PROPERTIES OF MATERIALS

9.2. PROPERTIES OF MATERIALS

Glyn Davies

To understand the behavior of aircraft structures, it is necessary to

appreciate the properties of the basic materials from which they are made.
We will look at the mechanical properties of metals and composites, but cost
also is a factor in choosing the material and the fabrication process. Actually,
the basic cost of materials is a small percentage of the whole (see Figure
9.4), so the cost should not be a main driver if significant weight savings can
be made or if maintenance costs can be reduced with a material whose wear
and tear or fatigue performance is superior. The situation in engines is quite
different where exotic materials, able to withstand high temperatures, for
example, are well worth the extra cost to the operator.

9.2.1. METALS

The most common aircraft materials are alloys of aluminum, titanium, and
steel, all of which would be described as ductile, so some definitions are
needed. Figure 9.5 shows a typical stress/strain curve. The initial portion is
−1 E ,
linear with a slope given by tan where E is Young’s Modulus, or material
stiffness. As the strain increases, eventually the dislocations propagate and
irreversible plasticity occurs. If the material is unloaded it will recover in a
linear fashion as shown where the residual strain of 0.002 has been selected.
The stress from which this unloading occurred is known as the 0.2% proof
stress. It is a convenient measure of the degree of early plasticity and is used
in estimating proof load deformations. The local tangent modulus, E t ,
decreases gradually with strain and ultimately a (tensile) failure occurs at
the ultimate strain ε u which is often used as a measure of the ductility.

Plasticity does lower the potential of a metal to resist high stresses, and most
alloying techniques are aimed at delaying the dislocations. However, there
are significant advantages as well. Clearly most metal-forming techniques
need irreversible plasticity, whether the materials are being pressed,
stretched, spun, or forged. Common mild steel is the most ductile of all
materials and most consumer goods use it still, as does the civil engineering
industry, which relies on cold-rolled sections. Another advantage of ductility
is to render the structure less sensitive to local stress concentrations caused
by discontinuities in geometry such as holes and fillets. If the stress
concentration exceeds the proof stress, then local plasticity will defuse the
stress field and leave behind an advantageous residual stress field. In the
case of cyclic loading and fatigue, the working stress field may be elastic
until a crack is formed, and then crack-tip plasticity can have a beneficial
effect, as discussed later.

Table 9.1 compares the mechanical properties of the three alloys. The values
for any alloy can be enhanced by heat treatment and work hardening, but
this can make the material more brittle. The values quoted in the table are
therefore average values for high-performance aerospace materials. We see
that both the stiffness and the strengths increase as we go from aluminum to
titanium to steel, but surprisingly the specific values (divided by density) do
not vary that much. Titanium is the most expensive and is difficult to
machine, but it does have superior toughness and fatigue properties. The
main reason for using it is that it retains its strength at much higher
temperatures than aluminum. For example, aluminum alloys should not be
used above 175°C, whereas titanium can be used up to 620°C. It is therefore
to be found wherever a firewall is needed between a hot engine and the
primary structure. As a matter of interest, titanium can be used for cheap
manufacture of complex multilevel sheet components using the properties of
superplastic forming which occur as the temperature approaches 1000°C.
The metal flows like a fluid, and large strains can be accommodated without
the sheet necking or creating voids. At the same time, diffusion bonding can
occur between mated surfaces in which the microstructure is as good as the
parent metal. Access doors in the Airbus wings are made this way, for
instance. Titanium is favored for diffusion bonding since for aluminum alloys
the surface oxidizes at the superplastic temperatures.
Figure 9.4. Relative capital cost of airframes and engines.

Steel has the superior strength and stiffness, but this can lead to very thin
sheet at the design stresses, and buckling resistance will be poor. However,
the smallest volume is a positive advantage for large forgings such as
undercarriages where the available space may be limited—particularly for
military aircraft.

Static strength is not the only requirement: fatigue and toughness can be
important, particularly for aircraft where the loading history is cyclic due to
gust maneuvers, pressurization, and landing and takeoff. A modern civil
aircraft may have to be in operation for 25 years, covering 80–120,000 hours
and 20,000 flights. For military aircraft however the figures would be about a
tenth of this. In selecting materials with good fatigue performance the S –N
curve is widely used. Figure 9.6 shows a typical family of these curves.
Figure 9.5. Stress–strain curve for ductile material.

Table 9.1. Comparison of Four Aerospace Metallic Materials

E(GPa) σ (UTS) ε U (%) ρ (Mg/m 3) E/ρ σ /ρ

(Mpa)

Aluminum 71 482 3.0 2.80 25 172

alloy

Titanium 110 1,000 2.5 4.43 25 226

alloy

Steel 207 1,723 7 7.78 26 221

Berylium 304 207 2.0 151 104

alloy

The S –N curve on its own does have limitations, primarily because the
loading for typical aircraft spectrum does not have a constant stress
amplitude, apart from pressure loading in the fuselage, and a way is
therefore needed for estimating the cumulative damage for quite complex
loading histories. A common way of estimating this is the Palmgren–Miner
law (Miner 1954). If the number of cycles is n i at a stress level where failure
would occur at N i cycles, the measure of cumulative damage is n i /N i , and
for a variety of levels the Palmgren–Miner criterion for failure is

This so-called safe life philosophy works quite well but suffers if a
particularly high tensile stress is applied, say early on in life. This could
cause residual compressive stresses in an area of high stress concentration,
and this would positively inhibit subsequent damage creation. The above
cumulative damage law would be pessimistic. The worst alternative likely,
particularly for military aircraft or civil aircraft operating from rough
terrains, would be damage in (say) an area of tensile stress concentration,
and then a high compressive load being applied. This would leave a residual
tensile stress concentration. Because of this uncertain empiricism, the
tendency in all fields is now to adopt a damage-tolerance criterion. This
comes from the original fail-safe ideas, where a structure was designed to be
so redundant that a failure of one component would not be disastrous if the
remaining load paths were adequate. The modern damage-tolerant view says
that there may be a crack which propagates only in a stable fashion, under
the given working stresses, until it reaches a critical length. The operator
therefore has to have an inspection period which will reveal such a crack well
before this length is reached. Experimental calibrations are usually
necessary to convert a local stress field to a crack propagation law, but
theoretical estimates are possible and indeed necessary for any designer.
The first need is to define the conditions for propagation of a crack where
the elastic stress concentration factor at a crack tip is theoretically infinite.
This is most easily done by evaluating the elastic strain energy released
when a crack increases by a small amount and equating this to the energy
needed to create the new crack surface. The following model, greatly
simplified, brings out the physics.
Figure 9.6. Typical S-N Curves for aluminum alloys.

Imagine in Figure 9.7 a crack of length 2 a , in a plate of thickness t , subjected

to an applied stress field σ . The stresses have to flow around the crack,
leaving a zone where the stresses are now much reduced. If we make the
radical assumption that this softened zone is circular and the stresses in it
are now negligible, then the strain energy lost is given by

Figure 9.7. Crack propagation.

A perturbation of the crack by δa will release further energy, so, taking an
increment of the above, we have

The critical propagation threshold will be given when this is equal to the
energy needed to create the new surface t · δa , i.e., G c · t · δa , where G c
is the critical energy release rate for the material and is a material property
akin to surface tension. Equating the two gives the critical stress as

so the critical stress varies like a − 1/2 , which explains why there is a rapid
decrease in strength as a crack grows. The stress field around the crack has
been oversimplified in this demonstration. One would expect some unknown
coefficient β to be present in the above formula. Catalogues are available of
β or the equivalent stress intensity factor (explained next) for various cases
such as cracks at the edges of holes or near a free edge or a stiffener—in
fact, any situation where the stress field differs from that around a small
crack in a large plate (Rooke and Cartwright 1976).

In the case of ductile metals, there is also a very efficient energy absorber in
the form of local crack tip plasticity, which will leave behind a thin skin of
residual stress as the crack propagates. This effect is included in the
effective value of G c found from coupon tests and is typically a factor of 10
greater than the energy needed simply to create a new surface.

An alternative explanation of the above uses the concept of stress intensity

factor, which is a measure of the strength of the singular stress field at the
crack tip. The stress intensity factor may be written as

so we see that K 2 = EG C. The concept of the strength of a stress field

which is infinite is not as obvious as the energy release explanation, but the
idea has been around longer. Rooke and Cartwright (1976) is in fact a
compendium of stress intensity factors. There are also a large number of
software tools which will evaluate stress intensity factors for cracks of any
size embedded in various stress concentrations.

All this discussion centers around a crack opening mode, but there are two
other shear modes which have different critical values. Figure 9.8 shows all
three possible modes. When we come to laminated composites, the possibility
of cracks occurring as a delamination between laminae is a weakness of
these materials, since a typical matrix will be brittle in nature. It is usually
found that the critical energy release rate, or fracture toughness, almost
doubles as the mode changes from I to II and then to III.

The energy release or stress intensity factor measure turns out to be useful
in formulating a cyclic fatigue law for the damage tolerance approach which
is able to cope with complex loading histories. The simple Paris et al. (1972)
law says that the cyclic crack growth rate is given by

where Δ K is the amplitude of the stress intensity factor ( K max −K min) and C
and n are constants to be found for the material. This law is not strictly true
for the initiation stage, nor for the final stages when the growth becomes
unstable, so several modifications have been made. The most comprehensive
(Foreman 1967) is

where K 1 C is the critical value for the initiation growth under the static
loading. A further mechanism for retarding the crack growth occurs if there
is an overload causing an increase in crack tip plasticity. Thus, crack growth
curves can be generated over a full range of stress and cycle numbers and
hence an accurate inspection interval selected.
Figure 9.8. Three crack modes.

9.2.2. COMPOSITES

Composite aerospace structures have been around for more than 30 years.
We shall mention only the most common type, where very fine fibers are
embedded in a resin, forming unidirectional laminae (prepreg). The fibres are
then laid up in a chosen stacking sequence to suit the designer. The most
popular combination has been carbon fibers in an epoxy resin which is cured
under a raised temperature and pressure (thermosets). Carbon is a brittle
material which does not suffer dislocations but is sensitive to the most minor
scratch or flaw. The fibers are only microns in diameter, and it has been
found that a typical unflawed length is of order 15 to 30 diameters. These
flaws in adjacent fibers will never coincide, so the resin is well able to provide
a load path bypassing the fiber fracture. This is the basis for laminated
composite structures.

Table 9.2 shows the values for strength and stiffness for pure fibres with no
flaws and for unidirectional composites having fiber volume fractions of 60%.
The previous metal figures are included in the table for comparison. Again,
these values are average for aerospace materials. The strengths are higher
than the metal figures and the stiffnesses comparable. Where composites
score, of course, is in their low density. The specific stiffness and strength
figures are much higher, with the exception of the glass stiffness. Of the
three composites, the highest value of specific stiffness is for carbon
composite and is better by a factor of 4 than that for any metal. It is stiffness
which controls the buckling performance of structures, and more than 70% of
the primary aircraft structure is designed by buckling. High-performance
military aircraft need stiffness for their aeroelastic margins, and few could
have been built without using carbon composites.

We have mentioned the poor transverse strength perpendicular to the fibers,

and for most structures where the stress field is not everywhere
unidirectional the designers will choose to assemble each ply into a laminate
with a stacking sequence carefully chosen for the local stress field. Lay-ups
vary from quasi-isotropic (+45, −45,0,90°) ns to cross-ply (0, 90) ns , and angle
ply (+45, −45) ns . Subscript n denotes the number of sublaminates, and
subscript s denotes symmetrical by using a mirror image repeated.
Symmetrical lay-ups are almost universal since they do not distort when
curing. Clever nonsymmetrical combinations have been used for aeroelastic
tailoring where wing flexure can be accompanied by beneficial twisting,
which would not happen in homogenous metals.

Table 9.2. Mechanical Properties of Metallic and Polymer Composites

E(GPa) σ (UTS) ε ρ (Mg/m 3) E/ρ σ /ρ

(Mpa) U (%)

Aluminum 71 482 3.0 2.80 25 172

Titanium 110 1,000 2.5 4.43 25 226

Steel 207 1,723 7 7.78 26 221

E glass 70 2,200 2.54 27.6 866

Kevlar 130 2,900 1.45 90 2,000

(aramid)

Carbon 380 2,700 1.86 204 1,450

E glass* 40 840 1.80 22.2 467

Kevlar* 82 1,500 1.39 59 1,080

Carbon* 200 1,500 1.55 129 970

 *Denotes fiber volume fraction of 60%.

The weak transverse strength has been mentioned, and it is equally true for
the through-thickness direction when the individual laminae can easily
delaminate by failure of the resin matrix. This is the Achilles’ heel of
laminated composite structures, and currently several solutions to this
problem are being evaluated. Textile technology such as weaving or stitching
is possible for the dry fibers, which are then impregnated by infusion of the
resin. This is also a cheap form of structure, but strengths will be down. Z-
pinning is an option for laminated composites already cured. Here small pins
of carbon, aramid, or metal are driven dynamically into the laminate from the
outer surface. Some in-plane strength is sacrificed by the process, but the
through-thickness strength and fracture toughness increase dramatically.
The other alternative is to design a structure which has no through-thickness
stresses, but this is virtually impossible near any 3D features such as joints
and panel stiffeners. Also, low-velocity impact damage is a threat which can
cause delamination and some 70% loss in the residual compressive strength.

Thermoplastic resins are an alternative which does not require a chemical

change during manufacture, just application of heat to a flat laminated sheet.
At the moment these appear to be too expensive, but they are much quicker
to mass produce since a thermoset composite spends 1 1/2–2 hours in an
autoclave. Their properties are better, particularly the through-thickness
fracture toughness where the thermoplastic matrix exhibits modest ductility.
The Airbus 380, for example, uses thermoplastic in the leading edge, and
these must withstand bird-strike without excessive losses in stiffness or
strength.

Composite structures do have several other virtues. Their fatigue life is

exceptional. Corrosion is not much of a problem, but beware moisture
absorption in the matrix. It is also easier to design a stealthy aircraft with a
small radar cross-section using composites than it is in metallic structures.
The coefficient of expansion for carbon is virtually zero. The high-
temperature performance (strength and creep) for thermoset materials is
better than for aluminum alloys.
Finally, the structural properties of stiffness and strength of laminated
composite structures are not formed until the manufacturing process is
complete. They can be theoretically deduced from the unidirectional lamina
properties, knowing the stacking sequence, but this is beyond the scope of
this book. Several texts on composite structures are available (Hull 1981;
Agarwal and Broutman 1990; Niu 1992).

9.2.3. SMART MATERIALS

Now that we have discussed composite structures, it is convenient to discuss

the possibility of smart materials and structures, even if only to dispel some
of the myths surrounding this high-profile name. There are two distinct forms
of smart structures. First are the health and usage monitoring systems
(HUMS), in which optic glass fibers are embedded in composite structures or
stuck on metallic. These can measure strain, temperature, moisture, and
fracture. The attraction of HUMS,

in which the signals are optical changes, is that the sensors are cheap and
can be used extensively in arrays and the data can be processed readily to
give the history and current state. The diameter of optic fibers is much
larger than that of carbon, and some weakening of a composite is possible.
Airlines have yet to be convinced. However, the same concept can be used to
monitor manufacturing processes, and industry has warmed to this aspect.

The second form of smart structure is the use of active materials, such as
piezoelectric and piezoceramic, which will strain under an applied voltage
gradient. They can also produce an electrical potential gradient when
strained. The strains are small, but design studies show it is feasible to have
a control surface, such as a leading or trailing edge, sufficiently deformed to
produce a respectable aerodynamic moment. Much larger strains can be
produced in shape memory alloys, but these essentially involve a change of
state which is not immediately reversible and thus would not be suitable for
conventional control. By having a feedback control it is possible to use smart
materials to suppress small strains, e.g., high-frequency (low-amplitude)
noise in fuselages and cockpits. Flutter suppression is also possible since it
is only necessary to inhibit the onset of flutter when the strains are small
and simply prevent the amplitudes growing. Smart materials have already
been used to dampen out and prevent growth of vibrations of large orbiting
satellite telescopes where there is no aerodynamic damping and precision of
the dish surface is crucial. Smart materials do certainly have a future, but the
world is waiting for materials which can produce strains of the order of 0.1%
at least. At the moment metals and ceramics will not do this, but there are
some electroactive polymers which can be persuaded to undergo very large
strains by an internally seeded liquid being forced to move under a voltage.
These could be called “artificial muscles,” and some applications are being
developed for moving surfaces on satellites and extraterrestrial vehicles
where conventional controls are at risk because the lubrication necessary
ceases to function at extremely low temperatures.

Citation
EXPORT
Mark Davies: Standard Handbook for Aeronautical and Astronautical Engineers.
PROPERTIES OF MATERIALS, Chapter (McGraw-Hill Professional, 2003),
AccessEngineering

© 2012 The McGraw-Hill Companies. All rights reserved.

Customer Privacy Notice. Any use is subject to the Terms of Use, Privacy Notice and
copyright information.
For further information about this site, contact us.

Designed and built using SIPP2 by Semantico.

This product incorporates part of the open source Protégé system. Protégé is
available at http://protege.stanford.edu//