid(a)
(a)immersed
immersedininaafluid
fluidmedium
mediumand
and(b)
(b)with
withits
itsgeometric
geometricparameters.
parameters.

2.2. Mathematical Model

2.2. Mathematical Model
For the analysis and solution of the physical problem studied in this paper, the following
For the analysis and solution of the physical problem studied in this paper, the following
considerations were made: a) the solid is homogeneous and isotropic, b) it has a parallelepipedic shape,
considerations were made: a) the solid is homogeneous and isotropic, b) it has a parallelepipedic
c) the mass diffusion coefficient remains constant throughout the process, d) the process is transient,
shape, c) the mass diffusion coefficient remains constant throughout the process, d) the process is
e) variations in material dimensions during the water absorption process were neglected, f) mechanism
transient, e) variations in material dimensions during the water absorption process were neglected ,
of water transport inside the material is purely diffusive, g) there is no mass generation, and h) the solid
f) mechanism of water transport inside the material is purely diffusive, g) there is no mass generation,
isand
completely dry atisthecompletely
h) the solid beginning of drytheatprocess. It is noteworthy
the beginning that the development
of the process. It is noteworthy of thethat
study
the
considered the composite homogeneous, isotropic, and without direct influence
development of the study considered the composite homogeneous, isotropic, and without direct of fiber orientation
within the composite.
influence This waswithin
of fiber orientation due to the
the composite.
manufacturing Thisprocess
was due of to
thethe
sisal fiber-reinforced
manufacturing polymeric
process of the
composite
sisal fiber-reinforced polymeric composite developed by Santos [37], used in the experimentalhand
developed by Santos [37], used in the experimental phase to validate this research: phase
lay-up (randomly
to validate this distributed fibers).
research: hand Thus, (randomly
lay-up to contemplate the properties
distributed of the
fibers). fibers
Thus, to in the numerical
contemplate the
solution, their characteristic was coupled to the mass diffusion coefficient.
properties of the fibers in the numerical solution, their characteristic was coupled to the mass
In thecoefficient.
diffusion Langmuir model, the non-Fickian moisture absorption behavior can be explained
quantitatively by assuming
In the Langmuir that the
model, moisture absorption
non-Fickian occursabsorption
moisture in the presence of two
behavior cansimultaneous
be explained
stages, one being the free water stage and the another the entrapped
quantitatively by assuming that moisture absorption occurs in the presence of twowater stage [26], described by the
simultaneous
following
stages, onemass transfer
being Equations:
the free water stage and the another the entrapped water stage [26], described by
the following mass transfer Equations:∂C   ∂S
= ∇· DC ∇C − (1)
∂t ∂C ∂t∂S
= ∇ · (D ∇C) − (1)
∂t ∂t
∂S
∂S
= λC − µS (2)
∂t = λC − µS (2)
∂t
In Equations (1) and (2), C represents the concentration of the free solute to diffuse into the
material, S represents the concentration of the entrapped solute, Dc is the mass diffusion coefficient
Polymers 2019, 11, 1847 5 of 30

In Equations (1) and (2), C represents the concentration of the free solute to diffuse into the
material, S represents the concentration of the entrapped solute, Dc is the mass diffusion coefficient
(free molecules), t is the time, λ is the probability of a free solute molecule to be entrapped inside the
solid, and µ is the probability that an entrapped solute molecule becomes free.

∂C ∂ C ∂C ∂ C ∂C ∂ C ∂C ∂S
! ! !
= D + D + D − (3)
∂t ∂x ∂x ∂y ∂y ∂z ∂z ∂t

∂S
= λC − µS (4)
∂t
For the solution of Equations (3) and (4), the following initial and boundary conditions were used:
Initial conditions: it was considered that the solid is completely dry at the beginning of the process.
Thus, one can write:
−Rx < x < Rx



y < y < Ry

C = S = 0;  −R (5)


 −R < z < R

z z

Boundary conditions: on the surface of the solid, the variation in the concentration of the solute in
the fluid medium is considered to be equal to the diffusive flux of solute at the surface of the material.
Thus, one can write:
∂C ∂C
(
x = ±Rx ; l = lx
l = ±D ; (6)
∂t ∂x t>0

∂C ∂C
(
y = ±Ry ; l = ly
l = ±D ; (7)
∂t ∂y t>0

∂C ∂C
(
z = ±Rz ; l = lz
l = ±D ; (8)
∂t ∂z t>0
Once determined C and S at any point inside the material, it is possible to calculate the total
moisture content present in the material at any position and instant of time. Thus, the total moisture
content is given by the sum of the concentrations of C and S, as follows:

M = C+S (9)

Furthermore, the average moisture content of the solid at any time of the process is given by:
Z
1
M̄ = MdV (10)
V V

in which V is the total volume of the solid and dV = dxdydz is the volume of an infinitesimal sample
of the porous solid.

2.3. Numerical Solution

Basically, the numerical solution of a partial differential equation consists of two steps:
a) discretizing the physical domain under study in several sub domains and b) transforming the
governing equation into a linear algebraic equation in the discretized form applied to each sub domain
contained in the solid under study. After these procedures, the result is a set of linear algebraic
equations whose solution provides the distribution of the variable of interest within the domain and
in time.
For the three-dimensional numerical solution of the governing equations applied to the absorption
of moisture in materials with parallelepipedic form, the finite volume method was used. In this
solution, a fully implicit formulation for the concentration of free solute and explicit formulation for
the concentration of entrapped solute were used. For this, due to symmetrical equivalence, only 1/8 of
Polymers 2019, 11, 1847 6 of 30

the solid was used. From this symmetric equivalence and the parallelepiped geometry of the studied
composite, the numerical mesh was developed to 1/8 of the solid, as shown in Figure 2, with the
refinement and choice of mesh to be determined later. Thus, for each control volume, a solution was
developed
Polymers 2019,for
Polymers 2019, the
1847properties of the mass phenomena studied from the Langmuir model.
11, 1847
11, 66 of
of 29
29

(a)
(a) (b)
(b)
Figure 2.
Figure 2. Parallelepipedic
Parallelepipedicsolid.
Parallelepipedic solid.(a)
solid. (a)Geometry
(a) Geometryofof
Geometry ofthe
the
the studied
studied
studied solid
solid
solid and
and
and (b)
(b)(b) mesh
mesh
mesh characterization
characterization
characterization of
of
of 1/8
of
1/8studied
1/8 of solid.
of studied
studied solid.
solid.

Figure
Figure 333 illustrates
Figure illustrates aaa control
illustrates control volume
control volume (sub
volume (sub domain)
(sub domain) used
domain) used for
used for the
for the discretization
the discretization of
discretization of governing
of governing
governing
equations. Further,
equations. Further,
equations. Figure
Further, Figure 3 shows
Figure 33 shows
shows thethe nodal
the nodal point
nodal point P (in
point PP (in the
(in the center
the center of
center of the
of the control
the control volume),
control volume), its
volume), its adjacent
its adjacent
adjacent
neighbors W,
W, E,
neighbors W,
neighbors E, S,
E, N,
S, N,
S, T,
T, and
N, T, and F,
and F,
F, the
the distances
the distances between
distances these
between these
between nodal
these nodal points,
nodal points, and
points, and the dimensions
and the
the dimensions Δx,
dimensions ∆x,Δy,
Δx, ∆y,
Δy,
and ∆z,
and Δz,
and of
Δz, of the
the control
of the control volume.
control volume.
volume.

Figure 3.
Figure
Figure 3. Control
3. Control volume
Control volume used
volume used for
used for the
for the numerical
the numerical solution.
numerical solution.
solution.

Following is
Following is the
the detailing
detailing of
of the
the whole
whole numerical
numerical procedure
procedure used
used for for the
the solution
solution ofof the
the
governing equations.
governing equations.
Solution for
Solution for the
the water
water concentration.
concentration.
The numerical
The numericalsolution
numerical solutionofof
solution of Equation
Equation
Equation (3)
(3) (3) is obtained
is obtained
is obtained by integrating
by integrating
by integrating itit in
it in volume in volume
volume and
and
and time, time,
astime, as
as
follows:
follows:
follows:  
∂C ∂ C ∂C ∂ C ∂C
R R R R R R R R   R R R R
∂C
∂C dtdzdydx = x y z ∂∂ D ∂C
∂C dtdzdydx + x y z ∂t∂∂y
D ∂C
∂C dtdzdydx+
Rx Ry Rz Rt ∂t dtdzdydx
dtdzdydx = = R Rt ∂xRD
D R  ∂xdtdzdydx
dtdzdydx +
+ D
D
∂y
dtdzdydx
dtdzdydx ++ (11)
∂
∂t
∂t C ∂C ∂x
∂x ∂S
∂x
∂x ∂y
∂y ∂y
∂y
x y z t ∂z
D ∂z
dtdzdydx − x y z t ∂t
dtdzdydx
∂∂ ∂C
∂C ∂S
∂S (11)
(11)
+
+ DD dtdzdydx −
dtdzdydx − dtdzdydx
dtdzdydx
∂z
∂z ∂z
∂z ∂t
∂t

Assuming aa fully
Assuming fully implicit
implicit formulation,
formulation, one
one can
can write
write the
the result
result of
of Equation
Equation (11)
(11) in
in the
the discretized
discretized
form as
form as follows:
follows:
A CC =
A =AA CC +
+AA CC +
+AA CC +
+AA CC +
+AA CC +
+AA CC +
+AA CC +
+ BB (12)
(12)
Polymers 2019, 11, 1847 7 of 30

Assuming a fully implicit formulation, one can write the result of Equation (11) in the discretized
form as follows:

AP CP = AE CE + AW CW + AN CN + AS CS + AF CF + AT CT + AoP CoP + BP C (12)

where
DCe
AE = ∆y∆z (13)
δxe
DCw
AW = ∆y∆z (14)
δxw
DCn
AN = ∆x∆z (15)
δyn

DCs
AS = ∆x∆z (16)
δys

DC
f
AF = ∆y∆x (17)
δzf
DC
t
AT = ∆y∆x (18)
δzt
∆x∆y∆z
AP o = (19)
∆t
∆x∆y∆z DC DC DC DC DC DC
AP = + e ∆y∆z + w ∆y∆z + n ∆x∆z + s ∆x∆z + f ∆y∆x + t ∆y∆x (20)
∆t δxe δxw δyn δys δzf δzt
  ∆x∆y∆z
BP C = − SP − SoP (21)
∆t
It should be noted that Equation (12) is only applied to the internal control volumes of the
computational domain (Figure 4). For the other control volumes (symmetry and border), it is proceeded
with a mass balance in each one of them. In total, there are 27 different types of control volume of
dimensions ∆x, ∆y, and ∆z. As an example, the result of Equation (11) applied to the control volume
of the right upper corner of the computational domain, as shown in Figure 5, is given by:

AP CP = AW CW + AS CS + AT CT + AoP CoP + BP C (22)

where
DCw
AW = ∆y∆z (23)
δxw
DCs
AS = ∆x∆z (24)
δys

DC
t
AT = ∆y∆x (25)
δzt
∆x∆y∆z
AP o = (26)
∆t
∆x∆y∆z DC DC DC ∆y∆x
δxw ∆y∆z + δys ∆x∆z + δzt ∆y∆x +
w s t
AP = ∆t + !+
δzf
+ ∆t
DC f lz
(27)
∆y∆z
 ∆z∆x  +  
δyn ∆t
C + ly
δxe
+ ∆t
D n DC e lx
 A = + ΔyΔz + ΔyΔz + ΔxΔz + ΔxΔz + ΔyΔx + ΔyΔx
Δt δx δx δy δy
ΔxΔyΔz δz δz (20)
B = −(S − S ) (21)
Δt
ΔxΔyΔz
B = −(S − S ) (21)
It should be noted that Equation (12) is only applied Δtto the internal control volumes of the
computational
It should be domain (Figure
noted that 4). For(12)
Equation theisother
only control
applied volumes (symmetry
to the internal controland border),
volumes of ittheis
Polymers 2019, 11, 1847 8 of 30
proceeded with a mass balance in each one of them. In total, there are
computational domain (Figure 4). For the other control volumes (symmetry and border), it is27 different types of control
volume of with
proceeded dimensions Δx, Δyin
a mass balance , and
eachΔz.oneAsofanthem.
example, the there
In total, resultare
of 27
Equation (11)
different applied
types to the
of control
o
Δ ∆x∆y
Δ
o
∆y∆z o
∆x∆z  ∆x∆y∆z
 + Δ z.of
volume of dimensions C
x, y , and C Asthe C
ancomputational
control volume of
BPthe
C right upper corner example, the result
− SPof
domain, Equation
as
o shown in (11) applied
Figure 5, isto the
given
e
f n

=  + −S (28)
by:
control volume of the rightδzupper
f ∆tcornerδx ofe the∆t δyn
computational
P
∆t domain, as shown ∆t in Figure 5, is given
+ lz + lx + ly
DC f DC e DC n
by:

Figure 4. Internal control volumes.

Figure 4. Internal control volumes.

Figure 5. Volume with mass flow conditions in the x, y, and z directions.

Bulleted lists look like this:

Solution for the entrapped water concentration.
The numerical solution of Equation (4) is obtained by integrating it in volume and time, as follows:

∂S
Z Z Z Z Z Z Z Z
dtdzdydx = (λC − µS)dtdzdydx (29)
x y z t ∂t x y z t

Assuming an explicit formulation, one can write the Equation (29), in the discretized form,
as follows:
AP SP = AoP SoP + BSP (30)

where
∆x∆y∆z
AoP = (31)
∆t
∆x∆y∆z
AP = + µ∆x∆y∆z (32)
∆t
BSP = λCP ∆x∆y∆z (33)

In the discretized form, the local and average moisture contents can be written, respectively,
as follows:
M = C+S (34)
Polymers 2019, 11, 1847 9 of 30

npx−1 npy−1 npz−1

1 X X X
M = Mi,j,k ∆Vi,j,k (35)
V
i=2 j=2 k=2

in which i, j and k represent the position of the nodal point in the x, y, and z directions, respectively,
and npx, npy, and npz are the nodal point numbers in the x, y, and z directions, respectively.
From the discretization of the governing equations, a system of algebraic equations is generated
that must be solved to obtain the values of C, S, and M within the material throughout the process.
The solution of this system of algebraic equations was done using the Gauss–Seidel iterative method.
In order to obtain the numerical results, a computer code was developed in Mathematica software
(version 9).

2.4. Studied Cases

2.4.1. Mesh and Time Step Refinements

For the refinement study of the mesh and time step, isothermal cases were used, with µ = 3.27972
× 10−6 s−1 , λ = 2.12852 × 10−6 s−1 , and D = 7.31787 × 10−12 m2 /s. A total time of approximately 56 h
was considered for analysis. The values of µ, λ, and D were initially estimated using the methodology
proposed by Joliff et al. [16], when applied to water absorption in composite materials reinforced by
sisal fiber [30,31].
Briefly, this method used as a starting point the estimates of the parameters of interest and the
−
experimental values of the average moisture content (M). Thus, in this research, it was used the set of
values reported by Santos [37] for sisal fiber reinforced polymer composites at T = 25 ◦ C and with the
−
thickness of 3 mm (2 Ry ). From the value of M, a linear regression of the characteristic equation was
made to find the values of λ and µ of the Langmuir model, and in possession of these values, the value
of the mass diffusion coefficient was reached. Thus, the values of D, λ, and µ were obtained and used
for the study of mesh refining and time.
Table 1 summarizes all the physical parameters of the composite and water used in the mesh
refining simulations.

Table 1. Cases studied for selecting an appropriated numerical mesh.

Water Composite
Case Mesh ∆t (s)
Te (◦ C) To (◦ C) lx (m) ly (m) lz (m) Rx (m) Ry (m) Rz (m)
20 × 20 × 20
1 20 25 25 0.126 0.0235 0.076 0.01 0.0015 0.01
Nodal Points
30 × 30 × 30
2 20 25 25 0.126 0.0235 0.076 0.01 0.0015 0.01
Nodal Points
40 × 40 × 40
3 20 25 25 0.126 0.0235 0.076 0.01 0.0015 0.01
Nodal Points

Table 2 summarizes all the physical and geometric parameters of the composite and water used in
the refining simulation of the time step.

Table 2. Cases studied for selecting an appropriated time step.

Water Composite
Case Mesh ∆t (s)
Te (◦ C) To (◦ C) lx (m) ly (m) lz (m) Rx (m) Ry (m) Rz (m)
4 10 25 25 0.126 0.0235 0.076 0.01 0.0015 0.01
5 20 × 20 × 20 20 25 25 0.126 0.0235 0.076 0.01 0.0015 0.01
Nodal Points
6 40 25 25 0.126 0.0235 0.076 0.01 0.0015 0.01
Polymers 2019, 11, 1847 10 of 30

2.4.2. Validation
For the validation of the mathematical modeling and numerical solution of the governing
equations, a comparison was made between the numerical results of the average moisture content
obtained in this paper and the analytical (Fick’s model, three-dimensional) and experimental data
reported by Santos [37], for the water absorption process in polymer composites reinforced by sisal
fibers. For preparation of the samples, a polymer material was used, an unsaturated polyester resin
(Resapol 10-316), with low viscosity, and pre-accelerated. This resin is reticulated by styrene peroxide,
using as initiator Methyl Ethyl Ketone (MEK-P) at a concentration of 1% by weight. The method used
was hand lay-up, in which the fibers were arranged randomly and the molding of the composite was
performed by compression.
To obtain the Fickian model from the Langmuir model, the ∂S ∂t
term, in Equation (1), was made null
and, in addition, high values of the sample distance to the container wall (lx , ly , and lz ), in Equation (6),
were used. The objective was to reduce the effect of these parameter on the obtained results, since this
effect is not considered in the Fick’s model. It has been found that this consideration allows that the
concentration of the free solute in the water does not change over time, being constant and equal to the
equilibrium concentration ∂S ∂t
' 0.
Equation (36) represents the three-dimensional (3D) analytical solution for the average moisture
content of a parallelepipedic solid with equilibrium boundary condition, as reported by Santos [37].
 − 
−∗  M − M  ∞ ∞ X
∞
2 2 2
X X
∞ 
M = 1 −   = Bx By Bz e−(βm +βn +βk )D.t (36)

 Mo − M∞ 
m = 1n = 1k = 1

in which M∞ is the equilibrium moisture content, Mo is the initial moisture content of the material,
and Bx , By , and Bz are coefficients given as follows:

2
Bx = (37)
(βm Rx )2

2
By =  2 (38)
βn Ry
2
Bz = (39)
(βk Rz )2
parameters βm , βn , and βk in Equation (34) are the calculated eigenvalues respecting the equilibrium
boundary conditions at the surface of the solid, as follows:

cos(βm Rx ) = 0 (40)

cos(βk Ry ) = 0 (41)

cos(βn Rz ) = 0 (42)

then, by using Equation (38)–(40), we obtain:

(2m − 1)π
βm = (43)
2Rx

(2n − 1)π
βn = (44)
2Ry
(2k − 1)π
βk = (45)
2Rz
Polymers 2019, 11, 1847 11 of 30

Table 3 shows the data used in the simulation for validation with the Fickian model. An isothermal
case was used, µ = λ = 0, D = 3.04 × 10−12 m2 /s and M∞ = 0.1468193 kg/kg, as reported by Santos [37].

Table 3. Parameters used in the simulation for the three-dimensional (3D) case.

Water Composite
Case Mesh ∆t (s)
Te (◦ C) To (◦ C) lx = ly = lz (m) Rx (m) Ry (m) Rz (m)
20 × 20 × 20
7 20 25 25 50 0.01 0.0015 0.01
Nodal Points

2.4.3. Arbitrary Cases

In this item, the influence of the geometrical parameters lx , ly , and lz (see Figure 1) in the process of
water absorption in polymeric composites reinforced by fiber was evaluated. This is an unprecedented
analysis, being the great contribution of this paper. For this analysis, it was considered µ = λ = 1 ×
10−6 s−1 and D = 1 × 10−12 m2 /s. Table 4 summarizes the process parameters used in the simulations.

Table 4. Parameters used in the simulation of arbitrary cases.

Water Composite
Case Mesh ∆t (s)
Te (◦ C) To (◦ C) lx (m) lz (m) ly (m) Rx (m) Rz (m) Ry (m)
8 20 50 25 0.01 0.1 0.01 0.01 0.01 0.0015
9 20 50 25 0.1 0.1 0.01 0.01 0.01 0.0015
10 20 50 25 1 0.1 0.01 0.01 0.01 0.0015
20 × 20 × 20
11 Nodal Points 20 50 25 0.1 0.1 0.001 0.01 0.01 0.0015
12 20 50 25 0.1 0.1 0.1 0.01 0.01 0.0015
13 20 50 25 0.1 0.01 0.01 0.01 0.01 0.0015
14 20 50 25 0.1 1 0.01 0.01 0.01 0.0015

3. Results and Discussion

3.1. Study of Mesh and Time Step

−
Figures 6–8 illustrate the transient behavior of the average free solute concentration (C), the average
− −
entrapped solute concentration (S), and the average moisture absorption (M), respectively, for a total
time of 56 h and different mesh types.
After analyzing these figures, it was observed that, for the three meshes used in the prediction of
water absorption, there is no significant variation in the predicted results with the greater refinement of
the mesh. Thus, considering that the greater the refinement of the mesh, the longer the computational
time to reach the conclusion of the simulation, the mesh with the smallest number of nodal points was
chosen, without loss of information on the behavior of the process parameters. Similar behavior was
found with the time step study (Figure 9), where ∆t = 20 s was chosen for the next analyzes.
3. Results and Discussion

3.1. Study of Mesh and Time Step

Figures 6–8 illustrate the transient behavior of the average free solute concentration (C ), the
average entrapped
Polymers 2019, 11, 1847 solute concentration (S), and the average moisture absorption (M), respectively,
12 of 30
for a total time of 56 h and different mesh types.

Polymers 2019,
Polymers 2019, 11,
11, 1847
1847 12 of
12 of 29
29
Figure
Figure6. Transientbehavior
6.Transient behavior of
of average
average free
free solute
solute concentration
concentration for
for different
different meshes.

Figure
Figure 7.
Figure7. Transient
7.Transient behavior
Transientbehavior of
behavior of the
of the average
the average entrapped
average entrapped solute
entrapped solute concentration
solute concentration for
concentration for different
for different meshes.
different meshes.

Figure 8.
Figure8.
Figure Transient behavior
Transientbehavior
8.Transient of
behavior of the
of the average
the average moisture
average moisture content for
moisture content for different
different meshes.
meshes.

After analyzing
After analyzing these
these figures,
figures, it
it was
was observed
observed that,
that, for
for the
the three
three meshes
meshes used
used in
in the
the prediction
prediction
of water
of water absorption,
absorption, there
there is
is no
no significant
significant variation
variation inin the
the predicted
predicted results
results with
with the
the greater
greater
refinement of the mesh. Thus, considering that the greater the refinement of the mesh, the longer
refinement of the mesh. Thus, considering that the greater the refinement of the mesh, the longer the the
computational time to reach the conclusion of the simulation, the mesh with the smallest
computational time to reach the conclusion of the simulation, the mesh with the smallest number of number of
Polymers 2019, 11, 1847 13 of 29
Polymers 2019, 11, 1847 13 of 30
Polymers 2019, 11, 1847 13 of 29

Figure 9. Transient behavior of the average moisture content for different time steps.
Figure
Figure 9.
9. Transient
Transient behavior
behavior of
of the
the average
average moisture
moisture content for different time steps.
3.2. Validation: Application of the Langmuir Model as Fick’s Model
3.2. Validation:
3.2. Validation: Application
Application of the
the Langmuir
Langmuir Model
Model asas Fick’s
Fick’s Model
Model
The numerical resultsofstrongly depend on the boundary conditions, thermophysical properties,
The numerical
The numerical
and geometry results strongly
considered.
results strongly depend
Figure 10depend on the
illustrates
on the boundary
theboundary
comparison conditions, thermophysical
betweenthermophysical
conditions, the results of theproperties,
average
properties,
and geometry
moisture considered.
content obtained Figure
with 10
the illustrates
numerical the comparison
solution, from between
the the results
approximation
and geometry considered. Figure 10 illustrates the comparison between the results of the average of of
thethe average
Langmuir
moisture
model
moistureto content
the
contentobtained
Fick’s model,with
obtained and
with the numerical
experimental
the numerical solution,
and fromfrom
analytical
solution, the approximation
data
the(Fick’s of the
model)
approximation Langmuir
reported
of model
by Santos
the Langmuir
to the Fick’s model, and experimental and analytical data (Fick’s model) reported by
[37]. to the Fick’s model, and experimental and analytical data (Fick’s model) reported by Santos
model Santos [37].
[37].

Figure 10. Comparison

Comparison between
between the
the average
average dimensionless
dimensionless moisture
moisture contents
contents predicted
predicted by
by the
numerical and analytical solutions and the experimental data as a function of the
Figure 10. Comparison between the average dimensionless moisture contentsprocess process time.
time.by the
predicted
numerical and analytical solutions and the experimental data as a function of the process time.
After analyzing
analyzing Figure 10, it was observed an excellent agreement between the data predicted
by the numerical
After solution,
analyzing Figurewith
10, itthose analyticalan
was observed and experimental,
excellent reported
reported
agreement by Santos
between the data[37], which
predicted
shows that the mathematical model presents an adequate description of the diffusion process
by the numerical solution, with those analytical and experimental, reported by Santos [37], which inside
the material.
shows that theTherefore,
Therefore, the Fick’s model is a particular case of the Langmuir diffusion model.
model. inside
mathematical model presents an adequate description of the diffusion process
the material. Therefore, the Fick’s model is a particular case of the Langmuir diffusion model.
3.3. Application
3.3. Application to
to Arbitrary
Arbitrary Cases
Cases
3.3. Application
The to Arbitrary
The absorption
absorption Cases
ofofwater
wateris facilitated when
is facilitated the polymer
when molecules
the polymer have groups
molecules capable capable
have groups of forming
of
hydrogen
forming bonds.
hydrogen Vegetable
bonds. fibers
Vegetable are rich
fibersin cellulose,
are rich in hemicellulose,
cellulose, and lignin
hemicellulose, which
and
The absorption of water is facilitated when the polymer molecules have groups capable of have
lignin hydroxy
which have
groups, hydrogen
hydroxy
forming thus have
groups, high
thus affinity
have
bonds. for fibers
high affinity
Vegetable water. The
forare richabsorption
water. The
in ofhemicellulose,
absorption
cellulose, water by the
of water resin,
byand in turn,
the lignin
resin, can
in turn,
which be
can
have
hydroxy groups, thus have high affinity for water. The absorption of water by the resin, in turn, can
Polymers 2019, 11, 1847 14 of 30

Polymers 2019, 11, 1847 14 of 29

considered practically
be considered null, since
practically it since
null, presents a considerable
it presents hydrophobic
a considerable hydrophobiccharacter [38].
character The
[38]. Theaddition
of the vegetable
addition offibers to the resin
the vegetable fibersgenerates
to the resinangenerates
increasean inincrease
the water absorption
in the levels, levels,
water absorption so an important
so
an important parameter to be analyzed is how much water is being
parameter to be analyzed is how much water is being absorbed by the composite over time. absorbed by the composite over
Fortime.
the analysis of the free (C) and entrapped (S) water mass distributions inside the composite,
For the analysis of the free (C) and entrapped (S) water mass distributions inside the composite,
the results of these parameters were plotted at different plans and times of process (x = Rx /2, y = Ry /2
the results of these parameters were plotted at different plans and times of process (x = Rx/2, y = Ry/2
and z =andRz z/2;= Rt1z/2;
 t61 ≅h, t2t2≅ 54
6 h, h,t3t≅3 211
54 h,  211 h, tand
h, and t  417 h). Figure 11 schematically illustrates the
4 ≅ 4174h). Figure 11 schematically illustrates the regions,
regions,ininthe
the Cartesian system, where the distributions
Cartesian system, where the distributions of the parameters of the parameters of interest
of interest will will be analyzed.
be analyzed.

(a) (b)

(c)
Figure
Figure 11. 11. Cartesian
Cartesian plans plans
usedused
forfor analyzing the
analyzing the free
freeand
andentrapped
entrappedsolute concentration
solute inside the
concentration inside the
composite.
composite. (a) x =(a)Rxx /2; /2; (b)
= Rx(b) y =y R
= R/2,
y/2, and (c) z = Rz/2.
and (c) z = R /2.
y z

3.3.1. of
3.3.1. Effect Effect of Distance
Distance lx lx
Figures 12–14 shows the kinetics of Langmuir model parameters C, S −and
− average moisture
Figures
content12–14
(M) forshows thelx kinetics
different distances, of Langmuir
keeping model
the other parameters
physical C, S parameters
and geometric and average moisture
of the
−
sample fixed.
content (M) for different lx distances, keeping the other physical and geometric parameters of the
sample fixed.
Polymers 2019, 11, 1847 15 of 29

Figure
Figure 12.12. Kinetics
Kinetics ofof the
the average
average free
free solute
solute concentrationfor
concentration fordifferent
differentdistances
distanceslxlx(Cases
(Cases8,8,9,9,and
and10).
10).
 Figure 12.
Figure 12. Kinetics of
of the
the average
average free
free solute
solute concentration
concentration for
for different
different distances
distances llxx (Cases
(Cases 8,
8, 9,
9, and
and
Polymers 2019, 11, Kinetics
1847 15 of 30
10).
10).

Figure
Figure 13.
Figure 13. Kinetics
13. Kinetics of
Kinetics of the average
of the average entrapped
entrapped solute
entrapped soluteconcentration
solute concentrationfort
concentration fortdifferent
fort differentdistances
different distancesl llx(Cases
distances x (Cases 8,
x (Cases8,8, 9,
9,
9,
andand
and 10).
10).
10).

Figure 14.
Figure 14.
Figure Kinetics of
Kineticsof
14. Kinetics the
ofthe average
theaverage moisture
averagemoisture content
moisture content for
content for different
for different
different distances (Cases
distances llxxx (Cases 8,
(Cases 8, 9,
8, 9, and
9, and 10).
and 10).
10).

From the analysis of these figures, it can be observed that, even with a significant change in the
values of the sample distance to the container wall in the x-direction (lx ), there is small change in the
− − −
average free (C) and entrapped (S) solute concentration and average moisture content (M) throughout
the process.
The variation of this parameter (lx ) demonstrates the amount of water in the x direction around
the sample, in other words, changing lx means varying the position, in the x direction, of the immersed
composite or the dimensions of the container used in the bath. Thus, with this analysis, it is possible to
verify that these variations do not directly influence the absorption of moisture under the conditions
studied. Thus, for the initial and boundary conditions established, the variation of the lx values does
not lead to a significant change in the three-dimensional and transient moisture absorption parameters
of the Langmuir model.
Figures 15–17 illustrate the distributions of free solute concentration in the plans x = Rx /2, z = Rz /2,
and y = Ry /2, respectively, in a total time t = 211 h, for the cases where variation in the distance of the
sample to the container wall (lx ) was stablished.
absorption parameters of the Langmuir model.
values does not lead to a significant change in the three-dimensional and transient moisture
Figures 15–17 illustrate the distributions of free solute concentration in the plans x = Rx/2, z =
absorption parameters of the Langmuir model.
Rz/2, and y = Ry/2, respectively, in a total time t = 211 h, for the cases where variation in the distance
Figures 15–17 illustrate the distributions of free solute concentration in the plans x = Rx/2, z =
of the sample to the container wall (lx) was stablished.
Rz/2, and y = Ry/2, respectively, in a total time t = 211 h, for the cases where variation in the distance
Polymers 2019, 11, 1847 16 of 30
of the sample to the container wall (lx) was stablished.

(a)

(a)

(b)

(b)

(c)
Figure 15. Distribution of free solute concentration (c)(C in kg/kg) in plan x = Rx/2 for different distances
lx (t = 211 h, Cases 8, 9, and 10). In which (a) lx = 0.01m; ly = 0.01 m and lz = 0.1 m; (b) lx = 0.1 m; ly = 0.01
Figure 15. Distribution
15. Distribution
Figure of of free
free soluteconcentration
solute concentration (C (Cininkg/kg)
kg/kg)inin
plan x =xR=xR
plan /2x/2
forfor
different distances
different distances
m and lz = 0.1 m; (c) lx = 1 m; ly = 0.01 m and lz = 0.1 m.
l (t = 211 h, Cases 8, 9, and 10). In which (a)
lx (t =x211 h, Cases 8, 9, and 10). In which (a) lx = 0.01m; lx = 0.01 m; l =
ly = 0.01
y 0.01 m and l
m and lz = 0.1 = 0.1 m; (b)
z m; (b) lx = 0.1l =
x m; lym;
0.1 = 0.01
m and l = 0.01 m and l = 0.1 m; (c) l = 1 m;
y lz = 0.1 m; (c)zlx = 1 m; ly = x0.01 m andl y = 0.01 m and
lz = 0.1 m. z l = 0.1 m.

(a)

(a)

Polymers 2019, 11, 1847 17 of 29

(b)

(b)

(c)
FigureFigure 16. Distribution
16. Distribution of of free
free soluteconcentration
solute concentration (C
(Cininkg/kg)
kg/kg) inin
plan z =zR=z /2
plan Rzfor different
/2 for distances
different distances
l (t = 211 h, Cases 8, 9, and 10). In which (a) l
lx (t =x211 h, Cases 8, 9, and 10). In which (a) lx = 0.01x = 0.01 m; l =
m; ly = 0.01
y 0.01 m and l
m and lz = 0.1 = 0.1 m; (b)
z m; (b) lx = 0.1 lx = m;
0.1lm;
y = 0.01
ly = 0.01 m and lz = 0.1 m; (c) lx = 1 m; ly = 0.01 m and lz = 0.1 m.
m and lz = 0.1 m; (c) lx = 1 m; ly = 0.01 m and lz = 0.1 m.

From the analysis of these figures, it can be seen that in the plans x = Rx/2 and z = Rz/2, the increase
in the value of lx does not affect the free solute concentration distribution, which confirms the
independence of the behavior of the water absorption phenomenon studied, with this parameter lx.
For the free solute concentration in the y = Ry/2 plan (Figure 16), in the case where lx = 0.01 m, there is
a small reduction in free solute concentration, in the x-direction, compared with that in the z-
 From the analysis of these figures, it can be seen that in the plans x = Rx/2 and z = Rz/2, the increase
in the value of lx does not affect the free solute concentration distribution, which confirms the
independence of the behavior of the water absorption phenomenon studied, with this parameter lx.
For the free solute concentration in the y = Ry/2 plan (Figure 16), in the case where lx = 0.01 m, there is
a small2019,
Polymers reduction
11, 1847 in free solute concentration, in the x-direction, compared with that in 17 the z-
of 30
directions.

(a) (b)

(c)
Figure17.
Figure 17. Distribution
Distributionof offree
freesolute
soluteconcentration
concentration(C (Cin
inkg/kg) planyy== RRyy/2
kg/kg)ininplan /2 for
for different
differentdistances
distances
llxx (t =
= 211
211h,h,Cases
Cases8,8,9,9,and
and10).
10).InInwhich
which(a)(a) = 0.01
lx =lx0.01 m; lm; ly = m
y = 0.01 and
0.01 mland
z = 0.1 = 0.1
lz m; (b) m;
lx =(b) x =ly0.1
0.1 lm; = 0.01
m;
lmy =and0.01lzm andm;
= 0.1 = 0.1
lz (c) lx =m; (c) llyx==0.01
1 m; = 0.01
1 m;mlyand lz = m m. lz = 0.1 m.
0.1and

From the analysis of these figures, it can be seen that in the plans x = Rx /2 and z = Rz /2, the increase
in the value of lx does not affect the free solute concentration distribution, which confirms the
independence of the behavior of the water absorption phenomenon studied, with this parameter lx .
For the free solute concentration in the y = Ry /2 plan (Figure 16), in the case where lx = 0.01 m, there is a
small reduction in free solute concentration, in the x-direction, compared with that in the z-directions.
The distributions of the entrapped solute concentration in the plans x = Rx /2, z = Rz /2 and y = Ry /2,
at t = 211 h, keeping the values of lz and ly constant and varying the value of lx can be observed in
Figures 18–20. In this case, in neither plan is observed significant variations of the parameter S with
the modification on the value of the sample distance to the container wall in the x direction (lx ).
Polymers 2019, 11, 1847 18 of 29
The distributions of the entrapped solute concentration in the plans x = Rx/2, z = Rz/2 and y =
Ry/2, The
at t =distributions
211 h, keeping of the
the entrapped
values of lzsolute
and ly concentration in the plans
constant and varying x = Rof
the value x/2,
lx zcan
= Rbe
z/2 and y =
observed
Ry/2,
in at t = 18–20.
Figures 211 h, Inkeeping thein
this case, values of plan
neither lz and
is lobserved
y constantsignificant
and varying the value
variations of of
thelxparameter
can be observed
S with
in Figures
the 2019,18–20.
modification
Polymers In this
11, 1847on the case,
valueinofneither plan is
the sample observed
distance significant
to the variations
container of the
wall in the parameter
x direction S with
(lx).18 of 30
the modification on the value of the sample distance to the container wall in the x direction (lx).

(a)
(a)

(b)
(b)

(c)
(c)
Figure 18. Distribution of entrapped solute concentration (S in kg/kg) in plan x = Rx/2 for different
Figure
Figure
distances 18.lxDistribution
18. Distribution of
of entrapped
(t = 211 h, Cases entrapped solute
solute
8, 9, and 10). concentration
concentration
In which (S
(S in
(a) lx = 0.01 kg/kg)
inm; in
in plan
ly = 0.01
kg/kg) plan x == R
m xand /2= for
Rlxxz/2 0.1 different
for m; (b) lx
different
distances l (t = 211 h, Cases 8, 9, and 10). In which l (a)
=
= 0.1 m; ly = 0.01 m and lz = 0.1 m; (c) lx = 1 m; ly = 0.01 m and lz = 0.1 m.
distances l x
x (t = 211 h, Cases 8, 9, and 10). In which (a) x lx =
0.01 0.01
m; ly m;
= l
0.01
y = m0.01
and m l and
z = l
0.1z =
m0.1
; (b)m;
lx
(b) l = 0.1 m; l = 0.01 m and l = 0.1 m; (c) l = 1 m; l = 0.01
= 0.1 m; ly = 0.01 m and lz = 0.1 m; (c) lx = 1 m; ly = 0.01 m and lz = 0.1 m.
x y z x y m and lz = 0.1 m.

(a)
(a)

(b)
(b)

(c)
(c)
Figure 19. Distribution of the entrapped solute concentration (S in kg/kg) in the plan z = Rz /2 and
different distances lx (t = 211 h; Cases 8, 9, and 10). In which (a) lx = 0.01 m; ly = 0.01 m and lz = 0.1 m;
(b) lx = 0.1 m; ly = 0.01 m and lz = 0.1 m; (c) lx = 1 m; ly = 0.01 m and lz = 0.1 m.
Polymers 2019, 11, 1847 19 of 29

Figure 19. Distribution of the entrapped solute concentration (S in kg/kg) in the plan z = Rz/2 and
different distances lx (t = 211 h; Cases 8, 9, and 10). In which (a) lx = 0.01 m; ly = 0.01 m and lz = 0.1
m; (b)2019,
Polymers lx =11,
0.11847
m; ly = 0.01 m and lz = 0.1 m; (c) lx = 1 m; ly = 0.01 m and lz = 0.1 m. 19 of 30

(a) (b)

(c)
Figure
Figure 20.20. Distribution
Distribution of the
of the entrapped
entrapped solute
solute concentration
concentration (S (S
in in kg/kg)
kg/kg) in in
plan y =yR=y/2
plan /2 for
Ryfor different
different
distances
distances = 211
lx (tlx=(t211 h; Cases
h; Cases 8, 9, 8,
and9, and 10).
10). In In which
which (a) lx (a) lx =m;
= 0.01 0.01
ly m; ly =m
= 0.01 0.01
andm land lz =m0.1
z = 0.1 m;lx(b)
; (b)
lx =m;
= 0.1 0.1lym; ly = 0.01
= 0.01 m andm and = 0.1
lz =lz0.1 m; m; lx =lx1=m;
(c) (c) 1 m; = 0.01
ly =ly0.01 mm andandlz l=z =
0.10.1mm.
.

It observed
It is is observed that
that thethe increase
increase in inthethe C and
C and S concentrations
S concentrations of of
thethe Langmuir
Langmuir model
model occurs
occurs from
from
the surface to the center of the composite. For the concentration of free
the surface to the center of the composite. For the concentration of free solute (C), it is observed thatsolute (C), it is observed
thethat
fluxtheof flux of moisture
moisture in the xinand
the zx direction
and z direction
occursoccurs symmetrically
symmetrically since, in since,
this in thisthe
case, case, the sample
sample has
thehas thedimensions
same same dimensions(Rx = Rz(R x =R
). For thez ).same
For the
timesame time ofthe
of process, process, the concentration
concentration of free solute of free solute
showed
showed
higher higher
values than values than the concentration
the concentration of entrapped of entrapped
solute in solute in the
the inner inner of
regions regions of the composite.
the composite. This
This behavior is probably due to the effects of the adsorptive forces being smaller
behavior is probably due to the effects of the adsorptive forces being smaller than those presented by than those presented
thebyforces
the forces
fromfrom the free
the free waterwater concentration
concentration gradient,
gradient, which
which are are responsible
responsible for for
thethe migration
migration of of
water into the composite. For the concentration of entrapped solute, it is observed that the changes in
water into the composite. For the concentration of entrapped solute, it is observed that the changes
in this
this variable
variable occur
occur more
moreslowly,
slowly,however
howeverwith withthe thegreatest
greatestintensity
intensityininthethehorizontal
horizontal directions
directions(x(x
and
z directions).
and z directions).
Polymers 2019, 11, 1847 20 of 30

3.3.2. Effect of the Distance ly

Figures 21–23 show that, unlike what happened for the lx distance effect, the average free solute
− −
concentration (C), average entrapped solute concentration (S), and the average moisture absorption
−
kinetics (M), for different distances from the sample to the walls of the container in the y direction (ly )
increase
Polymers over
Polymers 2019,
2019, 11,time.
11, 1847 In this analysis, others physical and geometric parameters were kept fixed.20
1847 20 of
of 29
29
From the analysis of the Figures 21–23, we can see that variations of ly values strongly influence the
3.3.2.
3.3.2. Effect
average freeof
Effect of the
the Distance
solute Distance llyy
concentration, the average entrapped solute concentration, and average moisture
content. The
Figures higher
Figures 21–23
21–23 show the l value
showythat, unlikehigher
the
that, unlike what the water layer
what happened
happened for close
for the
the to the composite
llxx distance
distance effect, thewall
effect, the and faster
average
average free shall
free solute
solute
be the water absorption
concentration
concentration (C
(C),), average
average rate. For initial
entrapped
entrapped times
solute
solute of the process,
concentration
concentration (S the differences
(S),), and
and the averageamong
the average moisture
moisturetheabsorption
predicted
absorption
results
kinetics are
kinetics (M not
(M),), for significant,
for different however
different distances
distances from in the
from the course
the sampleof
sample totime
to the it can
the walls
walls of be clearly
of the observed
the container
container inin thethe direction (l(lof
influence
the yy direction yy))
this parameter.
increase
increase over
over time.
time. In In this
this analysis,
analysis, others
others physical
physical and
and geometric
geometric parameters
parameters werewere kept
kept fixed.
fixed.

Figure
Figure 21.
Figure 21. Kinetics
21. Kineticsof
Kinetics ofofthe
the average
average
the free
free
average solute
solute
free concentration
concentration
solute for different
forfor
concentration different distances
distances
different llyy (Cases
distances (Cases 9,
9, 11,
ly (Cases11,9,and
and
11,
12).
12). 12).
and

Figure
Figure 22.
22. Kinetics
Kinetics of
Kinetics the
of the average
theaverage entrapped
averageentrapped solute
entrappedsolute concentration
soluteconcentration forfor
concentration for different
different
different distances
distances
distances llyy (Cases
ly (Cases(Cases 9,
9,
9, 11,
11,
11, and
and 12).12).
and 12).
Polymers 2019, 11, 1847 21 of 30
Polymers 2019, 11, 1847 21 of 29

Figure 23. Kinetics of the average moisture content for different ly (Cases 9, 11, and 12).

From the analysis of the Figures 21–23, we can see that variations of ly values strongly influence
the average free solute concentration, the average entrapped solute concentration, and average
moisture content. The higher the ly value the higher the water layer close to the composite wall and
faster shall be the water absorption rate. For initial times of the process, the differences among the
predictedFigure
results23.are not significant,
Kinetics of the average however
moistureincontent
the course of time
for different
different it can9,be
llyy (Cases 11,clearly
and 12).observed the
influence of this parameter.
Figures
From the24–26
Figures 24–26 illustrate
analysis of thethe
illustrate thefree
Figuresfreesolute
21–23,
solute concentration
we can see that
concentration distribution,
variations in
distribution, ofinthe
ly the plans
values plans x= xR= xR/2,
strongly z z= =RR
influence
x/2, z /2,
z/2,

and
the y y==RR
andaverage y /2, respectively,
free
y/2, atatt =
t =211
solute concentration,
respectively, 211h,h,forfordifferent
the averagevalues
different valuesofoflyl.ysolute
entrapped .After
Afteranalysis
analysisofofthese
concentration, figures,
and
these average
figures, ititisis
possible
possibletocontent.
moisture toobserve that,
The
observe inin
higher
that, the xly=
plans
plans = xR/2x/2and
R
xvalue the
andz=z =RzR/2,
higher thefor
z/2, forsmall
water value
layer
small close
value ofoflytol(ly(ly==composite
y the 0.001
0.001m),m),the wall
the fluxand
flux ofof
free
freesolute
faster shall practically
solute be the water
practically does not
notoccur
absorption
does occur ininyyFor
rate. direction, ininopposite
initial times
direction, of thewe
opposite wecancansee
process, the
see ititdifferences
ininthe
theother
other directions.
among the
directions.
Then,
Then,ititverified
predicted results the
verified thefree
are freewater
not watermolecules
significant, however
molecules move in horizontally
move the course of(x(x
horizontally and
timeand itz zcan
direction),
be clearly
direction), ininpractice,
practice,with
observed the
with
velocity
influence almost
velocity of equal.
this parameter.
almost equal.
Figures 24–26 illustrate the free solute concentration distribution, in the plans x = Rx/2, z = Rz/2,
and y = Ry/2, respectively, at t = 211 h, for different values of ly. After analysis of these figures, it is
possible to observe that, in plans x = Rx/2 and z = Rz/2, for small value of ly (ly = 0.001 m), the flux of
free solute practically does not occur in y direction, in opposite we can see it in the other directions.
Then, it verified the free water molecules move horizontally (x and z direction), in practice, with
velocity almost equal.

(a)

(b)
(a)

(b)
(c)

Figure 24. Distribution of free solute concentration (C in kg/kg) in plan x = Rx /2 for different distances
ly (t = 211 h, Cases 9, 11, and 12). In which (a) ly = 0.001 m; lx = 0.1 m and lz = 0.1 m; (b) ly = 0.01 m;
lx = 0.1 m and lz = 0.1 m; (c) ly = 0.1 m; lx = 0.1 m and lz = 0.1 m.

(c)
Polymers 2019, 11, 1847 22 of 29

Polymers 2019, 11, 1847 22 of 29

Figure 24. Distribution of free solute concentration (C in kg/kg) in plan x = Rx/2 for different distances
lyFigure
(t = 21124.
h, Distribution
Cases 9, 11, and 12).solute
of free In which (a) ly = 0.001
concentration m;kg/kg)
(C in lx = 0.1
in m and
plan x =lzR=x/20.1
formdifferent
; (b) ly = distances
0.01 m;
lxly=(t0.1 m and l z = 0.1 m; (c) ly = 0.1 m; lx = 0.1 m and lz = 0.1 m.
= 211 h, Cases 9, 11, and 12). In which (a) ly = 0.001 m; lx = 0.1 m and lz = 0.1 m; (b) ly = 0.01 m;
Polymers 2019, 11, 1847 22 of 30
lx = 0.1 m and lz = 0.1 m; (c) ly = 0.1 m; lx = 0.1 m and lz = 0.1 m.

(a)
(a)

(b)
(b)

(c)
Figure 25. Distribution of free solute concentration(c) (C in kg/kg) in plan z = Rz/2 for different distances
y (t = 211 h, Cases 9, 11, and 12). In which (a) ly = 0.001 m; lx = 0.1 m and lz = 0.1 m; (b) ly = 0.01 m;
lFigure
Figure25. 25. Distribution
Distributionof offree
freesolute
soluteconcentration
concentration(C (Cin
inkg/kg) planzz==RRzz/2
kg/kg)ininplan /2 for
for different
differentdistances
distances
llxlyy=(t(t0.1
= m and
= 211 lz = 9,
h, Cases
h, Cases 0.1 mand
9, 11,
11, and l12).
y = 0.1
; (c) 12). m; lx (a)
Inwhich
In which =(a)
0.1
lyly=m and m;
=0.001
0.001lm;
z = l0.1 m
0.1. m
lxx==0.1 and lz == 0.1
m and 0.1 m = 0.01
m;; (b) ly = m;
0.01 m;
llxx == 0.1 andlzl=
0.1mmand z =0.1
0.1m;m(c) ly l=
; (c) 0.1
y = lx =
m;m;
0.1 lx 0.1 mm
= 0.1 lz =lz0.1
andand m. m.
= 0.1

(a) (b)
(a) Figure 26. Cont. (b)
Polymers 2019, 11, 1847 23 of 30
Polymers 2019, 11, 1847 23 of 29

Polymers 2019, 11, 1847 23 of 29

(c)
(c)
Figure
Figure 26. 26. Distribution
Distribution of of free
free solute concentration (C in kg/kg) in plan y == RRyy/2/2for
solute concentration fordifferent
different distances
distances
llyy (t(t ==211
Figure
211h, 26.
h,Cases
CasesDistribution
9,9,11,
11, and
and of12).
free In
12). solute (a)llyy ==0.001
concentration
In which:
which: (a) m; llxx =
m;
(C in kg/kg)
0.001 =in0.1
0.1 myand
plan
m =and
Ry/2 =
lz lfor
z = 0.1m;m(b)
different
0.1 ly l=y 0.01
= 0.01
distances
; (b) m;
lx =lx0.1
m; m(t =and
=ly0.1 m z =Cases
211and
lh, l0.1 9, 11,
z = m;
0.1 mland
(c) = 12).
;y(c) 0.1 In
ly =m;
0.1 = 0.1
which:
lxm; lx(a) lyand
=m0.1 =m0.001= m;
lzand lzlm.
0.1 x==0.1
0.1mm. and lz = 0.1 m; (b) ly = 0.01
m; lx = 0.1 m and lz = 0.1 m; (c) ly = 0.1 m; lx = 0.1 m and lz = 0.1 m.
Figures 27–29
Figures 27–29 show
show thethe distribution
distribution of of entrapped
entrapped solute
solute concentration
concentration in time tt == 211
in time 211 h, h, for
for
different Figures
plans and27–29 show the
different l distributionItofcan
distances. entrapped
be solute aconcentration
observed similar in timethat
behavior t = presented
211 h, for by
different plans and different ly ydistances. It can be observed a similar behavior that presented by free
different plans and different ly distances. It can be observed a similar behavior that presented by free
free solute
solute concentration.
concentration. Then, Then, in general,
in general, thethethe geometric
geometric parameter ly affects distribution of the andfree
solute concentration. Then, in general, geometricparameter
parameter llyy affects
affects distribution
distribution of of
thethe
freefree
and
and entrapped
entrapped solute
solutesolute concentration
concentration and, and,
in in in consequence,
consequence, total moisture of the composite immersed
entrapped concentration and, consequence,totaltotal moisture
moisture of ofthe
thecomposite
composite immersed
immersed in in
in water.
water. Since
Since
water.
these
these
Since areare
these thethe first
first
are the
results
results
first onon
results
the
the
on
subject
thesubject
that demonstrate
subjectthat
demonstratethat
that demonstrate
that
that the
the
the parameter
parameter
parameter 11yy has
1y has
has
influence
influence on the
on
influencetheonkinetics
kinetics of moisture
of
the kinetics moisture
of moistureabsorption,
absorption,
absorption, the
the authors
theauthors recommend
authorsrecommend strongly llnew
strong
recommend strong new
y new
y studies
studies
studies related
related
related
to this topic,
to this to in order
this topic,
topic, to
in order
in order classify this
to classify
to classify phenomenon.
thisthis phenomenon.
phenomenon.

(a)
(a)

(b)

(b)

(c)

Figure 27. Distribution of entrapped solute concentration (S in kg/kg) in plan x = Rx /2 for different
distances ly (t = 211 h, Cases 9, 11, and 12). In (c)
which (a) ly = 0.001 m; lx = 0.1 m and lz = 0.1 m; (b)
ly = 0.01 m; lx = 0.1 m and lz = 0.1 m; (c) ly = 0.1 m; lx = 0.1 m and lz = 0.1 m.
 Polymers
Polymers 2019,
2019, 11,11, 1847
1847 2424
ofof
2929

Figure27.27.Distribution
Figure Distributionofofentrapped
entrappedsolute
soluteconcentration
concentration(S(Sininkg/kg)
kg/kg)ininplan
planx x= =RxR/2x/2for
fordifferent
different
distances ly (t = 211 h, Cases 9, 11, and 12). In which (a) ly = 0.001 m; lx = 0.1 m and lz = 0.1 mm
distances l y (t = 211 h, Cases 9, 11, and 12). In which (a) l y = 0.001 m; lx = 0.1 m and l z = 0.1 ly ly
; (b)
; (b)
= 0.01
= 0.01
Polymers
m; l
m;11,lx1847
2019,
= 0.1 m and l = 0.1 m ; (c) l = 0.1 m; l = 0.1 m
= 0.1 m and lz = 0.1 m; (c) ly = 0.1 m; lx = 0.1 m and lz = 0.1 m.
x z y x and l z = 0.1 m . 24 of 30

(a)(a)

(b)
(b)

(c)(c)
Figure28.
Figure 28.Distribution
Distributionofofofentrapped
entrappedsolute soluteconcentration
concentration(S(S (Sinin
inkg/kg)
kg/kg)inin inplan
planzzz===RRzR/2z/2
/2 for
fordifferent
different
Figure 28. Distribution entrapped solute concentration kg/kg) plan z for different
distances
distances
distances y (t
lyll(t
y (t
==211
= 211 h,h,
211 h,Cases
Cases 9,9,
Cases 11,11,
9, and
and
11, 12).
12).
and In In which
which
12). (a)(a)
In which ly l=y(a)=0.001
0.001
ly = m; m;lx l=xm;
0.001 =0.1
0.1
lxm=mand
and
0.1 mlz land
=z =0.10.1
lzm=m ; (b)
; (b)
0.1 y ly
lm;
=
=(b)0.01
ly =
0.01 m;
m;0.01 l
lx =m;
x = 0.1 m
= 0.1
0.1lx m and l
andmlzand =
= 0.1
z 0.1 m
lz =m0.1 ; (c)
; (c)m; l
ly (c) = 0.1 m;
ly =m;
= 0.1
y l
0.1lxm; = 0.1
lx =m
= 0.1
x m
0.1and and
m and l = 0.1
z=m
lz = l0.1
z m .
0.1. m.

(a)(a) (b)
(b)
Figure 29. Cont.
Polymers 2019, 11, 1847 25 of 29

Polymers 2019, 11, 1847 25 of 30

Polymers 2019, 11, 1847 25 of 29

(c)
Figure 29. Distribution of the entrapped solute concentration (S in kg/kg) in plan y = Ry/2 for different
distances ly (t = 211 h, Cases 9, 11, and 12). In which (a) ly = 0.001 m; lx = 0.1 m and lz = 0.1 m; (b) ly
= 0.01 m; lx = 0.1 m and lz = 0.1 m; (c) ly = 0.1 m; lx = 0.1 m and lz = 0.1 m.
(c)
From
Figurethe
Figure 29. analysis of of
29. Distribution
Distribution these
of figures, itsolute
the entrapped
the entrapped is observed
solute that (S
concentration
concentration the infree
(Sin solute
kg/kg)
kg/kg) concentration
ininplan
plan yy== RRyy/2
/2 for is going to
for different
different
increase from
distances sample
distances llyy (t ==211 surface
211h,h,Cases to
Cases9,9,11,center
11,and with
and12).12).InIna flux
which
which of moisture
ly =ly 0.001
(a)(a) that
= 0.001m;m; is typically
= 0.1
lx l=x 0.1 mm and horizontal,
= 0.1
andlz l=z 0.1 mm; thatly is, a
(b)
; (b)
greater
l=y =flux occurs
0.01m;
0.01 in
m;lxlx==0.10.1mthe
mand x and z
and lzz ==0.1 direction
0.1 m comparing
m;; (c) lyy = with
m;llxx == 0.1
= 0.1 m; 0.1m that in the
andlzlz==0.1
mand y-direction.
0.1m.m. The same behavior
repeats when it is considered the entrapped solute concentration (Figures 27–29), nevertheless this
From
From the
concentration analysis
analysis of
theincreases of these
more these figures,
slowly itit isis observed
in comparison
figures, observed that
to the the
the free
thatbehavior free solute
of
solute concentration
the free is
is going
going to
solute concentration,
concentration to
increase
mainly from
increaseinfrom sample
the initial
sample surface
times to
of process.
surface center with a flux of moisture that is
to center with a flux of moisture that is typically horizontal, that typically horizontal, thatis,is,a
agreater
greaterfluxfluxoccurs
occursin inthe
thexxand andzzdirection
direction comparing
comparing with with thatthat in
in the
the y-direction.
y-direction. The The same
same behavior
behavior
3.3.3.
repeats Effect of Distance l
repeats when it is considered the entrapped solute concentration (Figures 27–29), nevertheless this
when it is considered z the entrapped solute concentration (Figures 27–29), nevertheless this
concentration
concentration increases
increases more
more slowly
slowly in
in comparison
comparison to
to the
the behavior
behavior of
of the
the free
free solute
solute concentration,
concentration,
Figures 30–32 illustrate kinetics of average free and entrapped solute concentration, and the
mainly
mainly in in the
theinitial
initial times
timesfor ofofprocess.
process.
average moisture content different distances of the sample, to the walls of the container in the z
direction
3.3.3. Effect (lz)ofasDistance
a function lz of the process time.
3.3.3.ItEffect of Distance
is observed lz
no significant difference in the mass absorption phenomenon studied (moisture
Figures
absorption) 30–32
with the illustrate
variation kinetics
of l of average
z. Then,
free and
it is possible entrapped
to notice that only solute concentration,
ly parameter and the
can influence
Figures 30–32 illustrate kinetics of average free and entrapped solute concentration, and the
average on
directly moisture
the kineticscontent for different
of moisture distances
absorption of studied
in the the sample, to the walls
conditions. However,of the container
lz valuesinin the
average moisture content for different distances of the sample, to the walls of thefor container lower
the z
z direction
than the 0.01 (l ) as a function of the process time.
direction (lz)zm,
as athe effect of
function ofthis
the geometrical
process time.parameter must appear more clearly.
It is observed no significant difference in the mass absorption phenomenon studied (moisture
absorption) with the variation of lz. Then, it is possible to notice that only ly parameter can influence
directly on the kinetics of moisture absorption in the studied conditions. However, for lz values lower
than the 0.01 m, the effect of this geometrical parameter must appear more clearly.

Figure
Figure 30.
30. Kinetics
Kineticsofofthe
theaverage free
average solute
free concentration
solute for for
concentration different distances
different lz (Cases
distances 9, 13,9,and
lz (Cases 13,
14).
and 14).

Figure 30. Kinetics of the average free solute concentration for different distances lz (Cases 9, 13, and
14).
Polymers 2019, 11, 1847 26 of 30
Polymers 2019, 11, 1847 26 of 29

Figure
Figure 31.
31. Kinetics of the
Kinetics of the average
averageentrapped
entrappedsolute
soluteconcentration
concentrationforfor different
different distances
distances lzz (Cases
lz (Cases 9,
9, 13,
13,
andand
14).14).

Figure
Figure 32.
32. Kinetics
Kinetics of
of the
the average
average moisture
moisture content
content for
for different
different distances
distances llzzz (Cases
(Cases 9,
9, 13,
13, and
and 14).
14).

It
It is noteworthy
is observed that the distributions
no significant difference inofthe free
massandabsorption
entrappedphenomenon
solute concentrations,
studied (moisture for the
different
absorption) planes
withanalyzed (x = Rofxx/l2,
the variation z. y = Ryy/it2,iszpossible
Then, = Rzz/ 2), to
at notice
a giventhat
time andlyfor
only differentcan
parameter distances
influence lzz,
occur
directlysimilarly to the distributions
on the kinetics of C and Sinfound
of moisture absorption in the conditions.
the studied analysis of the influence
However, for lof lx variation
z values lower
(Figures
than the 15–20),
0.01 m, in theother
effectwords,
of this there are no significant
geometrical parameter must variations
appearin more
these clearly.
parameters changing the
l distance in the z direction.
It is noteworthy that theAdditionally,
distributions one must
of free andremember
entrappedthat, despite
solute the probabilities
concentrations, for the of λ and
different
µ to be analyzed
planes (x = Rof
equal, variation x / 2, y =
entrapped R y / solute
2, z = R z /
concentration
2), at a occurs
given more
time andslowly
for comparing
different with
distances the
lz ,
free
occursolute concentration.
similarly to the distributions of C and S found in the analysis of the influence of lx variation
Thus,
(Figures it can in
15–20), beother
said that onlythere
words, the lyyare
parameter directly
no significant influences
variations in the moisture
these absorption
parameters changing kinetics
the l
under
distance theinstudied conditions.
the z direction. The influence
Additionally, oneof this remember
must parameterthat,is notdespite
contemplated by the Fick
the probabilities of λmodel,
and µ
making the Langmuir
to be equal, variation ofmodel
entrappeda more complete
solute model since
concentration it can
occurs assimilate
more imperceptible
slowly comparing with effects
the freein
relation to the Fickian model (most used).
solute concentration.
Table
Thus, 5it summarizes theonly
can be said that casesthe treated for the directly
ly parameter analysisinfluences
of the effect of distance
the moisture lxx, lyy, andkinetics
absorption lzz and
their
underrespective
the studied variations
conditions.in theTheabsorption
influence parameters
of this parameterof the isstudied model.
not contemplated by the Fick model,
making the Langmuir model a more complete model since it can assimilate imperceptible effects in
Table
relation to 5.the
Langmuir
Fickianabsorption
model (most parameter
used). values found for t = 250 h and different lxx, lyy, and lzz (arbitrary
cases).

Case t (h) Water Composite

Polymers 2019, 11, 1847 27 of 30

Table 5 summarizes the cases treated for the analysis of the effect of distance lx , ly , and lz and their
respective variations in the absorption parameters of the studied model.

Table 5. Langmuir absorption parameter values found for t = 250 h and different lx , ly , and lz (arbitrary
cases).

Water Composite
− − −
Case t (h) C S M
Te (◦ C) To (◦ C) lx (m) lz (m) ly (m)
(kg/kg) (kg/kg) (kg/kg)
8 250 50 25 0.01 0.1 0.01 0.088246 0.042096 0.130342
9 250 50 25 0.1 0.1 0.01 0.088769 0.042282 0.131051
10 250 50 25 1 0.1 0.01 0.088826 0.042302 0.131128
11 250 50 25 0.1 0.1 0.001 0.055515 0.028919 0.084434
12 250 50 25 0.1 0.1 0.1 0.095215 0.044581 0.139796
13 250 50 25 0.1 0.01 0.01 0.088468 0.042255 0.130723
14 250 50 25 0.1 1 0.01 0.088666 0.042559 0.131225

In general, it can be concluded that the distances of the sample to the container wall in the x, y, and
z directions influence the variations of the average concentration of free water, average concentration
of entrapped water, and average moisture content. As can be seen in Table 5, the larger the dimension
− − −
of lx , ly , or lz , the higher the values of C, S, and M found. Moreover, it was verified, under the studied
conditions, that variations in ly promoted more relevant changes in the analyzed parameters, as can be
observed in cases 9, 11, and 12, confirming the decrease of moisture absorption with the decrease of
distance l. These effects may be related to the dimensions of the solid or to the relationship between
the distance of the sample to the container wall in a given direction and the surface area of the solid
perpendicular to that direction.
From the analysis of the results obtained, in general, it is possible to verify that in the regions near
the surface, water absorption is faster because there is a greater area of direct contact with water. Water
penetrates inside the material, generating a higher concentration gradient along the thickness, which
decreases with increasing immersion time. For this reason, anywhere inside the solid, the moisture
content increases with time until achieving equilibrium condition, that is, saturation point.
For longer times ( t → ∞ ), or when there is a greater amount of free molecules within the material,
the number of entrapped molecules will be greater, and this occurs until equilibrium. This condition
occurs as ∂S ∂t
= 0, or yet, λC = µS. The effects caused due to long time exposure to moisture may
be irreversible due to the water molecules affinity with specific functional groups of the polymeric
matrices. Thus, understanding this process is crucial to predict the quality of the material under
wet environments.

4. Conclusions
In this paper a three-dimensional and transient mathematical modeling and its numerical solution
(finite volume method) to describe the water absorption in polymer composites reinforced with
vegetable fibers using the Langmuir model was developed. The results obtained from the average
moisture content of polymer composites reinforced with sisal fiber were validated with experimental
and analytical data reported in the literature, which shows the physical coherence in the development
of the proposed study.
From the results, it can be concluded that:

1. For small process time, the water absorption rate is fast, and it decreases for longer process time.
Polymers 2019, 11, 1847 28 of 30

2. The gradients of water molecules concentration (free and entrapped) are larger near the surface
of the material, having a flux from the surface to the center of the sample, especially at the vertex
of the composite;
3. The higher the concentration of free solute, the higher the concentration of solute entrapped
inside the material;
4. In the distributions of the analyzed parameters (free solute concentration, entrapped solute
concentration, and total moisture content), it was observed that the distances of the sample to the
container wall in the x, y, and z directions directly influence the kinetics and distribution of the
process parameters, with higher intensity in the y direction (smaller sample thickness);
5. The lower ly value, the lower the absorption rate of free and entrapped water molecules.

Author Contributions: Conceptualization, M.K.T.d.B., W.R.G.d.S.; formal analysis, M.K.T.d.B., W.R.G.d.S.;

investigation, M.K.T.d.B., W.R.G.d.S.; methodology, M.K.T.d.B., W.R.G.d.S.; project administration, G.L.d.O.N.,
A.G.B.d.L.; software, M.K.T.d.B., W.R.G.d.S.; validation, M.K.T.d.B., W.R.G.d.S.; writing—original draft, M.K.T.d.B.,
W.R.G.d.S., B.R.B.C.; writing—review and editing, R.A.d.Q., B.R.d.B.C. and F.V.d.S.T.
Funding: This research was funded by CNPq Grant (165907/2017-7).
Acknowledgments: The authors thank the CNPq, CAPES, FINEP (Brazilian Research Agencies), the Federal
University of Campina Grande, and the referenced authors who, with their researches, helped to improve
this work.
Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the
study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to
publish the results.

References
1. ASTM. ASTM D3878—Standard Terminology for Composite Materials; ASTM International: West Conshohocken,
PA, USA, 2007.
2. Leão, M.A. Licuri fibers: An Alternative Reinforcement of Polymeric Composites. Master’s Thesis, Federal
University of Rio Grande do Norte, Natal, Brazil, 2008. (In Portuguese).
3. Correia, E.A.S. Geopolymer Matrix Composites Reinforced with Vegetable Fibers and Sisal Pineapple.
Ph.D. Thesis, Federal University of Paraiba, João Pessoa, Brazil, 2011. (In Portuguese).
4. Bodros, E.; Baley, C. Study of Tensile Properties of stinging nettle fibers (Urtica dioica). Mater. Lett. 2007, 62,
2143–2145. [CrossRef]
5. Espert, A.; Vilaplana, F.; Karlsson, S. Comparison of water absorption in natural cellulosic fibers from
wood and one-year crops in polypropylene composites and its influence on their mechanical properties.
Compos. Part A Appl. Sci. Manuf. 2004, 35, 1267–1276. [CrossRef]
6. Marinelli, A.L.; Monteiro, M.R.; Ambrósio, J.D.; Branciforti, M.C.; Kobayashi, M. Development of polymer
composites with natural plant fibers of biodiversity: A contribution to the sustainability of the Amazon.
Polymers 2008, 18, 92–99. (In Portuguese)
7. Alomayri, T.; Assaedi, H.; Shaikh, F.U.A.; Low, I.M. Effect of water absorption on the mechanical properties
of cotton fabric-reinforced geopolymer composites. J. Asian Ceram. Soc. 2014, 2, 223–230. [CrossRef]
8. Faruk, O.; Bledzki, A.K.; Fink, H.S.; Sain, M. Biocomposites reinforced with natural fibers: 2000–2010.
Prog. Polym. Sci. 2012, 37, 1552–1596. [CrossRef]
9. Mohammed, L.; Ansari, M.N.; Pua, G.; Jawaid, M.; Islam, M.S. A review on natural fiber reinforced polymer
composite and its applications. Int. J. Polym. Sci. 2015, 2015, 243947. [CrossRef]
10. Vaddadi, P.; Nakamura, T.; Singh, R.P. Inverse analysis for transient moisture diffusion through fiber-reinforced
composites. Acta Mater. 2003, 51, 177–193. [CrossRef]
11. Yu, B.; Yang, J. Hygrothermal effects in composites. In Reference Module in Materials Science and Materials
Engineering; Elsevier: Amsterdam, The Netherlands, 2017.
12. Cruz, V.C.; Nóbrega, M.M.; Silva, W.P.; Carvalho, L.H.; Lima, A.G. An experimental study of water absorption
in polyester composites reinforced with macambira natural fiber. Mater. Werkst. 2011, 42, 979–984. [CrossRef]
Polymers 2019, 11, 1847 29 of 30

13. Nóbrega, M.M.; Cavalcanti, W.S.; Carvalho, L.H.; Lima, A.G. Water absorption in unsaturated polyester
composites reinforced with caroá fiber fabrics: Modeling and simulation. Mater. Werkst. 2010, 41, 300–305.
[CrossRef]
14. Cunha, J.A.P.; Costa, M.L.; Rezende, M.C. Influence of different hygrothermal conditions on tensile strength
of modified carbon/epoxy composites. Polymers 2006, 16, 193–201. (In Portuguese)
15. Carvalho, L.H.; Canedo, E.L.; Farias Neto, S.R.; Lima, A.G.B.; Silva, C.J. Moisture transport process in
vegetable fiber composites: Theory and analysis for technological applications, Industrial and Technological
Applications of Transport in Porous Materials. Adv. Struct. Mater. 2013, 36, 37–62.
16. Joliff, Y.; Belec, L.; Chailan, J.F. Modified water diffusion kinetics in an unidirectional glass/fiber composite
due to the interphase area: Experimental, analytical and numerical approach. Compos. Struct. 2013, 97,
296–303. [CrossRef]
17. Bond, D.A. Moisture Diffusion in a fiber-reinforced composite: Part I—Non-Fickian transport and the effect
of fiber spatial distribution. J. Compos. Mater. 2004, 39, 2113–2141. [CrossRef]
18. Thominette, F.; Gaudichet-Maurin, E.; Verdu, J. Effect of structure on water diffusion in hydrophilic polymers.
Defect Diffus. Forum 2006, 258–260, 442–446.
19. El Yagoubi, J.; Lubineau, G.; Roger, F.; Verdu, J. A fully coupled dissusion-reaction scheme for moisture
sorpition-desorption in an anhydride-cured epoxy resin. Polymer 2012, 53, 5582–5595. [CrossRef]
20. Simar, A.; Gigliotti, M.; Grandidier, J.C.; Ammar-Khodja, I. Evidence of thermo-oxidation phenomena
occurring during hygrothermal aging of thermosetting resins for RTM composite applications. Composites
Part A 2014, 66, 175–182. [CrossRef]
21. Simar, A.; Gigliotti, M.; Grandidier, J.C.; Ammar-Khodja, I. Decoupling of water and oxygen diffusion
phenomena in order to prove the occurrence of thermo-oxidation during hygrothermal aging of thermosetting
resins for ETM composite applications. Polymers 2018, 5, 11855–11872. [CrossRef]
22. El Yagoubi, J.; Lubineau, G.; Saghir, A.; Verdu, J.; Askari, A. Thermomechanical and hygroelastic properties
of an epoxy system under humid and cold-warm cycling conditions. Polym. Degrad. Stab. 2014, 99, 146–155.
[CrossRef]
23. Quino, G.; El Yagoubi, J.; Lubineau, G. Characterizing the toughness of an epoxy resin after wet aging using
compact tension specimens with non-uniform moisture content. Polym. Degrad. Stab. 2014, 109, 319–326.
[CrossRef]
24. Maggana, P.; Pissis, P. Water sorption and diffusion studies in an epoxy resin system. J. Polym. Sci. Part B
Polym. Phys. 1998, 37, 1165–1182. [CrossRef]
25. Glaskova, T.I.; Guedes, R.M.; Morais, J.J.; Aniskevich, A.N. A comparative analysis of moisture transport
models as applied to an epoxy binder. Mech. Compos. Mater. 2007, 43, 555–570. [CrossRef]
26. Carter, H.G.; Kibler, K.G. Langmuir-Type model for anomalous moisture diffusion in composite resins.
J. Compos. Mater. 1978, 12, 118–131. [CrossRef]
27. Peret, T.; Clement, A.; Freour, S.; Jacquemin, F. Numerical transient hygro-elastic analyses of reinforced
Fickian and non-Fickian polymers. Compos. Struct. 2014, 116, 395–403. [CrossRef]
28. Laplante, G.; Ouriadov, A.V.; Lee-Sullivan, P.; Balcom, B.J. Anomalous moisture diffusion in na epoxy
adhesive detected by magnetic resonance imaging. J. Appl. Polym. Sci. 2008, 109, 1350–1359. [CrossRef]
29. Eymard, R.; Gallouet, T.; Herbin, R. Finite volume methods. Handb. Numer. Anal. 2000, 7, 713–1018.
30. Dhakal, H.N.; Zhang, Z.Y.; Richardson, M.O.W. Effect of water absorption on the mechanical properties
of hemp fibre reinforced unsaturared polyester composites. Compos. Sci. Technol. 2007, 67, 1674–1683.
[CrossRef]
31. Grace, L.R.; Altan, M.C. Characterization of anisotropic moisture absorption in polymeric composites using
hindered diffusion model. Compos. Part A Appl. Sci. Manuf. 2012, 43, 1187–1196. [CrossRef]
32. Melo, R.Q.C.; Santos, W.R.G.; Lima, A.G.B.; Lima, W.M.P.B.; Silva, J.V.; Farias, R.P. Water absorption process
in polymer composites: Theory analysis and applications. In Transport Phenomena in Multiphase Systems, 1st
ed.; Delgado, J.M.P.Q., Lima, A.G.B., Eds.; Springer: Cham, Switzerland, 2018; Volume 93, pp. 219–349.
33. Bonniau, P.; Bussell, A.R. A comparative study of water absorption theories applied to glass epoxy composites.
J. Compos. Mater. 1981, 15, 272–293. [CrossRef]
34. Apicella, A.; Estiziano, L.; Nicolais, L.; Tucci, V. Environmental degradation of the electrical and thermal
properties of organic insulating materials. J. Mater. Sci. 1988, 23, 729–735. [CrossRef]
Polymers 2019, 11, 1847 30 of 30

35. Cotinaud, M.; Bonniau, P.; Bunsell, A.R. The effect of water absorption on the electrical properties of
glass-fibre reinforced epoxy composites. J. Mater. Sci. 1982, 17, 867–877. [CrossRef]
36. Santos, W.R.G.; Melo, R.Q.C.; Lima, A.G.B. Water Absorption in Polymer Composites Reinforced with
Vegetable Fiber Using Langmuir-Type Model: An Exact Mathematical Treatment. Defect Diffus. Forum 2017,
371, 102–110. [CrossRef]
37. Santos, D.G. Thermohydric Study and Mechanical Characterization of Polymer Matrix Composites Reinforced
with Vegetable Fiber: 3D Simulation and Experimentation. Ph.D. Thesis, Federal University of Campina
Grande, Campina Grande, Brazil, 2017. (In Portuguese).
38. Sanchez, E.M.; Cavani, C.S.; Leal, C.V.; Sanchez, C.G. Composition of unsaturated polyester resin with
sugarcane bagasse: Influence of fiber treatment on properties. Polymers 2010, 20, 194–200. (In Portuguese)

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).