MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 1 of 16

MATC(MN)06

Through-Thickness Testing of Polymer Matrix

Composites
Interlaminar or through-thickness (TT) stresses and strains occur in laminated structures; irrespective of loading
mode or laminate thickness. Edge effects, holes, changes in thickness, curved edges, and bonded and bolted
joints all result in TT stress and strain gradients. The presence of interlaminar tensile (peel) and shear stresses
combined with inherently low TT strength properties may cause local microcracking and delaminations, and
eventual failure. As a result, TT stiffness and strength properties are increasingly being required by structural
engineers/designers for finite element analysis and other analysis methods in order to ensure efficient and safe
use of composite materials in secondary and primary load bearing structures.

This Measurement Note describes test methods suitable for determining TT properties under tensile, compressive
and shear loads. These methods can provide a full complement of TT elastic and strength property data, and
stress-strain response to failure. The methods described in this Measurement Note have been successfully used
with glass and carbon fibre-reinforced composite materials containing a wide range of different fibre formats,
including continuous aligned, woven fabric and random mat. The use and limitations of these methods for
characterising T-T properties under static and cyclic loads are discussed. The Measurement Note provides details
on specimen fabrication, test apparatus and procedure, and failure mode.

The research reported in this Measurement Note, which has led to the drafting of standards for TT tension,
compression and shear test methods, was undertaken as part of the 'Materials Measurement Programme' funded
by the Engineering Industries Directorate of the UK Department of Trade and Industry. Future work will be directed
towards validating the test methods by finite element analysis and interlaboratory comparative exercises.

W R Broughton, M R L Gower, M J Lodeiro and R M Shaw

May 2001

Introduction
There is a tendency to associate interlaminar stresses or through-thickness (TT) stresses with "thick" sections,
however interlaminar stresses and strains may be induced in "thin" laminates through the application of membrane
(i.e. in-plane) loads. Features, such as free edges, holes, changes in thickness, curved edges, and bonded and
bolted joints all act as stress concentrators (regions of high interlaminar tensile and shear stresses). The presence
of interlaminar tensile (peel) and shear stresses combined with inherently low TT strength properties may cause
local microcracking and delaminations. Delamination growth can cause severe reductions in stiffness and strength,
and eventual failure under static or fatigue loading conditions.

TT stiffness and strength properties are increasingly being required by structural engineers/designers for finite
element analysis (FEA) and other analysis methods to ensure efficient and safe use of composite materials in
secondary and primary load bearing structures. Recent developments have seen the emergence of tensile,
compressive and shear test methods suitable for generating TT elastic and strength property data [1-3].

This Measurement Note describes test methods suitable for determining TT properties under tensile, compressive
and shear loads. These methods can provide a full complement of TT elastic and strength property data, and
stress-strain response to failure (Table 1). The methods described in this Measurement Note have been
successfully used with glass and carbon fibre-reinforced composite materials containing a wide range of different
fibre formats, including continuous aligned, woven fabric and random mat. Details are provided on specimen
fabrication, test apparatus and procedure, data analysis (or reduction) and failure mode. The use and limitations of
these methods for characterising TT properties under static and cyclic loads are discussed.

Table 1: Short-Term Data for Orthotropic Materials

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 2 of 16

Load Type Elastic Moduli Poisson's Ratio Strength Failure Strains

E11, E22 ν12, ν21 S11T, S22T ε11, ε22

Tension
[E33] [ν13, ν23] [S33T] [ε33]

E11, E22 ν12, ν21 S11C, S22C ε11, ε22

Compression
[E33] [ν13, ν23] [S33C] [ε33]

G12 S12 ε12

Shear
[G13, G23] [S13, S23] [ε13, ε23]

Through-Thickness Testing
Many approaches have considered for determining TT properties of polymer matrix composites (PMCs) with
varying degrees of success [1]. Ideally, the interaction between specimen geometry and loading configuration
should aim to produce a region of pure, uniform, maximum stress in order to promote repeatable quantitative
measurements. The test method should also be applicable to a wide range of materials and loading modes (i.e.
static and fatigue).

Shear
The level of standardisation of TT shear test methods is well advanced when compared to TT tension and TT
compression. Although there is a multitude of test configurations, most of these are unsuitable for measuring TT
shear properties. Three test methods, the short beam shear (BS EN ISO 14,130 [4]), V-notched beam (ASTM D
5379 [5]) and double-notched shear (ASTM D 3846 [6]) have been evaluated as to their suitability for generating
TT shear properties. The short beam and double notch shear methods are only suitable for determining
interlaminar shear strength. The V-notched beam can be used to measure in-plane and TT shear properties for
isotropic and anisotropic materials.

Short Beam Shear

Short beam shear specimen preparation is straightforward and only a small amount of material is required. Tests
are carried out according to BS EN ISO 14,130, at a displacement rate of 1 mm/min. A schematic of the loading
configuration is shown in Figure 1. The standard specimen thickness h is 2 mm (nominally) with a width and
overall length of 10 and 20 , respectively. The recommended width is 5h and overall length is 10h for material
sectioned from panels with a thickness in excess of 2 mm. In all cases, the loading span is 5h ± 0.3 mm.

Figure 1: Short beam shear. unidirectional glass fibre-reinforced laminate.

The short beam shear test can be used for determining interlaminar shear strength of continuous aligned and 2-D
woven fabric laminates. The shear stress distribution through the beam and along the beam length, is non-uniform.
Failure frequently results from a combination of high shear and transfibrile compression gradients, present in the
orthotropic beams near the central loading point. Occasionally, compressive buckling or yielding occurs in the
upper portion of the beam. Testing and data analysis is relatively straightforward with the interlaminar shear

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 3 of 16

strength S given as:

where Pmax is the maximum applied, b and h are the width and thickness, respectively.

There was reasonable agreement between interlaminar shear strength measurements obtained for aligned
glass/epoxy and carbon/epoxy materials using the short beam shear and the V-notched beam methods (Table 2).

Double Notch Shear

The double notch shear test consists of loading a non-symmetrically notched composite specimen in either
uniaxial tension (BS 4994 [7] and BS 6464 [8]) or compression (ASTM D 3846). The tension specimens are
unsupported, whereas compression specimens (see Figures 2 and 3) are supported along the entire length of
specimen to minimise out-of-plane deformation.

Figure 2: Double notch shear specimen. unidirectional glass fibre-reinforced laminate.

The experimental evaluation showed that the interlaminar shear strength data obtained using the compressive
method (Figure 3) were generally consistent with those measured using the V-notched beam test (Table 2).
Fractographic analysis indicated that shear stresses dominated failure when these materials were loaded in
compression.

Figure 3: Compression fixture with double notch shear specimen.

The interlaminar shear strength measured using the compression test was consistently higher than corresponding
tension tests. The lower strength values for the tension test can be attributed to peel stresses and shear stress
concentrations at the notches. Large bending deformations were observed when double notch specimens were

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 4 of 16

loaded in tension.

The ASTM standard specifies a nominal length of 79.5 mm and a width of 12.7 mm. In the evaluation exercise,
10 mm wide specimens were used. This was to ensure equivalence with the short beam method. Specimen
thickness varies from 2.5-10 mm. Two parallel notches, one on each opposite face of the specimen and 6.4 mm
(0.25 in) apart, are cut across the entire width of the specimen and centrally located along its length. The notches
can be machined using either a diamond slitting wheel (water lubricated) or a tungsten carbide end-mill cutter. The
notch, which extends to the specimen mid-plane, is ~1.5 mm wide. The specimen is end-loaded in the supporting
fixture (see ASTM D695 [9]) at a displacement rate of 1 mm/min. The securing screws are only finger tight. It is
essential that the specimen is free to move along its longitudinal axis (free of friction).

Table 2: Through Thickness Shear Results

Shear Modulus (GPa) Shear Strength (MPa)

Test Method
Gxz Gyz Sxz Syz

Unidirectional Carbon/Epoxy

V-notched beam 5.3 ± 0.2 2.9 ± 0.3 111 ± 2 64 ± 9†

Double notch n/a n/a 75 ± 12 n/a

Short beam shear n/a n/a 108 ± 6 n/a

Chopped Strand Mat

V-notched beam 1.64 ± 0.09 [1.64 ± 0.09] 40.7 ± 1.7 [40.7 ± 1.7]

Double notch n/a n/a 38.3 ± 4.7 [38.3 ± 4.7]

Short beam shear n/a n/a 13.4 ± 3.0† [13.4 ± 3.0]†

2 x 2 Twill Glass Fabric/Epoxy

V-notched beam 4.12 ± 0.14 [4.12 ± 0.14] 68.4 ± 0.9 [68.4 ± 0.9]

Double notch n/a n/a 64.9 ± 1.8 [64.9 ± 1.8]

Short beam shear n/a n/a 52.5 ± 0.6† [52.5 ± 0.6]†

Discontinuous Glass/Nylon 66

V-notched beam 1.68± 0.06 [1.68 ± 0.06] 56.9 ± 3.6† [56.9 ± 3.6]†

Double notch n/a n/a 66.4 ± 4.8 [66.4 ± 4.8]

Short beam shear n/a n/a 18.6 ± 0.4† [18.6 ± 0.4]†

Random Mat/Polypropylene

V-notched beam 1.04 ± 0.04 [1.04 ± 0.04] 22.7 ± 0.8 [22.7 ± 0.8]

Double notch n/a n/a 38.3 ± 4.7 [18.1 ± 3.3]

Short beam shear n/a n/a 14.2 ± 1.5† [14.2 ± 1.5]†

*n/a = not applicable; † non-shear value; [ ] denotes value assumed from material symmetry.

This method provides consistent strength data (low scatter) with interlaminar failure occurring along the mid-plane
joining the notches. Notch depth must be accurately machined. Higher strength values are produced when the

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 5 of 16

notch depth is less than half the specimen thickness (i.e. < h/2). This is due to an increase in the cross-sectional
area of the fracture path (i.e. not along the mid-plane). Conversely, notch depths exceeding a depth of h/2 will fail
prematurely giving low strength values. Bending and peeling are frequently observed in over-cut specimens.

No allowances are made for non-uniformity of the stress distribution between notches or the effect of local stress
concentrations at the fibre ends associated with the notch geometry. Failure surfaces indicate shear dominated
mixed-mode failure.

The shear strength S is given by:

where Pmax is the maximum applied, b and l are the width and notch separation, respectively.

V-Notched Beam Shear

This test employs a double edge-notched, flat, rectangular specimen (Figure 4). Tests are carried out according to
ASTM D 5379 using the loading fixture shown in Figure 5. Standard specimens are 76 mm in length and 20 mm
wide with a nominal thickness of 5 mm. Two 90° angle notches with a notch root radius of 1.3 mm are machined at
the edge mid-length with faces oriented at ±45° to the longitudinal axis, to a depth of 20% of the specimen width
(i.e. 4 mm). Notches are machined using a water lubricated, diamond grinding wheel. Shear strain is measured by
bonding two biaxial strain gauges, one on each opposite face of the specimen at the specimen centre in the area
between the notches. The strain gauges should have an active gauge-length of 1 or 2 mm and must be aligned
±45° to the longitudinal axis of the specimen (Figure 4).

Figure 4: V-notched beam specimen. unidirectional glass fibre-reinforced laminate.

Provided adequate material thickness is available (i.e. 20 mm thick), this method can be used to measure shear
modulus and shear strength in all three material planes (x-y, x-z and y-z) for a diverse range of composite
materials. Differences between shear moduli, due to out-of plane deformation, could be as high as 10% for a batch
of nominally identical specimens. To ensure maximum accuracy, shear modulus is determined from the average
response of back-to-back biaxial rosettes. At present the standard requires only one specimen from a batch to be
tested in this manner provided the amount of twist for the test specimen is no greater than 3%.

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 6 of 16

Figure 5: V-notched beam loading fixture and specimen.

The average shear strength S and shear modulus G can be determined using:

where Pmax is the maximum applied, w is the distance between the notches and h is the thickness. The variables
∆P, ∆ε45 and ∆ε-45 are the change in applied load and +45° and -45° normal strains in the initial linear region of
the stress-strain (τ−γ) curve.

The following improvements to the test fixture will reduce differences in shear moduli obtained from the two sides
of specimen and possibly eliminate the need for two biaxial rosettes:

z Keying the bearing post or adding a second post to prevent twisting of the specimen due to the fixture
rotating on the bearing post; and

z Lateral adjustment to minimise out-of-plane deformation. At present, the specimen is not centrally loaded
through the specimen thickness.

In addition, the fixture needs to be modified to allow for large plastic deformations commonly encountered when
testing fibre-reinforced thermoplastics, by allowing larger deformations to be reached.

Shear property data obtained for a range of PMCs are shown in Tables 2 and 3. It is reasonable to assume
material property symmetry in the x-z and y-z planes for all the materials except the unidirectional carbon/epoxy
laminate. Various failure modes were encountered with the failure process being highly dependent on the
microstructure and orientation of the material. Most materials tended to fail in shear when tested in the x-z or y-z
planes.

Table 3: V-Notch Beam Results for Carbon and Glass Woven Fabrics

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 7 of 16

Shear Modulus (GPa) Shear Strength (MPa) Failure Strain (%)

Material
Gxz and Gyz Sxz and Syz γfxz and γfyz

Carbon Fabric/Epoxy 3.88 ± 0.05 65.0 ± 2.4 20.9 ± 1.4

Glass Fabric/Epoxy 3.23 ± 0.35 66.2 ± 5.4 20.5

Shear Fatigue Testing

Constant amplitude (sinusoidal waveform) shear fatigue tests were conducted on carbon and glass woven
fabric/epoxy materials using the V-notch beam method. The shear specimens were tested in load control using a
servo-hydraulic test machine at a frequency of 5 Hz. The stress ratio R (τmin/τmax) was equal to 0.1. Tests were
carried out at four different stress levels (90%, 80%, 70% and 55% Sxz). Fatigue strength data was normalised
with respect to the corrected interlaminar shear strength Sxz for an equivalent loading rate to the fatigue test.

The woven carbon laminate was observed to have excellent shear fatigue resistance. The fatigue life Nf exceeded
5 x106 cycles at a load of 90% of Sxz. The stress-life or S-N (S is the applied stress and N is the number of cycles
to failure) response for these materials was almost flat. In comparison, the woven glass laminate was less
resistant to fatigue loading (Figure 6). The normalised fatigue life (S-N data) for this material can be represented
by the following relationship:

where τmaxis the maximum applied load, Sxzis interlaminar shear strength, k (the slope) is the fractional loss in
interlaminar shear strength per decade of cycles and Nf is the number of cycles to failure. The value of k is
approximately 0.078.

Figure 6: S-N shear fatigue data for woven glass/epoxy laminate.

The change in stiffness with loading cycles due to damage formation (notch-root cracks and shear cracking in the
specimen mid-section) is shown in Figure 7. A failed glass fabric/epoxy specimen is shown in Figure 8.

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 8 of 16

Figure 7: Stiffness vs cycles (70% Sxz) for woven glass/epoxy.

Figure 8: Failed woven glass/ epoxy specimen.

Tension
There are two contrasting methods and several associated geometries, which can be employed in the
measurement of TT tension (i.e. direct and indirect tensile loading). The direct method introduces tensile load to
parallel sided (square cross-section) or waisted short block specimens via adhesively bonded bars or through
grips (Figures 9 and 10). The indirect method aims to induce TT tension, in significantly curved specimens by the
application of bending moments. The load is introduced via the specimen lever arms, in either tension or flexure
[1-3]. Indirect methods (e.g. C-section) tend to produce mixed mode failure, and not TT tension. In-plane tensile
methods identified in ISO 527-5 [10] are unsuitable for adaptation to TT tensile testing due to manufacturing
difficulties and costs associated with producing thick composite sections (i.e. 20 mm, or greater). ISO 527-5
specifies a specimen length of 250 mm.

Figure 9: Schematic of direct T-T tensile test methods From left to right: Parallel sided, circular, elliptical
and RARDE specimens.

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 9 of 16

Figure 10: Circular waisted specimen with end loading grips.

This section describes three direct methods that can be used for generating TT tensile and compressive
properties.

Short Parallel-Sided Block

Short parallel-sided blocks with a square cross-section can be used to measure TT tensile elastic properties. This
geometry is unsuitable for measuring TT tensile strength. Biaxial strain gauges (2 or 3 mm gauge-length),
measuring axial and transverse strains are bonded onto the mid-line of each of the four sides of the specimen.
This enables average strains to be calculated, thus accounting for bending due to small deviations in specimen or
load alignment.

Square cross-section specimens (15 mm square) with TT dimensions (i.e. thickness) ranging from 18 to 40 mm
have proved satisfactory for measuring TT tensile elastic properties (see Table 4). It is essential that the top and
bottom surfaces and faces are flat and parallel. This comment also applies to all surfaces. Specimens need to be
ground to size with a tolerance of ±0.1 mm. Reusable aluminium or hardened stainless steel (preferred option)
loading bars are bonded to the specimen using a high strength epoxy adhesive. Specimens need to be held in a
gluing fixture (e.g. V-blocks) to ensure good alignment between loading blocks and loading axis during the curing
process, and to maintain pressure on the bonding surfaces. Specimens need to be left for at least 24 hours before
testing.

Table 4: Through Thickness Tension Results

Test Method Poisson's Ratio Modulus, Ezz Strength, Szz

νzx νzy (GPa) (MPa)

UD Carbon/Epoxy

Parallel-sided block 0.019 ± 0.002 0.55 ± 0.01 9.9 ± 0.1 n/a

RARDE waisted block 0.020 ± 0.001 0.47 ± 0.01 9.5 ± 0.1 78 ± 7

Plain waisted block n/a n/a n/a 71 ± 6

Chopped Strand Mat

Parallel-sided block 0.23 [0.23] 6.3 n/a

RARDE waisted block 0.25 ± 0.05 0.21 ± 0.05 5.3 ± 0.4 8.2 ± 12

Plain waisted block n/a n/a n/a 9.1 ± 2.0

2 x 2 Twill Glass Fabric/Epoxy

Parallel-sided block 0.19 ± 0.01 [0.19 ± 0.01] 11.0 ± 0.4 n/a

RARDE waisted block 0.18 ± 0.04 0.17 ± 0.01 11.2 ± 0.1 36 ± 1

Plain waisted block n/a n/a n/a 41 ± 8

Discontinuous Glass/Nylon 66

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 10 of 16

Parallel-sided block 0.27 ± 0.04 0.41 ± 0.02 4.4 ± 0.1 n/a

Plain waisted block n/a n/a n/a 69 ± 7

Random Mat/Polypropylene

Parallel-sided block 0.16 ± 0.04 [0.16 ± 0.04] 3.5 ± 0.2 n/a

Plain waisted block n/a n/a n/a 5.7 ± 1.6

* n/a = not applicable; [ ] denotes value assumed from material symmetry.

This is a relatively straightforward and economic specimen to test, requiring no special loading fixture. Although,
care needs to be taken to ensure that misalignment, at both the fabrication and testing stages, is minimised.
Failure consistently occurs at the adhesive joints between the specimen and the metallic loading bars for all but
the weakest materials, and therefore this test geometry is unsuitable for measuring TT tensile strength. Difficulties
will be encountered when trying to bond to thermoplastic materials.

RARDE Waisted Block

This test geometry (developed by the Defence Evaluation and Research Agency (DERA)), can be used to
measure TT elastic and strength properties in both tension and compression. Specimens have 25 mm square
ends and a nominal height (thickness) of 38 mm. The gauge-section is 12 mm long, with a rectangular cross-
section of 10 mm (2-3 or y-z plane) x 16 mm (1-3 or x-z plane). Thinner material (19-20 mm thick) can be tested
provided the linear dimensions of the specimen are scaled in proportion to the standard geometry shown in
Figure 11.

Figure 11: RARDE waisted block specimen.

Specimen preparation, testing (Figure 12), strain gauging and data analysis are similar to those employed for the

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 11 of 16

parallel-sided short block specimens. A profile grinding wheel is used to waist the specimen. TT tensile property
data obtained using the standard RARDE geometry are shown in Tables 3 and 4. TT tensile strength SzzT is
determined using the following relationship:

where Pmax is the maximum applied and A is the cross-sectional area of the gauge-section.

Figure 12: RARDE tension test (courtesy of DERA)

RARDE specimens tend to fail in the gauge-section in the plane normal to the applied load. The presence of voids,
either locally or globally, will contribute to premature failure under static or fatigue loading conditions. Again, care
needs to be taken to ensure minimum misalignment, at both the fabrication and testing stages. Discounting the
problem of misalignment, this test method is relatively straightforward to perform. However, machining and
instrumentation is labour intensive and expensive. For the highest accuracy, the technique requires biaxial gauges
on all four faces. Extensometers are not particularly suitable for this test geometry.

Plain or Circular Waisted Block

In an attempt to simplify machining requirements, reduce production costs and minimise stress concentrations, a
plain/circular waisted specimen was developed for measuring TT tensile and compressive strengths. The
specimen has a square cross-section (25 mm base), which is waisted on all side faces by a plain radius along the
gauge-length (Figure 13).Specimen preparation and test procedure is identical to that employed for the other
specimens. Failure occurs consistently within the gauge-section of well bonded specimens (Figure 14). The large
(30 mm) radius ensures that the stress concentration at the mid-section is minimal (~1.10). Machining costs
compared with the RARDE test are marginally lower.

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 12 of 16

Figure 13: Plain waisted specimen.

Figure 14: Failed RARDE tension specimen.

Tensile Fatigue
Constant amplitude (sinusoidal waveform) tensile fatigue tests were conducted on carbon and glass woven
fabric/epoxy materials using the RARDE test method (Figure 12). The specimens were tested in load control at a
frequency of 5 Hz. The stress ratio R (σmin/σmax) was equal to 0.1. Tests were carried out at four different stress
levels (90%, 80%, 70% and 55% σUTS). Fatigue strength data is normalised with respect to the ultimate tensile
strength σUTS or in the case of TT tension SzzT (see Table 5 and Figure 15) for an equivalent loading rate to the
fatigue test.

Table 5: RARDE Tension/ and Compression Results for Carbon and Glass Woven Fabrics

Poisson's Ratio Modulus, Ezz Strength, Szz

Test Method
νzx νzy (GPa) (MPa)

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 13 of 16

Carbon Fabric/Epoxy

Tension 0.08 ± 0.01 0.08 ± 0.01 9.8 ± 0.5 51.4 ± 2.2

Compression 0.09 ± 0.01 0.09 ± 0.01 12.4 ± 0.9 763 ± 25

Glass Fabric/Epoxy

Tension 0.18 ± 0.01 0.17 ± 0.01 11.1 ± 0.6 48.5 ± 8.9

Compression 0.20 ± 0.01 0.20 ± 0.01 11.6 ± 0.9 526 ± 11

Provided the specimens are well aligned and the material is free of defects then the RARDE specimen can be
used to produce TT tensile fatigue (S-N) data (Figure 15). However, small misalignments during testing and the
presence of voids resulted in large uncertainties in the fatigue life of the woven fabric materials. The normalised
fatigue life for woven glass laminate can be represented by the following relationship:

where σmax is the maximum applied load and σUTS is ultimate TT tensile strength. The value of k ≈ 0.086. The
woven carbon laminate contained a high void content, which shortened the laminates fatigue life (Nf > 106 cycles
for σmax = 70% σUTS).

Figure 15: S-N tensile fatigue data for woven glass/epoxy laminate.

Compression
All three short block geometries described in the previous section have been evaluated in compression. The
results obtained using the three geometries (Tables 5 and 6) parallel those for tension. Specimens are loaded in
compression between flat, parallel hardened steel platens with recesses to reduce lateral movement of the
specimen. It is recommended for static tests only that a four pillar die set be used to maintain uniform compression
loading (Figure 16). Specimens 20-40 mm thick can be tested provided linear dimensions are scaled in proportion,
and that the material does not buckle under compressive loads. Thermoplastic composites tend to buckle, and
hence for these materials the plain waisted specimen is preferred method for measuring TT compressive strength.
Stress concentrations at the fillet root of waisted specimens can be expected to reduce strength measurements.
The effects of stress concentrations are less pronounced for the plain waisted blocks, and hence the generally
higher strength values compared with the RARDE geometry.

Table 6: Through Thickness Compression Results

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 14 of 16

In-plane tests (i.e. IITRI, Celanese or end block methods [11]), are unsatisfactory due to the difficulty in obtaining
material of the required thickness (110 mm) and the sensitivity of the specimens to premature failure during
handling and testing. Parallel-sided blocks are only suitable for measuring TT elastic properties. Stress
concentrations induced at the ends of specimens can result in premature failure. Shear is the predominant cause
of failure in all cases; independent of material microstructure, loading configuration or specimen size [3].

Figure 16: RARDE compression specimen and loading fixture (courtesy of DERA).

Compression Fatigue
Fatigue tests are normally carried out at the highest frequency possible in order to minimise test duration.
Restrictions on test frequency can arise due to test equipment limitations, (response time), time-dependent
processes and hysteretic heating. Hysteretic heating can precipitate a rise in temperature causing thermal
degradation of the material. The increase in temperature is dependent on strain or frequency rate, stress
amplitude, specimen geometry, loading mode and thermal conductivity of the material. Hysteretic heating, which
increases with stress amplitude and frequency, was observed to adversely affect the compression fatigue
performance of carbon and glass woven RARDE specimens. Test frequencies of the order of 1 to 5 Hz resulted in
a substantial increase in the surface temperature (> 170°C) and short fatigue life. Reducing the frequency to
0.1 Hz and stress amplitude to 25% ultimate compressive strength results in a fatigue life Nf > 5 x 106 cycles. The
maximum temperature rise of the material should be less than 10°C.

Concluding Remarks
The V-notched beam test has been shown to be suitable for measuring interlaminar shear properties under static

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 15 of 16

and fatigue loading conditions, although improvements to the ASTM test fixture are proposed. NPL has developed
draft procedures that can be used to measure TT elastic and strength properties under tensile and compressive
static loading conditions. The RARDE specimen has been shown to be suitable for use under tension-tension
fatigue loading conditions, provided the material is free of voids and the specimen is well aligned. This geometry is
also suitable for compression fatigue loading, providing hysteretic heating is avoided. The normalised S-N data for
TT shear and tension fatigue for a glass woven/epoxy were similar, and can be represented by a simple empirical
relationship.

References
1. An Overview of Through-Thickness Test Methods for Polymer Matrix Composites, W R Broughton and G D
Sims, NPL Report DMM(A) 148, 1994.
2. Mechanical Testing of Advanced Fibre Composites, Edited by J M Hodgkinson, Woodhead Publishing
Limited, 2000.
3. Understanding Limitations of Through Thickness Test Methods, M J Lodeiro, W R Broughton and G D
Sims, Plastics, Rubber and Composites, Volume 28, Number 9, 1999, pp 416-424.
4. BS EN ISO 14,130, Fibre-Reinforced Plastic Composites - Determination of Apparent Interlaminar Shear
Strength by Short-Beam Method.
5. ASTM D 3846, Standard Test Method for In-Plane Shear of Reinforced Plastics, Volume 8.02, ASTM
Standards, 2000, pp 471-473.
6. ASTM D 5379, Standard Test Method for Shear Properties of Composite Materials by the V-Notched Beam
Method, Volume 15.03, ASTM Standards, 2000, pp 241-253.
7. BS 4994, Design and Construction of Vessels and Tanks in Reinforced Plastics, 1987.
8. BS 6464, Reinforced Plastics Pipes, Fittings and Joints for Process Plants, 1984.
9. ASTM D 695, Test Method for Compressive Properties of Rigid Plastics, Volume 8.01, ASTM Standards,
2000, pp 78-84.
10. ISO 527-5, Plastics - Determination of Tensile Properties. Part 5: Test Conditions for Unidirectional Fibre-
Reinforced Plastic Composites.
11. BS EN ISO 14126, Fibre Reinforced Plastic Composites - Determination of Compression Properties.

Acknowledgements
The research reported in this Measurement Note was carried out by NPL, in conjunction with DERA, as part of the
'Materials Measurement Programme' funded by the Engineering Industries Directorate of the UK Department of
Trade and Industry. The authors acknowledge the contribution of their colleagues Dr G D Sims, Mr R Mera and Mr
S Gnaniah at the National Physical Laboratory, and Mr R Ferguson (now at BAE SYSTEMS), Dr C Messenger and
Mr M Hiley at DERA.

For further information contact:

Dr Bill Broughton
Centre for Materials Measurement and Technology
Tel: 020 8943 6834
Email: [email protected]

Experts

National Physical Laboratory

Queens Road
Teddington
Middlesex
United Kingdom
TW11 0LW

Tel: 020 8977 3222

Fax: 020 8943 6458
Email: [email protected]

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005
MATC(MN)06 Through-Thickness Testing of Polymer Matrix Composites Page 16 of 16

MATC(MN)06
Through-Thickness Testing of Polymer Matrix Composites
May 2001

© Crown Copyright 2001. Reproduced by permission of the Controller of HMSO.

http://midas.npl.co.uk/midas/content/ma06.html 22/06/2005