MATERIAL TECHNOLOGY : SKMM 3622

ASSIGNMENT 2

TITLE:

FATIGUE LIFE TESTING FOR CARBON FIBER

NAME MUHAMMAD FADHIL BIN JAMALUDIN

MATRIC NO. A14KM0121

LECTURER DR ENGKU MOHAMMAD NAZIM BIN ENGKU ABU BAKAR

SECTION 02

DATE OF 15TH DECEMBER 2016

SUBMISSION
CONTENT

1.0 INTRODUCTION........................................................................................................1

1.1 APPLICATION OF CARBON FIBER.................................................................1

1.2 PROPERTIES OF CARBON FIBER....................................................................1

1.3 PROCESSING METHOD OF CARBON FIBER..................................................3

2.0 TESTING METHOD.....................................................................................................5

2.1 MATERIAL PROCESSING...................................................................................5

2.2 FATIGUE TESTING..............................................................................................6

3.0 RESULT AND DISCUSSION......................................................................................7

4.0 CONCLUSION............................................................................................................14

5.0 REFERENCES............................................................................................................14
1.0 INTRODUCTION
Carbon fiber reinforced polymers (CFRPs), or more commonly referred to as just
carbon fiber is a rapidly expanding field in the engineering design world, being used
everywhere from the structures of spacecraft down to high end sporting equipment such
as cycling. To produce a carbon fiber, the carbon atoms are bonded together in crystals
that are more or less aligned parallel to the long axis of the fiber as the crystal alignment
gives the fiber high strength-to-volume ratio (making it strong for its size). Several
thousand carbon fibers are bundled together to form a tow, which may be used by itself or
woven into a fabric.

1.1 Application of Carbon Fiber

Every day, a new application is found for carbon fiber. What started out forty years
ago as a highly exotic material is now a part of our everyday lives. These thin filaments, a
tenth the thickness of a human hair, are now available in a wide range of useful forms.
The fibers are bundled, woven and shaped into tubes and sheets (up to thick) for
construction purposes, supplied as cloth for molding, or just regular thread for filament
winding.

Carbon fiber has gone to the moon on spacecraft, but it is also used widely in aircraft
components and structures, where its superior strength to weight ratio far exceeds that of
any metal. 30% of all carbon fiber is used in the aerospace industry. From helicopters to
gliders, fighter jets to microlights, carbon fiber is playing its part, increasing range and
simplifying maintenance. The applications in the military are very wide ranging from
planes and missiles to protective helmets, providing strengthening and weight reduction
across all military equipment. It takes energy to move weight whether it is a soldiers
personal gear or a field hospital, and weight saved means more weight moved per gallon
of gas.

1.2 Properties of the Carbon Fiber

1.2.1 High strength to weight ratio

Strength of a material is the force per unit area at failure, divided by its density. Any
material that is strong and light has a favourable strength/weight ratio. Materials such as
aluminium, titanium, magnesium, Carbon and glass fiber, high strength steel alloys all
have good strength to weight ratios. Because of the way the crystals of carbon fibre orient
in long flat ribbon or narrow sheets of honeycomb crystals, the strength is higher running
lengthwise than across the fibre. That is why designers of carbon fibre objects specify the
direction the fibre should be laid to maximize strength and rigidity in a specific direction.
The fibre being aligned with the direction of greatest stress.

1
1.2.2 Very rigid

Rigidity or stiffness of a material is measured by its Young Modulus and measures

how much a material deflects under stress. Carbon fiber reinforced plastic is over 4 times
stiffer than Glass reinforced plastic, almost 20 times more than pine, 2.5 times greater
than aluminium.

1.2.3 Corrosion resistant and chemically stable

Although carbon fiber themselves do not deteriorate measurably, Epoxy is sensitive

to sunlight and needs to be protected. Other matrices (whatever the carbon fiber is
embedded in) might also be reactive. Composites made from carbon fiber must either be
made with UV resistant epoxy (uncommon), or covered with a UV resistant finish such as
varnishes.

1.2.4 High fatigue resistance

Resistance to Fatigue in Carbon Fiber Composites is good. However when carbon

fiber fails it usually fails catastrophically without significant exterior signs to announce
its imminent failure. Damage in tensile fatigue is seen as reduction in stiffness with larger
numbers of stress cycles unless the temperature is high. The orientation of the fibers and
the different fiber layer orientation, have a great deal of influence on how a composite
will resist fatigue.

1.2.5 Good tensile strength

The ultimate tensile strength is the maximum stress that a material can withstand
while being stretched or pulled before necking, or failing.

1.2.6 Non flammable

Carbon fiber can be be made to feel quite soft to the hand and can be made into or
more often integrated into protective clothing for firefighting. Nickel coated fiber is an
example. Because carbon fiber is also chemically very inert, it can be used where there is
fire combined with corrosive agents. Carbon Fiber Blanket used as welding protection.

1.2.7 Low coefficient of thermal expansion

Low Coefficient of Thermal expansion makes carbon fiber suitable for applications
where small movements can be critical. Telescope and other optical machinery is one
such application.

2
1.3 Processing method of carbon fiber

The raw material used to make carbon fiber is called the precursor. About 90% of the
carbon fibers produced are made from polyacrylonitrile (PAN). The remaining 10% are
made from rayon or petroleum pitch. All of these materials are organic polymers,
characterized by long strings of molecules bound together by carbon atoms. The process
for making carbon fibers is part chemical and part mechanical. The precursor is drawn
into long strands or fibers and then heated to a very high temperature with-out allowing it
to come in contact with oxygen. Without oxygen, the fiber cannot burn. Instead, the high
temperature causes the atoms in the fiber to vibrate violently until most of the non-carbon
atoms are expelled. This process is called carbonization and leaves a fiber composed of
long, tightly inter-locked chains of carbon atoms with only a few non-carbon atoms
remaining. The sequence operation to form carbon fiber from PAN is as below:

Spinning:

Acrylonitrile plastic powder is mixed with another plastic, like methyl acrylate or
methyl methacrylate, and is reacted with a catalyst in a conventional suspension or
solution polymerization process to form a polyacrylonitrile plastic.

The plastic is then spun into fibers using one of several different methods. In some
methods, the plastic is mixed with certain chemicals and pumped through tiny jets into a
chemical bath or quench chamber where the plastic coagulates and solidifies into fibers.
This is similar to the process used to form polyacrylic textile fibers. In other methods, the
plastic mixture is heated and pumped through tiny jets into a chamber where the solvents
evaporate leaving a solid fiber. The spinning step is important because the internal atomic
structure of the fiber is formed during this process.

The fibers are then washed and stretched to the desired fiber diameter. The stretching
helps align the molecules within the fiber and provide the basis for the formation of the
tightly bonded carbon crystals after carbonization.

Stabilizing:

Before the fibers are carbonized, they need to be chemically altered to convert their
linear atomic bonding to a more thermally stable ladder bonding. This is accomplished by
heating the fibers in air to about 390-590 F (200-300 C) for 30-120 minutes. This
causes the fibers to pick up oxygen molecules from the air and rearrange their atomic
bonding pattern. The stabilizing chemical reactions are complex and involve several steps,
some of which occur simultaneously. They also generate their own heat, which must be
controlled to avoid overheating the fibers. Commercially, the stabilization process uses a

3
variety of equipment and techniques. In some processes, the fibers are drawn through a
series of heated chambers. In others, the fibers pass over hot rollers and through beds of
loose materials held in suspension by a flow of hot air. Some processes use heated air
mixed with certain gases that chemically accelerate the stabilization.

Carbonizing:

Once the fibers are stabilized, they are heated to a temperature of about 1,830-5,500
F (1,000-3,000 C) for several minutes in a furnace filled with a gas mixture that does not
contain oxygen. The lack of oxygen prevents the fibers from burning in the very high
temperatures. The gas pressure inside the furnace is kept higher than the outside air
pressure and the points where the fibers enter and exit the furnace are sealed to keep
oxygen from entering. As the fibers are heated, they begin to lose their non-carbon atoms,
plus a few carbon atoms, in the form of various gases including water vapor, ammonia,
carbon monoxide, carbon dioxide, hydrogen, nitrogen, and others. As the non-carbon
atoms are expelled, the remaining carbon atoms form tightly bonded carbon crystals that
are aligned more or less parallel to the long axis of the fiber. In some processes, two
furnaces operating at two different temperatures are used to better control the rate of
heating during carbonization.

Surface treating:

After carbonizing, the fibers have a surface that does not bond well with the epoxies
and other materials used in composite materials. To give the fibers better bonding
properties, their surface is slightly oxidized. The addition of oxygen atoms to the surface
provides better chemical bonding properties and also etches and roughens the surface for
better mechanical bonding properties. Oxidation can be achieved by immersing the fibers
in various gases such as air, carbon dioxide, or ozone; or in various liquids such as
sodium hypochlorite or nitric acid. The fibers can also be coated electrolytically by
making the fibers the positive terminal in a bath filled with various electrically
conductive materials. The surface treatment process must be carefully controlled to avoid
forming tiny surface defects, such as pits, which could cause fiber failure.

Sizing:

After the surface treatment, the fibers are coated to protect them from damage during
winding or weaving. This process is called sizing. Coating materials are chosen to be
compatible with the adhesive used to form composite materials. Typical coating materials
include epoxy, polyester, nylon, urethane, and others. The coated fibers are wound onto
cylinders called bobbins. The bobbins are loaded into a spinning machine and the fibers
are twisted into yarns of various sizes.

4
2.0 TESTING METHOD
Structural components are usually subjected to complex fatigue load histories
characterized by changes in the amplitude, stress ratio (R), frequency and waveform of
the cycling stresses. The fatigue behaviour of composites, for example, has been shown to
be highly dependent on the stress ratio, R. This project is to present fatigue behaviour of a
balanced woven carbon/epoxy composite. The stress ratio effect on fatigue life is
modelled based on experimental SN curves performed with R values from 1 to +0.4, at
room temperature. The damage process was monitored in terms of the stiffness loss
correlated with the failure sites and the mechanisms analysed using microscopy. A series
of sequence block tests were performed using, for each sequence block, different stress
ratios corresponding to the constant amplitude fatigue tests. For each block loading test
the fatigue life is estimated based on the constant stress amplitude SN curves using
Palmgren Miners law and compared with experimental results.

2.1 Material processing

Composite laminate sheets were manufactured using 12 balanced woven

bi-directional layers of carbon fibers (196 g/m2 ) all of them with the same orientation
0/90, and an epoxy resin matrix.

The geometry of the specimen for static and fatigue test as below;

With R > 0 With R < 0

iii. L =160 mm i. L =100 mm

iv. Lo = 80 mm ii. Lo = 25 mm

5
 Figure 2: Geometry and dimension

2.2 Fatigue Testing

Fibers and resin were hand placed in a mould subjected to low level compression.
The mould was then put into a vacuum bag during 8 h for room temperature curing.
Figure 1 is a schematic view of the manufacturing process. Airtech commercial vacuum
consumables, supplied from SP, were used: peel plies Stitch Ply A in Nylon 6-6, release
film WL3600 in polypropylene, breather Econoweave 1010 W produced with polyester
fibre and vacuum bag films WL7400 in PA6. The fibre weight fraction was 0.66 and the
average plate thickness was 3 mm. The quality control of the plates was by visual
inspections in terms of finish surface and void content.

The specimens used in the tensile, compressive and fatigue tests were cut from these
plates and aligned with one of the fiber direction. The geometry and dimensions of the
specimens are shown in Fig. 2. The specimens used in compressive tests and fatigue with
R < 0 were smaller to escape buckling effects. Few compressive tests were carried out
with two strain gauges bonded onto both sides of the specimen in order to check the
occurrence of buckling. The static tests were performed in tension using an
electromechanical machine (Instron model 4206) according ISO standard, with a strain
rate of 5 mm/s. The compressive tests were performed in a servohydraulic machine
(Instron model 1341) at a strain rates of 0.05 mm/s, 5 mm/s and 50 mm/s, according to
the recommendations of ASTM D 3410/D 3410 M Standard. Four specimens were tested
for each condition. The fatigue tests were carried out at room temperature in the same
servohydraulic Instron machine using a sinusoidal wave load at constant amplitude and a
frequency of 10 Hz. One specimen loaded with the higher displacement was instrumented
with thermocouples in order to analyze the increase of temperature during the fatigue
tests. No significant variation was observed and, consequently, the other fatigue
specimens were not instrumented. Six series of fatigue tests were performed using the
stress ratios: 1, 0.5, 0.25, 0.05, 0.2 and 0.4, respectively.

6
3.0 RESULT AND DISCUSSION
The static properties of the laminated composites were obtained by tensile and
compressive tests. Two typical stressstrain curves for both conditions are plotted in Fig.
3. It can be seen that for the tensile tests, Fig. 3a, the curves are practically linear until
brittle failure. For compressive tests, Fig. 3b, nonlinear behaviour occurs from the
beginning.

Figure 3a: Tensile stress versus strain

Figure 3b: Compressive stress versus strain

In fact some studies reveal that the nonlinear elastic behaviour, in both tension and
compression ranges, is due to viscoelastic behaviour of the matrix and the nonlinear
behaviour of the carbon fibers . The material non-linearity of carbon fibers are a result of
their specific microstructure. However, as the tensile tests shows, this material presents
linear-elastic behaviour. Therefore the nonlinear elastic behaviour observed in
compression tests must be due other explanations. The hypothesis of the general buckling

7
occurrence was discarded by compressive tests with two strain gauges bonded onto both
sides of the specimen.

Figure 4: Fiber distribution along the longitudinal direction of speciment

Fig. 4 shows that fibers are not perfectly aligned with longitudinal direction in all
specimens. While in tension this misalignment tends to decreases with load increasing, in
compression tests the opposite phenomenon is expected and consequent microbuckling
tends to occur. The increase of the load during the tests promotes an extension of this
damage mechanism, which results in a progressive lose of specimen stiffness. Failure
mechanisms will be observed later and are in agreement with this explanation.

Figure 5: Compressive load-displacement curves

In order to better understand material nonlinearity in compressive conditions, Fig. 5

shows typical load-displacement curves for strain rates of 0.5 mm/s, 5 mm/s and 50 mm/s,

8
respectively. This effect is more pronounced for lower strain rates (0.5 mm/s), where
nonlinearity is strongly observed from the beginning. It can also be seen that
loaddisplacement curves for strain rates of 5 mm/s and 50 mm/s are very similar and the
nonlinearity is less significant than for 0.5 mm/s. Table 1 presents all the experimental
results obtained from the static tests plus average values and the respective standard
deviations.

It is possible to observe that ultimate strength for a strain rate of 5 mm/s, defined as
the stress at peak load, for the tensile tests, is around 69% higher than for the compressive
tests. On the other hand, the difference between the compressive strengths observed for
strain rates of 0.5 mm/s and 50 mm/s is about 7%. The compressive strength increases
with the increasing in strain rate. The lower compressive strength of the composite was
expected and the differences observed from the tensile values are very close to the
literature. The fatigue tests were carried out at constant stress amplitude for the six stress
ratio R (R = minimum stress/maximum stress), and the results are plotted in Fig. 6a and
6b in terms of stress amplitude versus number of cycles to failure and maximum stress
versus number of cycles to failure, respectively.

Figure 6a: Stress amplitude versus cycle Figure 6b: Maximum stress versus cycle
9
 The present results confirm the trend reported in literature that the fatigue strength in
terms of maximum stress increases with R values, even for negative values. The effect of
mean stress observed in this case is very significant. The negative stress ratio tests present
a drastic decrease of fatigue strength as a consequence of the lower compressive strength
of the laminate composites and different main failure mechanisms. When fatigue strength
is quantified in terms of the stress amplitude it decreases with the increasing of stress
ratio for positive values. However for negative R values a decrease of fatigue strength
was also observed as a consequence of the much lower ultimate compressive strength
associated with the failure mechanisms. The tensile and compressive parts of the stress
cycle do not contribute equally to the composites damage. The compressive loads
produce higher levels of damage and consequently reduce significantly the fatigue life.

Figure 7:Failure following fatigue test: (a) R=0.05, (b) R=-0.05

Fig. 7 presents optical photos showing the upper and lateral views of the main
significant sites and failure mechanisms observed during the tests. The failure mode
observed for tensiontension fatigue tests depends on the magnitude of the applied stress.
For relatively high stress levels, close to the ultimate tensile strength, short delaminations
were observed before final fiber breakage, while for lower stress levels long
delaminations occurred over an extended period of time. On the other hand, for
tensioncompression fatigue tests, the failure starts with the breakage of the fibers, as a
consequence of microbuckling and kinking of the fibers aligned with the load direction,
followed by cracking parallel to the load direction. The stiffness modulus was measured
during fatigue tests. The peak load and displacement were monitored and used to
calculate the stiffness modulus (E), defined by the ratio axial stress/axial strain.

10
 Figure 8: E/E0 versus N/Nf for test different value of R

Fig. 8 plots E/E0 versus N/Nf, for different tests with the six R values, where E0 is the
initial value of E, N the current number of cycles and Nf is the number of cycles to failure.
The results show in all cases a slight drop in stiffness modulus during the first fatigue
cycles, followed by a region with a stable decrease of E/E0 until the final failure. The
slope and shape of the plot in the stable region appears to be not dependent on both the R
value and the stress amplitude. The results for different loads and R values show lower
differences than the scatter of results usually observed.

Figure 9:(a)stress amplitude-mean stress dagram, (b)correlation of graph (a)

11
 The effect of the mean stress on the fatigue strength in terms of the stress amplitude
is presented in Fig. 9a for fixed cyclic lives of 105 and 106 cycles. It is at once evident that,
in terms of the stress amplitude, the fatigue strength losses for low mean stresses
(negative R values). This plot also shows that the linear relationship of the Goodman Law
between stress amplitude and mean stress does not fit the experimental results for low
mean stresses. A quadratic equation, including a linear term to account the loss of fatigue
strength for negative stress ratios, is in better agreement with experimental results plotted
in Fig. 10b) as the stress amplitude against the dimensionless parameter m / UTS . The
fitted equation is: (where a is the stress amplitude, f is the stress amplitude
strength for reverse loadings, m is the mean stress, UTS is the ultimate tensile
strength)
2

a f A m B m
UTS UTS

A and B are constants dependent of the material and fatigue life. The fitted values are:
f = 108.35 MPa, A = 358.36 MPa, B = 468.08 MPa for 105 and f = 102.07 MPa, A =
309.95 MPa and B = 412.09 MPa for 106 cycles.

Figure 10: Block loading sequence Figure 11: experimental and estimated test result

A series of sequence block tests were performed, using for each, sequence blocks
with different stress ratios corresponding to the constant amplitude fatigue tests. Fig. 10
shows schematically an example of block loading sequence. Table 2 presents the
sequence parameters and fatigue life for each test. For valid variable amplitude loading
(VAL) test, the sequence must be repeated at least 510 times in order to achieve a
service-like load mixing. In the present work the sequence was repeated at least eight
times. For each variable stress amplitude test, the fatigue life is estimated using constant
stress amplitude SN curves using linear damage accumulation PalmgrenMiners Law

12
and then compared to the experimental values. The comparison of estimated and
experimental results is presented in Fig. 11, where are also plotted two curves
corresponding to a tolerance range three times higher or lower then the estimated against
the experimental values. Estimated points are within the tolerance range and the average
ratio between predictions and experimental lives are close to unit. Therefore, it can be
concluded that the method is accurate enough for predictive fatigue design. The tests
were carried out under variable block loading. However, in practice the loading has a
random character. The comparison of results obtained for other materials under same
spectrum shape, but once applied with blocked loads and once with randomly distributed
loads shows a lower fatigue life for random loading, if the sequence length of the
spectrum is too long and the repetition is too low. This aspect should be considered for
the tested materials in future investigations.

13
4.0 CONCLUSION
Based on this mechanical testing for fatigue life evaluation for carbon fiber
composite, there is several things that can be conclude.

1. The tensile ultimate strength obtained for woven balanced bidirectional laminated
carbon/epoxy composites is significantly higher which is about 69% than the compressive
ultimate strength. Under tensile loading the composites exhibit brittle behaviour, while in
compressive tests some nonlinear behaviour was observed, which may be consequence of
progressive fiber buckling.

2. Fatigue behaviour studied for stress ratios ranging from 0.5 to 0.4 shows a strong
dependency on dependent of stress ratio. For negative stress ratios a drastic decay of the
fatigue strength was observed as a consequence of much lower ultimate compressive
strengths and fiber buckling. A quadratic equation, including a linear term to account the
loss of fatigue strength for negative stress ratios, fitted as a function of the stress rate
shows good agreement with experimental results.

3. The fatigue lives resulting from block loading tests and estimated lives based on
PalmgrenMiners Law were found to be in a good agreement.

5.0 REFERENCES
Carbon fibers. (2016). En.wikipedia.org. Retrieved 10 December 2016, from
https://en.wikipedia.org/wiki/Carbon_fibers
Johnson, T. (2016). Applications of Carbon Fiber. About.com Money. Retrieved 10
December 2016, from http://composite.about.com/
Carbon Fiber Properties. (2016). Christinedemerchant.com. Retrieved 10 December
2016, from http://www.christinedemerchant.com/carboncharacteristics.html

How is it made? - Zoltek Carbon Fiber. (2016). ZOLTEK Carbon Fiber. Retrieved 10
December 2016, from http://zoltek.com/carbonfiber/how-is-it-made/

Reis et al. (2008). Composite science and technology on Fatigue Life Evaluation.
Retrieved 11 December 2016 from www.elsevier.com/locate/compscitech

14