Critical Void Content for Thermoset

Composite Laminates

ZHAN-SHENG GUO,1,* LING LIU,1 BO-MING ZHANG2 AND SHANYI DU2

1
School of Aerospace Engineering and Applied Mechanics
Tongji University, Shanghai 200092, China
2
Center for Composite Materials, Harbin Institute of Technology
Harbin, 150001, China

(Received September 25, 2005)

(Accepted January 10, 2006)

ABSTRACT: An experimental program to characterize the effect of voids on the

strength of composite laminates is presented. The adequacy of a fracture criterion to
represent the experimental data for the effect of voids on the flexure strength, tensile
strength, and interlaminar shear strength of composite laminates is assessed. The
experimental program investigates the effect of different pressures and dwell times
on the critical void content. Laminates produced with carbon fiber/epoxy resin
unidirectional prepreg have been produced with an intentionally high void content.
Short beam shear, three-point flexure, and tensile testing are used for mechanical
evaluation and the results correlate to void volume fraction and ultrasonic
absorption coefficient. The ultrasonic absorption coefficient is measured for all
the specimens and its variation is approximately linear with the void content,
corroborating previous experimental results. The effects of these factors on the
strength of the composite laminates are discussed in terms of the fracture parameters
involved in the fracture criterion. The critical void content is estimated for each case
both in terms of void content and ultrasonic attenuation.

KEY WORDS: polymeric composites, voids, water absorption, mechanical strength,

failure criteria.

INTRODUCTION

COMMON PROBLEM in the manufacturing of polymer composites is the formation

A of defects such as resin-rich regions, crimped and distorted fibers, foreign inclusions,
and voids. Among those defects, voids are arguably the greatest problem because they
are difficult to avoid, particularly at the corners of composite components, and are
detrimental to the mechanical properties [1–11]. Voids are formed because of a number
of reasons, including the formation of bubbles from volatile byproducts produced during

*Author to whom correspondence should be addressed. E-mail: guowinboy@163.com

Figures 1, 3–6, 10 and 11 appear in color online: http://jcm.sagepub.com

Journal of COMPOSITE MATERIALS, Vol. 43, No. 17/2009 1775

0021-9983/09/17 1775–16 $10.00/0 DOI: 10.1177/0021998306065289
ß 2009 SAGE Publications
1776 Z.-S. GUO ET AL.

the cure reaction of the polymeric matrix, the use of a high-viscosity resin combined
with closely packed fibers that are not completely wetted by resin, the entrapment of air
in the material system, and fabrication mishaps such as a leaking vacuum bag or a poor
vacuum source [1–10].
Most aircraft composite parts are inspected after fabrication with nondestructive
techniques such as ultrasound either by pulse echo or through transmission. These
techniques are able to detect defects that cause ultrasonic attenuation such as voids,
delaminations, interlaminar cracks, inclusions, foreign object damage, resin-rich regions,
and others [3–5,9].
The behavior of a composite laminate with voids under different types of mechanical
loading has been widely studied. Most works consider the interlaminar shear strength
[3–5,9–15] but the interlaminar fracture toughness [12], bending strength under static and
fatigue loading [9], and compressive strength [3–5] have also been studied. However, there
is no general agreement over the magnitude of the effect of porosity on the mechanical
properties of composites. The difficulty lies in the large number of parameters involved
in the problem.
Different types of prepreg materials used in the manufacture of the laminates affect
the material toughness; the processing parameters (such as temperature and pressure) and
type of reinforcement affect the distribution, location, shape, and size of the voids in the
laminate.
All of these factors, in turn, produce different effects on the laminate strength. The type
of mechanical loading, its nature (static or fatigue), and inspection technique used are also
significant factors.
For example, the use of different frequencies in the ultrasound equipment results
in different values of attenuation. Therefore, only the results obtained from the
characterization of composites produced and tested in a similar way can directly be
compared. The experimental studies mentioned aim at correlating the void content to
the laminate strength for a specific type of loading. In those studies, it is implicitly
assumed that the void content is uniform at least over the critical section of the specimen.
However, in practice, voids are not uniformly distributed but are random in nature.
Note that void content is a measurement associated with a finite volume of material rather
than measurement to a point. The void content measurement by either matrix digestion
or ultrasonic inspection captures some sort of average value over a given volume,
without retaining the information on the shape, size, and distribution of the voids.
However, these features play an important role in determining the effect of voids on the
mechanical behavior of laminate and are primarily controlled by matrix material,
type of reinforcement, and manufacturing problem that originated the defects. The
important issue of the effect of the size of the area of the laminate affected by voids
is not addressed in this work. The voids are assumed to be uniformly distributed over
the laminate.
A fracture criterion that correlates fracture stress with void content or, alternatively,
to ultrasonic attenuation, is needed to establish an acceptance level for the inspection.
Establishing the acceptable level of defects is a critical issue in designing composite
structures. An overly conservative acceptance criterion causes many parts that could
perform satisfactorily to be unnecessarily discarded, increasing the manufacturing cost.
On the other hand, if the deleterious effects of defects are underestimated, in-service failure
of some parts may occur. Both situations are avoided by a judicious choice of acceptable
level of defects in the part. This should be based on reliable fracture criteria supported
Critical Void Content for Thermoset Composite Laminates 1777

by extensive experimental characterization and an in-depth understanding of the effect

of defects on the mechanical behavior of the laminate [3–5].
Almeida and Neto [9] proposed a fracture criterion that presented good correlation
with the experimental data on the bending strength of composite laminates with voids.
The same idea was successfully applied to predict the interlaminar shear strength by
Costa et al. [3–5], Jeong [7], and Almeida and Santacreu [8]. The fracture criterion
proposed by Almeida and Neto [9] is based on ultrasonic attenuation rather than directly
using the void content. The rationale for this is that, in practical applications, the
acceptable level of defects must be established in terms of a nondestructive technique.
However, laminate thickness, type of matrix, and reinforcement affect the shape
and size of the voids and, consequently, ultrasonic attenuation. Therefore, the
void content corresponding to a given ultrasonic attenuation level depends on those
factors; as a consequence, the fracture parameters obtained from the application of
the fracture criterion will reflect the dependence of the ultrasonic measurements on
those factors.
The purpose of this work is (1) to obtain an optimal cure cycle as well as to evaluate
the effects of pressure conditions on the void contents and mechanical properties,
(2) to present the results of an experimental program to investigate the effect of voids on
the mechanical strength of composite laminates, (3) to assess the adequacy of the fracture
criterion to represent the experimental data for the effect of voids on the flexure, tensile,
and shear strength of composite laminates, and (4) to discuss the effect of the type
of loading (flexure vs. interlaminar shear) on the critical void content of composite
laminates both in terms of volume fraction and ultrasonic attenuation. Moreover,
the fracture parameters are estimated by the fracture criterion.

EXPERIMENTAL PROCEDURE

Fabrication of the Specimens and Void Content Measurements

The material under research was T700/TDE85 carbon fiber reinforced epoxy prepreg.
The initial fiber volume fraction Vf is 60  2%.
The viscosity was measured on an improved NDJ-7 rheometer. The viscosity was
measured at the manufacturer’s temperature cure cycle. The temperature was increased
up to 120 C at a rate of 2 C/min and was held for 2 h. Then it was increased to 160 C
at the same rate and was held for 4 h . Finally, it was cooled down to room temperature.
Polymer composites with high void content were manufactured using a procedure based
on the rheological analyses of the wet prepregs [3–5]. Laminates presenting intentionally
high porosity levels were produced, combining the technique proposed by Almeida
and Neto [9] and Olivier et al. [16]. This procedure involved the control of effective
pressure on liquid resin during cure with the simultaneous introduction of moisture
between layers during layup, as suggested by Gurdal et al. [17]. Moisture was introduced
by spraying water finely and uniformly to produce laminate plates with homogeneous
porosity. The effective pressure on the liquid resin and the amount of moisture dispersed
into the laminate were used to control the void content [5]. All specimens were cured in an
autoclave at 160 C with different pressures.
Two groups of cure pressure cycles were selected to assess the influence of voids on the
mechanical properties. One group kept the first isothermal time as 120 min and changed
1778 Z.-S. GUO ET AL.

the autoclave pressures as 0.0, 0.1, 0.2, 0.4, and 0.6 MPa. The other group selected the
temperature held at 120 C for 0, 30, 90, and 120 min. The autoclave pressure was set
as 0.2 MPa. The time for which the pressure was applied was chosen with respect to the
minimum viscosity and gelation points. The pressure inside the vacuum bag is 0.1 MPa and
remains so throughout the entire cycle for all experiments. A reference specimen
was produced to represent the behavior of low void content laminates. All laminates
with 12 plies, [0/90]3S, were manufactured as described above for each of the cure routes
considered respectively. The size of each plate was about 300  300  2 mm.
All plates were ultrasonically inspected to assess the resulting distribution of porosities.
Areas of uniform porosity within each plate were identified, and different levels of void
and fiber content specimens and at least 10 interlaminar shear and flexure specimens
were cut from each of those areas.

C-scan Ultrasonic Inspection

All plates were inspected by double-through-transmission technique (Figure 1). The

host machine was SONIC—138VFD and the probe diameter was 25 mm. The scanning
was implemented with a 5 MHz ultrasonic failure detector by an automation M400D
system to generate an actual size map of the plate, associating a color with each
attenuation level. This feature was used to identify areas of constant porosity level.
The measured ultrasonic attenuation is the result of three factors: front surface loss,
transmission loss, and back surface losses. The front and back surface losses do not
depend on the condition of the panel, apart from its surface finish, and would be
expected to be independent on plate thickness [18]. On the other hand, transmission
loss depends on the defects present in the laminate. A calibration procedure was used
to compensate for the front and back surfaces, such that the measured attenuation
corresponds to transmission losses through the specimen only [3]. Three independent
scans of each plate were performed to measure the absorption coefficient of selected
areas with approximately uniform porosity level. The average value of these
measurements is considered as the absorption coefficient of the samples respectively.

Figure 1. Ultrasonic double-through-transmission technique.

Critical Void Content for Thermoset Composite Laminates 1779

The ultrasonic absorption coefficient  may be defined as

At
¼ ð1Þ
t

where  is measured in decibels per millimeter and depends on the internal condition of
the laminate, particularly on void content, and where transmission loss At (measured
in decibels) is assumed to increase linearly with plate thickness t.
Three independent scans of each plate were performed to measure the absorption
coefficient of selected areas with approximately uniform porosity level.

Characterization of Void Content

After the C-scan inspection, specimens were cut and their densities were determined
in order to estimate void content. The density of each composite sample was determined
using the water displacement method by measuring its weight in air and in water. The void
content and volume fiber content of each specimen were measured by matrix digestion
according to ASTM D3171. The void content measurement was made with five specimens
on each plate and the average was taken as the normal void content associated with
each porosity level. Specimens with different porosity levels ranging from 0 to 3.5%
were obtained for these carbon/epoxy laminates. The fiber and resin density used for
computations were provided by the manufacturer.

Microstructural Analysis

Microscopic image analysis is reported as one of the best methods to measure void
contents [1]. In addition, this technique provides detailed information on other vital
parameters such as void location, shape, and size that cannot be assessed by other
methods. The void morphology in composite laminates is usually assessed by microscopic
image analysis. A ZEISS MC80DX Microscope equipped with a camera and an image
processing system was used for all analyses. Void features are obtained from images
acquired at 200 magnification using a PC-based CCD camera mounted on a MEIJI
optical microscope. The selected magnification of 200 enables the assessment of voids as
small as the radius of a single fiber of 7 mm. Consequently, all identifiable voids
throughout the entire composite samples were included in the analysis of void content
and morphology.

Mechanical Tests

INTERLAMINAR SHEAR STRENGTH TESTS

The results of the interlaminar shear strength (ILSS) tests were measured according
to short beam shear (ASTM D2344). There were 10 specimens, with dimensions
25  8  2 mm, tested to assess the effect of void content on the ILSS. A nominal span
length of 20 mm and a 6.4 mm-diameter loading nose were used. The diameter of the
supports was 3.2 mm. The tests were performed in an Instron mechanical testing machine
using a test speed of 1 mm/min.
1780 Z.-S. GUO ET AL.

FLEXURE STRENGTH TESTS

The flexure strengths were obtained according to three-point flexure (ASTM D790).
There were 10 specimens, with dimensions 40  15  2 mm, tested to assess the effect
of void content on the flexure strength. The tests were also performed in an Instron
mechanical testing machine using a test speed of 1 mm/min.

TENSILE STRENGTH TESTS

The tensile strengths were obtained according to tensile testing (ASTM D3039).
There were 10 specimens, with dimensions 180  12  2 mm, tested to assess the effect of
void content on the tensile strength. The tests were also performed in an Instron
mechanical testing machine using a test speed of 0.5 mm/min.

RESULTS AND DISCUSSION

The Viscosity Curve of TDE85 Epoxy Resin

Rheological tests provide important information used to determine the time during
which the pressure must be applied. The results of viscosity are plotted against time
and temperature in Figure 2. It shows that there exists one minimum viscosity plateau
(from Point A to B) in the viscosity versus time curve. At the end of the plateau, the
viscosity increases quickly with time or temperature owing to the initiation of cure
reaction. In those cure temperature cycles, a—b—c—d is the cure cycle recommended
by the prepreg’s manufacturer.

Ultrasonic Attenuation Properties of Polymeric Composites

Figure 3 shows a C-scan of the studied laminates in a double-through-transmission

technique with pressures 0.0 MPa (a) and 0.6 MPa (b). It shows the distribution and
location of porosity. Figure 3(c) gives the presentation of grey levels associating to each

200
12 Viscosity
-------- Temperature
d
10 160
g
Temperature (°C)

f e c
Viscosity (Pa s)

8 b
120

6 a
80
4

40
2
A B
0
0 30 60 90 120 150 180 210
Time (min)

Figure 2. Plot of viscosity vs time and temperature.

Critical Void Content for Thermoset Composite Laminates 1781

attenuation level, denoted as the echo amplitude. The area porosity corresponding to echo
amplitude is also shown. It was calculated from the results of ultrasonic attenuation and
void contents through experimental method. Hence, area porosity levels are affected
by frequency and echo amplitude of ultrasonic inspection, material type such as the fiber
and the matrix, and the thickness of composite laminates, and so on. This method
is a quantitative grading evaluation criterion that will be very useful in engineering
application.
Figure 4 correlates the void content determined by acid digestion (ASTM D3171) to the
measured absorption coefficients  for all laminates studied. As expected, the smallest

Figure 3. C-scan showing areas with different void contents: (a) cure pressure is 0.0 MPa; (b) cure pressure
is 0.6 MPa; and (c) presentation of grey level and its porosity.

Figure 4. Correlation between void contents and absorption coefficient.

1782 Z.-S. GUO ET AL.

absorption coefficient corresponds to the low porosity laminates. This suggested that
greater void content causes increased attenuation and a linear correlation between the
porosity and the absorption coefficient can be observed for laminates with a porosity
range from 0 to 3.5%. Few measurements could be made of very high void content
laminates because of the difficulty in consistently obtaining laminates with a uniform
distribution of voids for high void contents. However, the results of this work demonstrate
that the range of void content obtained for all specimen types sufficed to determine the
maximum allowable void content.
Ultrasonic attenuation depends on a number of factors [3–5,9]: laminate
thickness, type of reinforcement, matrix material, fiber content, and the internal condition
of the material, which includes the void content. Moreover, the shape, size, and
distribution of voids in the laminate and parameters such as the ultrasound frequency,
size of the probe, and calibration procedure also affect the absorption coefficient
measurements. Therefore, it is difficult to compare ultrasonic attenuation results
obtained by other researchers without full knowledge of all the factors that affect the
measurements.

Optical Image Analysis

As illustrated in Figure 5, voids are seen at two different locations within the laminates.
The first location is defined as areas rich in matrix away from fibers. Voids encountered
in this location are completely surrounded by the epoxy matrix. The shapes of voids are
mostly circular (Figure 5(a)). The second location is defined as areas rich in interface,
where the area is primarily composed of reinforcing fibers. The shapes of voids are more
elliptical or of different irregular geometry (Figure 5(b)). A photomicrograph of the
sample of low porosity laminate is shown in Figure 5(a), which has a void content of 0.6%
with a curing pressure of 0.6 MPa. In this view, the voids occur mostly at the resin-rich
region. Figure 5(b) shows the occurrence of much larger, flattened, and elongated voids
distributed in the samples with a cure pressure of 0.0 MPa. As can be seen, the voids occur

Figure 5. Photomicrograph of voids with different cure pressures: (a) 0.6 MPa and (b) 0.0 MPa.
Critical Void Content for Thermoset Composite Laminates 1783

mostly at the ply interface. The differences in the morphology of the voids are determined
by the process parameters and physical properties of the resin.

Cure Cycles and Void Contents

Physically, void removal by pressurization is feasible that pressurization results

in void dissolution. Efforts were made to describe the effects of cure pressure
conditions on void contents. A set of cure cycle pressures, 0.0, 0.1, 0.2, 0.4, and
0.6 MPa, were used. The void content of each type of laminate is plotted in Figure 6 as
a function of the cure pressure. It shows that an exponential decrease fitting curve can
be well characterized by the relationship between void content and cure pressure.
This result is similar to the works of Tang et al. [10], Boey et al. [11], and Olivier et al. [16].
Hence, for a particular application, one can determine an appropriate magnitude of
curing pressure according to void levels required when other process parameters are
unchanged.
In order to determine the optimum time of first dwell time, some experiments
were carried out. The pressure within the range of minimum viscosity was applied
for 30, 60, 90, and 120 min, respectively. Figure 7 shows the results of void contents
varying with the first dwell time. It shows that there exists one time span for applying
pressures, and if the level of porosity limits to less than 1.5%, the time span chosen
between the 50 and 90 min is better. If pressure is applied too early, a resin-rich region may
appear in the laminates. It usually resulted in high void content and low mechanical
strength (Figures 7 and 8). Of course, pressure must not be applied too late because of the
existing resin-starved region which also resulted in high void content and low mechanical
strength (Figures 7 and 8). Hence, the optimum time of applying pressure is not 120 min
but 90 min. The resin viscosity affects the resin flow and also affects the transport of voids,
void formation, and growth. Experimental results show that the resin can flow fully in the

Figure 6. Void contents as a function of cure pressures.

1784 Z.-S. GUO ET AL.

3.0

2.5

Void content (%)

2.0

1.5

1.0

0.5

0.0
0 30 60 90 120 150
Dwell time (min)

Figure 7. Void contents vs the first dwell time.

1.2% 0.9%
100
(Void content) 2.0%
2.2%
Strength fraction (%)

95

90

85

Shear strength
80 Flexural strength
Tensile strength
75
0 30 60 90 120
Dwell time (min)

Figure 8. Mechanical strength vs the first dwell time.

time span from 50 to 90 min. An improved cure cycle is obtained as a—b—e—d (Figure 2).
The time was shortened to 30 min.

Fracture Criterion

A fracture criterion that correlates fracture stress with void content or, alternatively,
to ultrasonic attenuation, is needed to establish an acceptance level for the inspection.
Criteria to predict the laminate strength under certain loading conditions in the presence
of voids are scarce in the literature. Almeida and Neto [9] proposed a fracture criterion
that presented good correlation with experimental data on the bending strength of
composite laminates with voids. The same idea was successfully applied to predict the
interlaminar shear strength by Costa et al. [3–5], Jeong [7], and Almeida and Santacreu [8].
Critical Void Content for Thermoset Composite Laminates 1785

Almeida and Neto [9] also took advantage of the form of the equation to estimate the
critical void content defined as the void content below which the strength of the laminate is
not significantly affected by the presence of the voids. A similar approach was used by
Soriano and Almeida [19] to analyze the fracture strength data of composite laminates
with circular notches.
The considered fracture criterion for the strength of composite laminates containing
voids is given by

f ¼ HðÞm ð2Þ

where  f is the fracture stress, H is the laminate toughness,  is the ultrasonic absorption
coefficient in decibels per millimeter, and m is the slope parameter.
Equation (2) provides a good fit to experimental results for specimens with voids
[3–5,7–9,20]. However, it predicts infinite fracture stress for void-free laminates. To avoid
this inconsistency, the fracture criterion assumes that, for low void content, fracture
occurs according to classical fracture mechanisms with no influence of void content
(fiber microbuckling for compression and shear failure for ILSS tests). Therefore, for low
void content, fracture stress  f is assumed to be given by

f ¼ f0 ð3Þ

where  f0 is the fracture stress of a laminate with low void content.

Equations (2) and (3) yield a critical value for the ultrasonic absorption coefficient cr,
below which the void content does not affect the laminate strength. The value of cr can
be computed from

1  
f0
logðcr Þ ¼  log ð4Þ
m H

where cr is the critical value of the ultrasonic absorption coefficient. Note that the
definition of the critical ultrasonic absorption coefficient provides a systematic approach
to establish a maximum allowable value for void content. The approach is derived from
the mathematical form of the fracture criterion, which, in turn, is a consequence of the
basic assumptions described earlier. Therefore, the critical ultrasonic absorption
coefficient cr should be interpreted as a reference value for the minimum value of void
content that affects laminate strength rather than as a physical characteristic of the
laminate.
The proposed criterion in logarithmic form becomes

logðf0 Þ   cr
logðf Þ ¼ ð5Þ
logðH Þ  m logðÞ  > cr

Therefore, this criterion implies that the logarithmic plot of fracture stress  f,
as a function of ultrasonic absorption coefficient , should be approximately linear
for >cr and constant for   cr. Logarithmic plots of experimental results are presented
to assess the validity of this hypothesis for all cases considered in the present work (Figures
9–11). A straight line obtained from a best-fit procedure is included in all plots. Note that,
1786 Z.-S. GUO ET AL.

44
σf = 47.3(α)−0.364

42

Interlaminar strength (MPa)

40

38

36

34

32
0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
Absorption coefficient (dB/mm)

Figure 9. Interlaminar shear strength as a function of ultrasonic absorption coefficient.

560 σ = 609(α)−0.365

540
Flexural strength (MPa)

520

500

480

460

440

420

0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
Absorption coefficient (dB/mm)

Figure 10. Flexure strength vs ultrasonic absorption coefficient.

to be consistent with Equation (5), the best-fit procedure must not include the reference
laminate that corresponds to the fracture stress of a laminate with low void content,  f 0.
To verify the robustness of the fracture criterion, it was also applied to experimental
data obtained independently. Figure 12 shows the application of the proposed fracture
criterion to results obtained by Costa et al. [3–5] for the interlaminar shear strength of
carbon tape/epoxy laminates with voids. Figure 13 shows the application of the proposed
fracture criterion to the results obtained by Stone and Clarke [18] for the interlaminar
shear strength of carbon tape/epoxy laminates with voids.
Critical Void Content for Thermoset Composite Laminates 1787
1450
σ =1536(α)−0.310
1400

Tensile strength (MPa)

1350

1300

1250

1200

1150

1100
0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
Absorption coefficient (dB/mm)

Figure 11. Tensile strength vs ultrasonic absorption coefficient.

Figure 12. Logarithmic plot of interlaminar shear strength as a function of ultrasonic absorption coefficient for
carbon tape/epoxy laminates by Costa et al. [3–5].

The results shown in Figures 9–13 demonstrate that a good agreement with
the experimental results was obtained from the application of the fracture criterion to
fit the fracture stress as a function of the absorption coefficient  for all the
cases considered. Note that the experimental data include the effect of load type
(flexure and interlaminar shear) and experimental data from other researchers. Therefore,
the proposed fracture criterion yields good estimates of laminate fracture stress for a wide
1788 Z.-S. GUO ET AL.

Figure 13. Logarithmic plot of interlaminar shear strength as a function of ultrasonic absorption coefficient for
carbon tape/epoxy laminates by Stone and Clarke [18].

Table 1. Fracture parameters for all laminates under interlaminar shear,

flexure, and tensile strength tests.
m H (MPa) acr (dB/mm) Vcr (%)
ILSS 0.364 47.3 1.46 1.11
Flexure strength 0.365 609 1.38 1.05
Tensile strength 0.310 1536 1.45 1.10

range of situations, provided that adequate values for fracture parameters are used in each
case. The fracture parameters are the laminate toughness H, slope parameter m, and
critical void content, either in terms of the absorption coefficient cr (decibels per
millimeter), or in terms of volume fraction Vcr (percent).
The values of the fracture parameters which are listed in Table 1 are obtained from the
linear regression procedure based on Equation (5). The value of m is about 0.365 either
under flexure or interlaminar shear loading, 0.310 under tensile loading for all specimens.
It seems that the slope parameter m depends on the loading type. The laminate toughness
H depends on the loading type as demonstrated by the results in Table 1.
An important fracture parameter is the critical void content, which establishes
an acceptance criterion for the nondestructive inspection of composite laminates.
When expressed in terms of the critical absorption coefficient, it depends on the type of
loading. For laminates under ILSS and tensile loading, the critical absorption coefficient
is 1.45 dB/mm; for flexure strength measurements, the critical absorption coefficient
is 1.38 dB/mm. These results indicate that the type of loading has a significant effect on
the critical absorption coefficient.
When the critical void content is expressed in terms of critical volume fraction Vcr, the
load type has little influence on the critical void content. Note, however, that from
Critical Void Content for Thermoset Composite Laminates 1789

the point of view of nondestructive inspection, the critical void content must be established
in terms of the parameters of ultrasonic inspection, that is, the critical absorption
coefficient cr.

CONCLUSIONS

An experimental program aiming at establishing acceptance levels for the attenuation

level in the ultrasonic inspection of composite laminates is described. Laminates produced
with carbon tape/epoxy were produced with intentionally high void content. The
absorption coefficient was measured for all specimens and shown to vary approximately
linearly with the void content, corroborating previous results. The magnitude and time of
applied pressure have a significant effect on void content. If they were selected properly,
the entrapped air, water vapor, and excessive resin can be squeezed out from the laminates
and the first dwell time can be shortened. Microscopic image analysis was utilized to
examine void content, location, shape, etc.
A fracture criterion was applied to the experimental data to correlate the interlaminar
shear strength, tensile strength, and flexure strength of the laminates to the absorption
coefficient. The theory presented good correlation with the experimental data for all cases,
including experimental data independently obtained by other researchers. Therefore,
it was demonstrated to be a useful nondestructive inspection aid in certifying the structural
integrity and safety of composite laminates containing porosity. The application of the
fracture criterion to the experimental data establishes a systematic approach to estimate
an inspection acceptance criterion based on a critical void content below which the
strength of the laminate is not significantly affected by the presence of the voids.
The critical void content may be described either in terms of the absorption coefficient
in the ultrasonic inspection or in terms of the volume fraction. It is shown that, when
expressed in terms of the critical absorption coefficient, the critical void content depends
on the type of loading. However, the resulting values are also dependent on parameters
of the ultrasonic inspection, namely, the frequency, diameter of the probe, calibration
procedure, etc. Therefore, the estimated critical absorption coefficient may only be used
as an inspection criterion for the same conditions used in this work. These data, however,
cannot directly be used as an acceptance criterion in a nondestructive inspection of
composite laminates.

REFERENCES

1. Hamidi, Y.K., Aktas, L. and Altan, M.C. (2005). Effect of Packing on Void Morphology in Resin
Transfer Molded E-glass/epoxy Composites, Polymer Composites, 26(5): 614–627.
2. Cantwell, W.J. and Morton, J. (1992). The Significance of Damage and Defects and Their
Detection in Composite Materials: A Review, J. Strain Analysis, 27(1): 29–42.
3. Costa, M.L., Rezende, M.C. and Almeida, S.F.M. (2001). Influence of Porosity on the
Interlaminar Shear Strength of Carbon/Epoxy and Carbon/Bismaleimide Fabric Laminates,
Compos. Sci. Tech., 61(14): 2101–2108.
4. Costa, M.L., Almeida, S.F.M and Rezende, M.C. (2005). Critical Void Content for Polymer
Composite Laminates, AIAA J., 43(6): 1336–1341.
5. Costa, M.L., Rezende, M.C. and Almeida, S.F.M. (2005). Strength of Hygrothermally
Conditioned Polymer Composites with Voids, J. Compos. Mater., 39(21): 1943–1961.
1790 Z.-S. GUO ET AL.

6. Suarez, J.C., Molleda, F. and Guemes, A. (1993). Void Content in Carbon Fiber/Epoxy
Resin Composites and Its Effects on Compressive Properties, Proceedings of ICCM-9, Vol. 6,
pp. 589–596.
7. Jeong, H. (1997). Effects of Voids on the Mechanical Strength and Ultrasonic Attenuation
of Laminated Composites, J. Compos. Mater., 31(3): 276–292.
8. Almeida, S.F.M. and Santacreu, A.C.M. (1995). Environmental Effects in Composite Laminates
with Voids, Polym. Compos., 3(3): 193–204.
9. Almeida, S.F.M. and Neto, Z.S.N. (1994). Effects of Void Content on the Strength of
Composite Laminates, Compos. Struc., 28(2): 139–148.
10. Tang, T.M., Lee, I.W. and Springer, G.S. (1987). Effects of Cure Pressure on Resin Flow,
Voids and Mechanical Properties, J. Compos. Mater., 21: 421–440.
11. Boey, F.Y.C. and Lye, S.W. (1992). Void Reduction in Autoclave Processing of Thermoset
Composites, Composites, 23(4): 261–265.
12. Mouritz, A.P. (2000). Ultrasonic and Interlaminar Properties of Highly Porous Composites,
J. Compos. Mater., 34(3): 218–239.
13. Wisnom, M.R., Reynolds, T. and Gwilliam, N. (1996). Reduction in the Interlaminar Shear
Strength by Discrete and Distributed Voids, Compos. Sci. Tech., 56(1): 93–101.
14. Ghiorse, S.R. (1993). Effect of Void Content on the Mechanical Properties of Carbon/Epoxy
Laminates, SAMPE Quarterly, 24(2): 54–59.
15. Bowles, K.J. and Frimpong, S. (1992). Voids Effects on the Interlaminar Shear Strength
of Unidirectional Graphite-Fiber-Reinforced Composites, J. Compos. Mater., 26(10):
1487–1509.
16. Olivier, P., Cottu, J.P. and Ferret, B. (1995). Effects of Cure Cycle Pressure and Voids on Some
Mechanical Properties of Carbon/Epoxy Laminates, Composites, 26(7): 509–515.
17. Gurdal, Z., Tomasino, A. and Biggers, S.B. (1991). Effects of Processing Induced Defects on
Laminate Response: Interlaminar Tensile Strength, SAMPE J., 27(4): 3949–3953.
18. Stone, D.E. and Clarke, B. (1975). Ultrasonic Attenuation as a Measure of Void Content in
Carbon-Fibre Reinforced Plastics, Non Destructive Testing, 8(3): 137–145.
19. Soriano, E.A. and Almeida, S.F.M. (1999). Notch Sensitivity of Carbon/Epoxy Fabric
Laminates, Compos. Sci. Tech., 59(8): 1143–1151.
20. Judd, N.C.W. and Wrigth, W.W. (1978). Voids and Their Effects on the Mechanical Properties
of Composites, An Appraisal, SAMPE J., 14(1): 10–14.