Beruflich Dokumente
Kultur Dokumente
V
'(x) oV. (3)
The average strain 4 is dened similarly.
It is also convenient to dene volume-average stresses
and strains for the ber and matrix phases. To obtain
these, rst partition the averaging volume V into the
volume occupied by the bers V
l
and the volume occu-
pied by the matrix V
m
. We consider only two-phase
composites, so that
V = V
l
V
m
. (4)
The ber and matrix volume fractions are simply
v
l
= V
l
V and v
m
= V
m
V and, since only bers and
matrix are present, v
m
v
l
= 1.
The average ber and matrix stresses are the averages
over the corresponding volumes,
'
l
1
V
l
V
l
'(x) oV ano '
m
1
V
m
V
m
'(x) oV. (5)
The average strains for the ber and matrix are dened
similarly.
The relationships between the ber and matrix
averages and the overall averages can be derived from
the preceding denitions; they are:
' = v
l
'
l
v
m
'
m
(6)
4 = v
l
4
l
v
m
4
m
. (7)
An important related result is the average strain theo-
rem. Let the averaging volume V be subjected to surface
displacements u
0
(x) consistent with a uniform strain 4
0
.
Then the average strain within the region is
4 = 4
0
. (8)
This theorem is proved [2] by substituting the denition
of the strain tensor 4 in terms of the displacement vector
u into the denition of average strain 4, and applying
Gauss's theorem. The result is
c
ij
=
1
V
S
u
0
i
n
j
n
i
u
0
j
oS (9)
where S denotes the surface of V and n is a unit vector
normal to oS. The average strain within a volume V is
completely determined by the displacements on the sur-
face of the volume, so displacements consistent with a
uniform strain must produce the identical value of
average strain. A corollary of this principle is that, if we
dene a perturbation strain 4
C
(x) as the dierence
between the local strain and the average,
4
C
(x) 4(x) 4 (10)
then the volume-average of 4
C
(x) must equal zero
4
C
=
1
V
V
4
C
(x) oV = 0 (11)
The corresponding theorem for average stress also
holds. Thus, if surface tractions consistent with a '
0
are
exerted on S then the average stress is
' = '
0
. (12)
2.3. Average properties and strain concentration
The goal of micromechanics models is to predict the
average elastic properties of the composite, but even
these need careful denition. Here we follow the direct
approach [3]. Subject the representative volume V to
656 C.L. Tucker III, E. Liang / Composites Science and Technology 59 (1999) 655671
surface displacements consistent with a uniform strain
4
0
; the average stiness of the composite is the tensor C
that maps this uniform strain to the average stress.
Using Eq. (8) we have
' = C4. (13)
The average compliance S is dened in the same way,
applying tractions consistent with a uniform stress '
0
on the surface of the averaging volume. Then, using Eq.
(12),
4 = S'. (14)
It should be clear that S = C
1
. Other authors dene
the average stiness and compliance through the inte-
gral of the strain energy over V; this is equivalent to the
direct approach [2,4].
An important related concept, rst introduced by Hill
[2], is the idea of strain- and stress-concentration tensors
A and B. These are essentially the ratios between the
average ber strain (or stress) and the corresponding
average in the composite. More precisely:
4
l
= A4. (15)
'
l
= B'. (16)
A and B are fourth-order tensors and, in general, they
must be found from a solution of the microscopic stress
or strain elds. Dierent micromechanics models pro-
vide dierent ways to approximate A or B. Note that A
and B have both the minor symmetries of a stiness or
compliance tensor, but lack the major symmetry. That
is,
A
ijkl
= A
jikl
= A
ijlk
. (17)
but in general,
A
ijkl
,= A
klij
. (18)
For later use it will be convenient to have an alternate
strain concentration tensor
A that relates the average
ber strain to the average matrix strain,
4
l
=
A4
m
. (19)
This is related to A by
A =
A (1 v
l
)I v
l
A
1
(20)
so the two forms are easily interchanged.
Using equations now in hand, one can express the
average composite stiness in terms of the strain-con-
centration tensor Aand the ber and matrix properties [2].
Combining Eqs. (1), (2), (6), (7), (13) and (15), one obtains
C = C
m
v
l
C
l
C
m
( )A. (21)
The dual equation for the compliance is
S = S
m
v
l
S
l
S
m
B. (22)
Eqs. (21) and (22) are not independent, since S = C [ [
1
.
Hence, the strain-concentration tensor A and the stress-
concentration tensor B are not independent either. The
choice of which one to use in any instance is a matter of
convenience.
To illustrate the use of the strain-concentration and
stress-concentration tensors, we note that the Voigt
average corresponds to the assumption that the ber
and the matrix both experience the same, uniform
strain. Then 4
l
= 4, A = I, and from Eq. (21) the com-
posite modulus is
C
Voigl
= C
m
v
l
C
l
C
m
= v
l
C
l
v
m
C
m
. (23)
Recall that the Voigt average is an upper bound on the
composite modulus. The Reuss average assumes that
the ber and matrix both experience the same, uniform
stress. This means that the stress-concentration tensor B
equals the unit tensor I, and from (22) the compliance is
S
Reuss
= S
m
v
l
S
l
S
m
= v
l
S
l
v
m
S
m
. (24)
This represents a lower bound on the stiness of the
composite.
3. Theories
3.1. Eshelby's equivalent inclusion
A fundamental result used in several dierent models
is Eshelby's equivalent inclusion [5,6]. Eshelby solved for
the elastic stress eld in and around an ellipsoidal par-
ticle in an innite matrix. By letting the particle be a
prolate ellipsoid of revolution, one can use Eshelby's
result to model the stress and strain elds around a
cylindrical ber.
Eshelby rst posed and solved a dierent problem,
that of a homogeneous inclusion (Fig. 1). Consider an
innite solid body with stiness C
m
that is initially
stress-free. All subsequent strains will be measured from
this state. A particular small region of the body will be
called the inclusion, and the rest of the body will be
called the matrix. Suppose that the inclusion undergoes
Fig. 1. Eshelby's inclusion problem. Starting from the stress-free state
(a), the inclusion undergoes a stress-free transformation strain 4
T
(b).
Fitting the inclusion and matrix back together (c) produces the strain
state 4
C
(x) in both the inclusion and the matrix.
C.L. Tucker III, E. Liang / Composites Science and Technology 59 (1999) 655671 657
some type of transformation such that, if it were a
separate body, it would acquire a uniform strain 4
T
with
no surface traction or stress. 4
T
is called the transfor-
mation strain, or the eigenstrain. This strain might be
acquired through a phase transformation, or by a com-
bination of a temperature change and a dierent ther-
mal expansion coecient in the inclusion. In fact the
inclusion is bonded to the matrix, so when the trans-
formation occurs the whole body develops some com-
plicated strain eld 4
C
(x) relative to its shape before the
transformation. Within the matrix the stress '
m
is sim-
ply the stiness times this strain,
'
m
(x) = C
m
4
C
(x) (25)
but within the inclusion the transformation strain does
not contribute to the stress, so the inclusion stress is
'
!
= C
m
4
C
4
T
. (26)
The key result of Eshelby was to show that within an
ellipsoidal inclusion the strain 4
C
is uniform, and is
related to the transformation strain by
4
C
= E4
T
. (27)
E is called Eshelby's tensor, and it depends only on the
inclusion aspect ratio and the matrix elastic constants.
A detailed derivation and applications are given by
Mura [7], and analytical expressions for Eshelby's ten-
sor for an ellipsoid of revolution in an isotropic matrix
appear in many papers [812]. The strain eld 4
C
(x) in
the matrix is highly non-uniform [13], but this more
complicated part of the solution can often be ignored.
The second step in Eshelby's approach is to demon-
strate an equivalence between the homogeneous inclu-
sion problem and an inhomogeneous inclusion of the
same shape. Consider two innite bodies of matrix, as
shown in Fig. 2. One has a homogeneous inclusion with
some transformation strain 4
T
; the other has an inclu-
sion with a dierent stiness C
l
, but no transformation
strain. Subject both bodies to a uniform applied strain
4
A
at innity. We wish to nd the transformation strain
4
T
that gives the two problems the same stress and
strain distributions.
For the rst problem the inclusion stress is just Eq.
(26) with the applied strain added,
'
!
= C
m
4
A
4
C
4
T
(28)
while the second problem has no 4
T
but a dierent
stiness, giving a stress of
'
!
= C
l
4
A
4
C
. (29)
Equating these two expressions gives the transformation
strain that makes the two problems equivalent. Using
Eq. (27) and some rearrangement, the result is
C
m
C
l
C
m
E
4
T
= C
l
C
m
4
A
. (30)
Note that 4
T
is proportional to 4
A
, which makes the
stress in the equivalent inhomogeneity proportional to
the applied strain.
3.2. Dilute Eshelby model
One can use Eshelby's result to nd the stiness of a
composite with ellipsoidal bers at dilute concentra-
tions. Recall from Eq. (21) that to nd the stiness one
only has to nd the strain-concentration tensor A. To
do this, rst note that for a dilute composite the average
strain is identical to the applied strain,
4 = 4
A
. (31)
since this is the strain at innity. Also, from Eshelby,
the ber strain is uniform, and is given by
4
l
= 4
A
4
C
. (32)
where the right-hand side is evaluated within the ber.
Now write the equivalence between the stresses in the
homogeneous and the inhomogeneous inclusions, Eqs.
(28) and (29),
C
l
4
A
4
C
= C
m
4
A
4
C
4
T
. (33)
then use Eqs. (27), (31) and (32) to eliminate 4
T
, 4
A
and
4
C
from this equation, giving
I ES
m
C
l
C
m
4
l
= 4. (34)
Comparing this to Eq. (15) shows that the strain-con-
centration tensor for Eshelby's equivalent inclusion is
A
!slelLy
= I ES
m
C
l
C
m
1
. (35)
This can be used in Eq. (21) to predict the moduli of
aligned-ber composites, a result rst developed by
Russel [14]. Calculations using this model to explore the
eects of particle aspect ratio on stiness are presented
by Chow [15].
While Eshelby's solution treats only ellipsoidal bers,
the bers in most short-ber composites are much better
approximated as right circular cylinders. The relation-
ship between ellipsoidal and cylindrical particles was
considered by Steif and Hoysan [16], who developed a
very accurate nite element technique for determining
Fig. 2. Eshelby's equivalent inclusion problem. The inclusion (a) with
transformation strain 4
T
has the same stress '
!
and strain as the in-
homogeneity (b) when both bodies are subject to a far-eld strain 4
A
.
658 C.L. Tucker III, E. Liang / Composites Science and Technology 59 (1999) 655671
the stiening eect of a single ber of given shape. For
very short particles, d = 4, they found reasonable
agreement for E
11
by letting the cylinder and the ellipsoid
have the same d. The ellipsoidal particle gave a slightly
stier composite, with the dierence between the two
results increasing as the modulus ratio E
l
E
m
increased.
Henceforth we will use the cylinder aspect ratio in place
of the ellipsoid aspect ratio in Eshelby-type models.
Because Eshelby's solution only applies to a single
particle surrounded by an innite matrix, A
!slelLy
is
independent of ber volume fraction and the stiness
predicted by this model increases linearly with ber
volume fraction. Modulus predictions based on Eqs.
(35) and (21) should be accurate only at low volume
fractions, say up to v
l
of 1%. The more dicult pro-
blem is to nd some way to include interactions between
bers in the model, and so produce accurate results at
higher volume fractions. We next consider approaches
for doing that.
3.3. MoriTanaka model
A family of models for non-dilute composite materi-
als has evolved from a proposal originally made by
Mori and Tanaka [17]. Benveniste [18] has provided a
particularly simple and clear explanation of the Mori
Tanaka approach, which we use here to introduce the
approach.
We have already introduced the strain-concentration
tensor in Eq. (15). Suppose that a composite is to be
made of a certain type of reinforcing particle, and that,
for a single particle in an innite matrix, we know the
dilute strain-concentration tensor A
!slelLy
,
4
l
= A
!slelLy
4. (36)
The MoriTanaka assumption is that, when many
identical particles are introduced in the composite, the
average ber strain is given by
4
l
= A
!slelLy
4
m
. (37)
That is, within a concentrated composite each particle
`sees' a far-eld strain equal to the average strain in the
matrix. Using the alternate strain concentrator dened
in Eq. (19), the MoriTanaka assumption can be re-
stated as
A
MT
= A
!slelLy
(38)
Eq. (20) then gives the MoriTanaka strain con-
centrator as
A
MT
= A
!slelLy
(1 v
l
)I v
l
A
!slelLy
1
. (39)
This is the basic equation for implementing a Mori
Tanaka model.
The MoriTanaka approach for modeling composites
was rst introduced by Wakashima, Otsuka and Ume-
kawa [19] for modeling thermal expansions of composites
with aligned ellipsoidal inclusions. (Mori and Tanaka's
paper [17] treats only the homogeneous inclusion pro-
blem, and says nothing about composites). Mori
Tanaka predictions for the longitudinal modulus of a
short-ber composite were rst developed by Taya and
Mura [8] and Taya and Chou [9], whose work also
included the eects of cracks and of a second type of
reinforcement. Weng [20] generalized their method,
and Tandon and Weng [11] used the MoriTanaka
approach to develop equations for the complete set of
elastic constants of a short-ber composite. Tandon and
Weng's equations for the plane-strain bulk modulus k
23
and the major Poisson ratio v
12
must be solved iteratively.
However, this iteration can be avoided by using an alter-
ate formula for v
12
; details are given in Appendix A.
The usual development of the MoriTanaka model
[811] diers somewhat from Benveniste's explanation.
For an average applied stress ', the reference strain 4
0
is dened as the strain in a homogeneous body of matrix
at this stress,
' = C
m
4
0
. (40)
Within the composite the average matrix strain diers
from the reference strain by some perturbation 4
m
,
4
m
= 4
0
4
m
(41)
A ber in the composite will have an additional strain
perturbation 4
l
, such that
4
l
= 4
0
4
m
4
l
(42)
while the equivalent inclusion will have this strain plus
the transformation strain 4
T
. The stress equivalence
between the inclusion and the ber then becomes
C
l
4
0
4
m
4
l
= C
m
4
0
4
m
4
l
4
T
. (43)
Compare this to the dilute version, Eq. (33), noting that
4
A
in the dilute problem is equivalent to (4
0
4
m
) here.
The development is completed by assuming that the
extra ber perturbation is related to the transformation
strain by Eshelby's tensor,
4
l
= E4
T
(44)
Combining this with Eqs. (41) and (42) reveals that Eq.
(44) contains the essential MoriTanaka assumption:
the ber in a concentrated composite sees the average
strain of the matrix.
Some other micromechanics models are equivalent to
the MoriTanaka approach, though this equivalence
has not always been recognized. Chow [21] considered
Eshelby's inclusion problem and conjectured that in a
concentrated composite the inclusion strain would be
the sum of two terms: the dilute result given by Eshelby
(27) and the average strain in the matrix.
(4
C
)
l
= E4
T
(4
C
)
m
(45)
C.L. Tucker III, E. Liang / Composites Science and Technology 59 (1999) 655671 659
This can be combined with the denition of the average
strain from Eq. (7) to relate the inclusion strain (4
C
)
l
to
the transformation strain 4
T
:
(4
C
)
l
= (1 v
l
)E4
T
(46)
Chow then extended this result to an inhomogeneity
following the usual arguments, Eqs. (28)(35). This
produces a strain-concentration tensor
A
Clow
= I (1 v
l
)ES
m
C
l
C
m
1
(47)
which is equivalent to the MoriTanaka result (39).
Chow was apparently unaware of the connection between
his approach and the MoriTanaka scheme, but he seems
to have been the rst to apply the MoriTanaka approach
to predict the stiness of short-ber composites.
A more recent development is the equivalent poly-in-
clusion model of Ferrari [22]. Rather than use the strain-
concentration tensor A, Ferrari used an eective Eshelby
tensor
E, dened as the tensor that relates inclusion
strain to transformation strain at nite volume fraction:
(4
C
)
l
=
E4
T
. (48)
Once
E has been dened, it is straightforward to derive a
strain-concentration tensor A and a composite modulus.
Ferrari considered admissible forms for
E, given the
requirements that
E must (a) produce a symmetric sti-
ness tensor C, (b) approach Eshelby's tensor E as
volume fraction approaches zero, and (c) give a com-
posite stiness that is independent of the matrix stiness
as volume fraction approaches unity. He proposed a
simple form that satises these criteria,
E = (1 v
l
)E. (49)
The combination of Eqs. (48) and (49) is identical to
Chow's assumption (46) and, for aligned bers of uni-
form length, Ferrari's equivalent poly-inclusion model,
Chow's model, and the MoriTanaka model are identical.
Important dierences between the equivalent poly-
inclusion model and the MoriTanaka model arise when
the bers are misoriented or have dierent lengths, a
topic that will be addressed in a subsequent paper [1].
3.4. Self-consistent models
A second approach to account for nite ber volume
fraction is the self-consistent method. This approach is
generally credited to Hill [23] and Budiansky [24], whose
original work focused on spherical particles and con-
tinuous, aligned bers. The application to short-ber
composites was developed by Laws and McLaughlin
[25] and by Chou, Nomura and Taya [26].
In the self-consistent scheme one nds the properties
of a composite in which a single particle is embedded in
an innite matrix that has the average properties of the
composite. For this reason, self-consistent models are
also called embedding models.
Again building on Eshelby's result for a ellipsoidal
particle, we can create a self-consistent version of Eq.
(35) by replacing the matrix stiness and compliance
tensors by the corresponding properties of the compo-
site. This gives the self-consistent strain-concentration
tensor as
A
SC
= I ES C
l
C
1
. (50)
Of course the properties C and S of the embedding
`matrix' are initially unknown. When the reinforcing
particle is a sphere or an innite cylinder, the equations
can be manipulated algebraically to nd explicit
expressions for the overall properties [23,24]. For short
bers this has not proved possible, but numerical solu-
tions are easily obtained by an iterative scheme. One
starts with an initial guess at the composite properties,
evaluates E and then A
SC
from Eq. (50), and substitutes
the result into Eq. (21) to get an improved value for the
composite stiness. The procedure is repeated using this
new value, and the iterations continue until the results
for C converge.
An additional, but less obvious, change is that Eshel-
by's tensor E depends on the `matrix' properties, which
are now transversely isotropic. Expressions for Eshelby's
tensor for an ellipsoid of revolution in a transversely
isotropic matrix are given by Chou, Nomura and Taya
[27] and by Lin and Mura [28]. With these expressions
in hand one can use Eq. (50) together with Eq. (21) to
nd the stiness of the composite. This is the self-con-
sistent approach used for short-ber composites [25,26].
A closely-related approach, called the `generalized
self-consistent model,' also uses an embedding
approach. However, in these models the embedded
object comprises both ber and matrix material. When
the composite has spherical reinforcing particles, the
embedded object is a sphere of the reinforcement
encased in a concentric spherical shell of matrix; this is
in turn surrounded by an innite body with the average
composite properties. The generalized self-consistent
model is sometimes referred to as a `double embedding'
approach. For continuous bers the embedded object is
a cylindrical ber surrounded by a cylindrical shell of
matrix. The rst generalized self-consistent models were
developed for spherical particles by Kerner [29], and for
cylindrical bers by Hermans [30]. Both of these papers
contain an error, which is discussed and corrected by
Christensen and Lo [31]. While the generalized self-
consistent model is widely regarded as superior to the
original self-consistent approach, no such model has
been developed for short bers.
3.5. Bounding models
A rather dierent approach to modeling stiness is
based on nding upper and lower bounds for the com-
posite moduli. All bounding methods are based on
660 C.L. Tucker III, E. Liang / Composites Science and Technology 59 (1999) 655671
assuming an approximate eld for either the stress or
the strain in the composite. The unknown eld is then
found through a variational principle, by minimizing or
maximizing some functional of the stress and strain.
The resulting composite stiness is not exact, but it can
be guaranteed to be either greater than or less than the
actual stiness, depending on the variational principle.
This rigorous bounding property is the attraction of
bounding methods.
Historically, the Voigt and Reuss averages were the
rst models to be recognized as providing rigorous
upper and lower bounds [32]. To derive the Voigt
model, Eq. (23), one assumes that the ber and matrix
have the same uniform strain, and then minimizes the
potential energy. Since the potential energy will have an
absolute minimum when the entire composite is in
equilibrium, the potential energy under the uniform
strain assumption must be greater than or equal to the
exact result, and the calculated stiness will be an upper
bound on the actual stiness. The Reuss model, Eq.
(24), is derived by assuming that the ber and matrix
have the same uniform stress, and then maximizing the
complementary energy. Since the complementary energy
must be maximum at equilibrium, the model provides a
lower bound on the composite stiness. Detailed deri-
vations of these bounds are provided by Wu and
McCullough [33].
The Voigt and Reuss bounds provide isotropic results
(provided the ber and matrix are themselves isotropic),
when in fact we expect aligned-ber composites to be
highly anisotropic. More importantly, when the ber
and matrix have substantially dierent stinesses then
the Voigt and Reuss bounds are quite far apart, and
provide little useful information about the actual com-
posite stiness. This latter point motivated Hashin and
Shtrikman to develop a way to construct tighter
bounds.
Hashin and Shtrikman developed an alternate varia-
tional principle for heterogeneous materials [34,35].
Their method introduces a reference material, and bases
the subsequent development on the dierences between
this reference material and the actual composite. Rather
than requiring two variational principles, like the Voigt
and Reuss bounds, their single variational principle
gives both the upper and lower bounds by making
appropriate choices of the reference material. For an
upper bound the reference material must be as sti or
stier than any phase in the composite (ber or matrix),
and for a lower bound the reference material must have
a stiness less than or equal to any phase. In most
composites the ber is stier than the matrix, so choos-
ing the ber as the reference material gives an upper
bound and choosing the matrix as the reference material
gives a lower bound. If the matrix is stier than the
ber, the bounds are reversed. The resulting bounds are
tighter than the Voigt and Reuss bounds, which can be
obtained fromthe HashinShtrikman theory by giving the
reference material innite or zero stiness, respectively.
Hashin and Shtrikman's original bounds [35] apply to
isotropic composites with isotropic constituents. Fre-
quently the bounds are regarded as applying to compo-
sites with spherical particles, though a ber composite
with 3-D random ber orientation must also obey the
bounds.
Walpole re-derived the HashinShtrikman bounds
using classical energy principles [36], and extended them
to anisotropic materials [37]. Walpole also derived
results for innitely long bers and innitely thin disks
in both aligned and 3-D random orientations [38].
The HashinShtrikmanWalpole bounds were exten-
ded to short-ber composites by Willis [39] and by Wu
and McCullough [33]. These workers introduced a two-
point correlation function into the bounding scheme,
allowing aligned ellipsoidal particles to be treated.
Based on these extensions, explicit formulae for aligned
ellipsoids were developed by Weng [40] and by Eduljee
et al. [41,42].
The general bounding formula, shown here in the
format developed by Weng, gives the composite stiness
C as
C = v
l
C
l
Q
l
v
m
C
m
Q
m
v
l
Q
l
v
m
Q
m
1
. (51)
where the tensors Q
l
and Q
m
are dened as
Q
l
= I E
0
S
0
(C
l
C
0
)
1
ano
Q
m
= I E
0
S
0
(C
m
C
0
)
1
.
(52)
Here E
0
is Eshelby's tensor associated with the proper-
ties of the reference material, which has stiness C
0
and
compliance S
0
.
When the matrix is chosen as the reference material,
Eq. (51) gives a strain concentrator of
A
lowei
= I E
m
S
m
(C
l
C
m
)
1
(53)
This result is labeled here as the lower bound, on the
presumption that the ber is stier than the matrix. The
composite stiness is found by substituting
A
lowei
into
Eqs. (20) and (21). Eduljee and McCullough [41,42]
argue that the lower bound provides the most accurate
estimate of composite properties, and recommend it as a
model. Note that this lower bound prediction is iden-
tical to the MoriTanaka model, Eq. (39) [20,40]. This
correspondence lends theoretical support to the Mori
Tanaka approach, and guarantees that it will always
obey the bounds.
The other bound, found by using Eq. (51) with the
ber as the reference material, has a strain concentrator
of
A
uei
= I E
l
S
l
(C
m
C
l
)
. (54)
Note that the Eshelby tensor E
l
is now computed for
inclusions of matrix material surrounded by the ber
C.L. Tucker III, E. Liang / Composites Science and Technology 59 (1999) 655671 661
material. Eq. (54) is labeled as the upper bound, pre-
suming that the ber is stier than the matrix. An iden-
tical result can be obtained from the MoriTanaka
theory by assuming that ellipsoidal particles of the
matrix material are embedded in a continuous phase of
the ber material.
If the matrix is stier than the bers, then the right-
hand sides of Eqs. (53) and (54) are unchanged but Eq.
(53) becomes the upper bound and Eq. (54) becomes the
lower bound. All of the preceding bounding formulae
have been given for two-component composites, but the
theory readily accommodates multiple reinforcements.
At ber volume fractions close to unity, the matrix
stiness strongly inuences the composite stiness for
the lower bound/MoriTanaka models, despite the tiny
amount of it that is present. Packing considerations
suggest that the only way to approach such high volume
fractions is for the ber phase to become continuous,
and Lielens et al. [43] suggest that at very high ber
volume fractions the composite stiness should be much
closer to the upper bound, or equivalently to the Mori
Tanaka prediction using the ber as the continuous
phase. This insight prompted Lielens and co-workers to
propose a model that interpolates between the upper
and lower bounds, such that the lower bound dominates
at low volume fractions and the upper bound dominates
at high volume fractions (again presuming the ber is
the stier phase). They perform this interpolation on the
inverse of the strain-concentration tensor
A, producing
the predictive equation [43]
A
Iielens
= (1 f)|
A
lowei
]
1
f|
A
uei
]
1
1
. (55)
The interpolating factor f depends on ber volume
fraction, and they propose
f =
v
l
v
2
l
2
. (56)
This theory reproduces the lower bound and Mori
Tanaka results at low volume fractions, but is said to
give improved results at reinforcement volume fractions
in the 4060% range.
3.6. HalpinTsai equations
The HalpinTsai equations [44,45] have long been
popular for predicting the properties of short-ber
composites. A detailed review and derivation is pro-
vided by Halpin and Kardos [46], from which we sum-
marize the main points.
The HalpinTsai equations were originally developed
with continuous-ber composites in mind, and were
derived from the work of Hermans [30] and Hill [47].
Hermans developed the rst generalized self-consistent
model for a composite with continuous aligned bers
(see Section 3.4). Halpin and Tsai found that three of
Hermans' equations for stiness could be expressed in a
common form
P
P
m
=
1 rnv
l
1 nv
l
will n =
(P
l
P
m
) 1
(P
l
P
m
) 1
. (57)
Here P represents any one of the composite moduli lis-
ted in Table 1, and P
l
and P
m
are the corresponding
moduli of the bers and matrix, while r is a parameter
that depends on the matrix Poisson ratio and on the
particular elastic property being considered. Hermans
derived expressions for the plane-strain bulk modulus
k
23
, and for the longitudinal and transverse shear mod-
uli G
12
and G
23
. The r parameters for these properties
are given in Table 1. Note that for an isotropic matrix
k
m
= (E
m
2(1 v
m
)(1 2v
m
)).
Hill [47] showed that for a continuous, aligned-ber
composite the remaining stiness parameters are given
by
E
11
= v
l
E
l
v
m
E
m
4
v
l
v
m
1
k
l
1
k
m
2
1
k
23
v
l
k
l
v
m
k
m
. (58)
v
12
= v
l
v
l
v
m
v
m
v
l
v
m
1
k
l
1
k
m
1
k
23
v
l
k
l
v
m
k
m
. (59)
This completes Hermans' model for aligned-ber com-
posites; note that one must know k
23
to nd E
11
and v
12
.
We now know that Hermans' result for G
23
is incorrect,
in that it does not satisfy all of the ber/matrix continuity
conditions [3]. It is, however, identical to a lower bound
Table 1
Correspondence between HalpinTsai Equation (57) and generalized self-consistent predictions of Hermans [30] and Kerner [29]. After Halpin and
Kardos [46]
P P
l
P
m
r Comments
k
23
k
l
k
m
1vm2v
2
m
1vm
Plane strain bulk modulus, aligned bers
G
23
G
l
G
m
1vm
3vm4v
2
m
Transverse shear modulus, aligned bers
G
12
G
l
G
m
1 Longitudinal shear modulus, aligned bers
K K
l
K
m
2(12vm)
1vm
Bulk modulus, particulates
G G
l
G
m
75vm
810vm
Shear modulus, particulates
662 C.L. Tucker III, E. Liang / Composites Science and Technology 59 (1999) 655671
on G
23
derived by Hashin [48]. Hermans' remaining
results are identical to Hashin and Rosen's composite
cylinders assemblage model [49], so Hermans' k
23
, and
thus his E
11
and v
12
, are identical to the self-consistent
results of Hill [23].
The HalpinTsai form (57) can also be used to
express equations for particulate composites derived by
Kerner [29], who also used a generalized self-consistent
model. Table 1 gives the details. Kerner's result for
shear modulus G is also known to be incorrect, but
reproduces the HashinShtrikmanWalpole lower
bound for isotropic composites, while Kerner's result
for bulk modulus K is identical to Hashin's composite
spheres assemblage model [50]. See Christensen and Lo
[31] and Hashin [3] for further discussion of Kerner's
and Hermans' results.
To transform these results into convenient forms for
continuous-ber composites, Halpin and Tsai made
three additional ad hoc approximations:
. Eq. (57) can be used directly to calculate selected
engineering constants, with E
11
or E
22
replacing P.
. The r parameters in Table 1 are insensitive to v
m
,
and can be approximated by constant values.
. The underlined terms in Eqs. (58) and (59) can be
neglected.
In Eq. (58) the underlined term is typically negligible,
and dropping it gives the familiar rule of mixtures for
E
11
of a continuous-ber composite. However, drop-
ping the underlined term in Eq. (59) and using a rule of
mixtures for v
12
is not necessarily accurate if the ber
and matrix Poisson ratios dier. Halpin and Tsai argue
for this latter approximation on the grounds that lami-
nate stinesses are insensitive to v
12
.
In adapting their approach to short-ber composites,
Halpin and Tsai noted that r must lie between 0 and .
If r = 0 then Eq. (57) reduces to the inverse rule of
mixtures [46],
1
P
=
v
l
P
l
v
m
P
m
(60)
while for r = the HalpinTsai form becomes the rule
of mixtures,
P = v
l
P
l
v
m
P
m
. (61)
Halpin and Tsai suggested that r was correlated with
the geometry of the reinforcement and, when calculating
E
11
, it should vary from some small value to innity as a
function of the ber aspect ratio d. By comparing
model predictions with available 2-D nite element
results, they found that r = 2(d) gave good predic-
tions for E
11
of short-ber systems. Also, they suggested
that other engineering constants of short-ber compo-
sites were only weakly dependent on ber aspect ratio,
and could be approximated using the continuous-ber
formulae [45]. The resulting equations are summarized
in Table 2. The early references [44,45] do not mention
G
23
. When this property is needed the usual approach is
to use the r value given in Table 1. While the Halpin
Tsai equations have been widely used for isotropic ber
materials, the underlying results of Hermans and Hill
apply to transversely isotropic bers, so the Halpin
Tsai equations can also be used in this case.
The HalpinTsai equations are known to t some
data very well at low volume fractions, but to under-
predict some stinesses at high volume fractions. This
has prompted some modications to their model.
Hewitt and de Malherbe [51] proposed making r a
function of v
l
, and by curve tting found that
r = 1 40v
10
l
. (62)
gave good agreement with 2-D nite element results for
G
12
of continuous ber composites.
Nielsen and Lewis [52,53] focused on the analogy
between the stiness G of a composite and the viscosity
j of a suspension of rigid particles in a Newtonian uid,
noting that one should nd jj
m
= GG
m
when the
reinforcement is rigid (G
l
G
m
) and the matrix is
incompressible. They developed an equation in which
the stiness not only matches dilute theory at low
volume fractions, but also displays GG
m
as v
l
approaches a packing limit v
lmax
. This leads to a mod-
ied HalpinTsai form
P
P
m
=
1 rnv
l
1 [(v
l
)nv
l
(63)
with n retaining its denition from Eq. (57). Here the
function [(v
l
) contains the maximum volume fraction
v
l max
as a parameter. [ is chosen to give the proper
behavior at the upper and lower volume fraction limits,
which leads to forms such as
[(v
l
) = 1
1 v
l max
v
2
l max
v
l
. (64)
[(v
l
) =
1
v
l
1 ex
v
l
1 (v
l
v
l max
)
. (65)
The Nielsen and Lewis model improves on the Halpin
Tsai predictions, compared to experimental data for G
of particle-reinforced polymers [52] and to nite element
calculations for G
12
of continuous-ber composites [53],
using v
l max
values from 0.40 to 0.85.
Table 2
Traditional HalpinTsai parameters for short-ber composites, used in
Eq. (57)
P P
l
P
m
r Comments
E
11
E
l
E
m
2(d) Longitudinal modulus
E
22
E
l
E
m
2 Transverse modulus
G
12
G
l
G
m
1 Longitudinal shear modulus
v
12
Poisson ratio, = v
l
v
l
v
m
v
m
For G
23
see Table 1.
C.L. Tucker III, E. Liang / Composites Science and Technology 59 (1999) 655671 663
Recently Ingber and Papathanasiou [54] tested the
HalpinTsai equation and its modications against
boundary element calculations of E
11
for aligned short
bers. They found the Nielsen modication to be better
than the original HalpinTsai form. Hewitt and de
Malherbe's form could be adjusted to t data for any
single d, but was not useful for predictions over a
range of aspect ratios. These results are discussed fur-
ther in Section 4.
3.7. Shear lag models
Historically, shear lag models were the rst micro-
mechanics models for short-ber composites [55], as
well as the rst to examine behavior near the ends of
broken bers in a continuous-ber composite [56,57].
Despite some serious theoretical aws, shear lag models
have enjoyed enduring popularity, perhaps due to their
algebraic simplicity and their physical appeal.
Classical shear lag models only predict the long-
itudinal modulus E
11
, so they do not meet our criterion
of predicting a complete set of elastic constants. How-
ever, we include them here because of their historical
importance and their widespread use. One could obtain
a complete stiness model by using the shear lag pre-
diction for E
11
and some continuous-ber model (such
as Hermans') for the remaining elastic constants. If the
ber is anisotropic then its axial modulus should be
used in the shear lag equations.
Following Cox [55], the shear lag analysis focuses on
a single ber of length and radius r
l
, which is encased
in a concentric cylindrical shell of matrix having radius
R. The ber is aligned parallel to the z axis, as shown in
Fig. 3. Only the axial stress o
11
and axial strain c
11
are
of interest, and Poisson eects are neglected so that
o
l
11
= E
l
c
l
11
. The outer cylindrical surface of the matrix
is subjected to displacement boundary conditions con-
sistent with an average axial strain c
11
, and one solves
for the ber stress o
l
11
(z). (More rigorously, o
l
11
(z) is the
average stress over the ber cross-section at z.) Axial
equilibrium of the ber requires that
oo
l
11
oz
=
2t
rz
r
l
. (66)
where t
rz
is the axial shear stress at the ber surface.
The key assumption of shear lag theory is that t
rz
is
proportional to the dierence in displacement w
between the ber surface and the outer matrix surface:
t
rz
(z) =
H
2r
l
w(R. z) w(r
l
. z) | ]. (67)
where H is a constant that depends on matrix properties
and ber volume fraction. Solving Eq. (66) for o
l
11
(z)
and applying boundary conditions of zero stress at the
ber ends gives an average ber stress of
o
l
11
= E
l
c
11
1
lanl ([2)
([2)
(68)
with
[
2
=
H
r
2
l
E
l
. (69)
It is convenient to rewrite this as an expression for the
average ber strain,
c
l
11
= n
c
11
. (70)
where n
= 1
lanl ([2)
([2)
. (71)
Note that n
K
R
v
l
. (74)
where K
R
is a constant that depends on the assumption
used to nd R. Table 3 summarizes the choices for K
R
. Fig. 3. Idealized ber and matrix geometry used in shear lag models.
664 C.L. Tucker III, E. Liang / Composites Science and Technology 59 (1999) 655671
Cox [55] assumed a hexagonal packing, and chose R as
the distance between centers of nearest-neighbor bers
(Fig. 4(a)). It seems more realistic to let R equal half of
the distance between nearest neighbors (Fig. 4(b)), a
choice labeled `hexagonal' in Table 3. Rosen [56,57],
and later Carman and Reifsnider [58], chose r
2
l
R
2
= v
l
so that the concentric cylinder model in Fig. 3 would
have the same ber volume fraction as the composite.
This is the same R as the composite cylinders model of
Hashin and Rosen [49]. More recently, Robinson and
Robinson [59,60] assumed a square array of bers, and
chose R as half the distance between centers of nearest
neighbors (Fig. 4(c)) [61]. Each of these choices gives a
somewhat dierent dependence of n
on ber volume
fraction, with larger values of K
R
producing lower
values of E
11
.
Shear lag models are usually completed by combining
the average ber stress in Eq. (68) with an average
matrix stress to produce a modied rule of mixtures for
the axial modulus
E
11
= n
v
l
E
l
(1 v
l
)E
m
. (75)
However, the matrix stress in this formula is not con-
sistent with the basic concepts of average stress and
average strain. Note that Eq. (7) must hold for c
11
, as
for any other component of strain. Combining this with
Eq. (70) to nd the average matrix strain, and following
through to nd the composite stiness (with Poisson
eects neglected), gives a result that is consistent with
both the assumptions of shear lag theory and the basic
concepts of average stress and strain
E
11
= n
v
l
E
l
(1 n
v
l
)E
m
(76)
= E
m
v
l
(E
l
E
m
)n
.
This equation is an exact scalar analog of the general
tensorial stiness formula, eqn (21). For the cases in this
paper, the dierence between Eqs. (75) and (76) is small,
and we will use the classical shear lag result (75) when
testing the models.
A model by Fukuda and Kawata [62] for the axial
stiness of aligned short-ber composites is closely
related to shear lag theory. They begin with a 2-D elas-
ticity solution for the shear stress around a single slen-
der ber in an innite matrix. The usual shear lag
relation, Eq. (66), is used to transform this into an
equation for the ber stress distribution, which is then
approximated by a Fourier series. The coecients of a
truncated series are evaluated analytically using Galer-
kin's method. This is a dilute theory, in which modulus
varies linearly with ber volume fraction.
Like any shear lag theory, Fukuda and Kawata's
theory predicts that E
11
approaches the rule of mixtures
result as the ber aspect ratio approaches innity. But
for short bers Fukuda and Kawata's theory gives
much lower E
11
values than shear lag theory. In Fukuda
and Kawata's theory, the ratio of ber strain to matrix
strain is governed by the parameter (d)(E
m
E
l
). In
contrast, for shear lag theory, Eq. (71), the governing
parameter is [2, which is proportional to
(d)
E
m
E
l
=3.628
Composite cylinders 1=1.000
Hexagonal 2
=0.907
Square 4=0.785
Fig. 4. Fiber packing arrangements used to nd R in shear lag models.
(a) Cox. (b) Hexagonal. (c) Square.
C.L. Tucker III, E. Liang / Composites Science and Technology 59 (1999) 655671 665
results only for the square array, noting that this choice
for R gives the highest stiness. The models which are
not shown are: the dilute Eshelby model, which is lim-
ited to small volume fractions; the HashinShtrikman
Walpole lower bound, which is identical to the Mori
Tanaka model; and the upper bound, which is not
claimed to be useful by itself.
4.1. Finite element modeling
Using the nite element method we analyzed two
types of periodic, three-dimensional arrays of bers,
which we call regular and staggered arrays. The repre-
sentative volume elements (RVE's) are shown in Fig. 5.
The unit cell dimensions were chosen with b = [a,
where [ is a constant. We used both [ = 1 to obtain
square packing, and [ =