7-1

C H A P T E R 7

F R A C T U R E A N D F A T I G U E

7.1 FRACTURE 7.1.1 Stress Concentrat ion

7.1.2 Duct i le Fracture

7.1.3 Br i t t le Fracture

7.1.4 E f fects of Temperature & Stra in Rate: Duct i le - to-Br i t t le Trans i t ion

7.1.5 Impact Test ing

7.2 FATIGUE 7.2.1 Fat igue Mechanisms

7.2.2 Fat igue Test ing

7.2.3 Non-Destruct ive Test ing (NDT)

7-2

7.1 FRACTURE

• Fracture is the separation of a piece of material into two or

more pieces when subject to stress.

• Fracture involves the initiation and propagation (i.e.

growth) of a crack, and may occur in several ways, such as:

◆ slow application of external loads (e.g. tensile test)

◆ rapid application of external loads (impact)

◆ cyclic or repeated loading (fatigue)

◆ time-dependent deformation (creep)

7.1.1 Stress Concentration

• The theoretical fracture strength of a material is the stress

required to cause simultaneous breaking of all atomic

bonds across the fracture plane, resulting in the creation of

two new surfaces.

• However, actual fractures in bulk materials occur at stress

levels 10–1,000 times lower than their theoretical fracture

strengths. This is because real materials contain cracks,

which are unable to carry any tensile load; instead, the

applied load is transferred to regions around the cracks.

7-3

• For a long, thin crack, its geometry results in the amplifica-

tion or concentration of the applied stress at its crack tip,

such that the local stress near the tip is much higher that

the average applied stress (Fig. 7.1-1).

Fig. 7.1-1 Thin plate under uniaxial

tension with (a) edge crack of length a, and (b) centre crack of length 2a.

(c) Concentration of stress at the crack tip. (d) Photoelastic fringe pattern showing concentration of stress at a crack tip.

[Note: the closer the lines, the higher the stress]

(d)

• Cracks may be formed in a material during manufacture, or

through wear on the surface. In metals, cracks may also be

introduced during plastic deformation.

• The mechanism by which the crack propagates (grows)

determines whether the fracture is brittle or ductile.

7-4

7.1.2 Brittle Fracture

• Brittle fracture is sudden and catastrophic, with little or no

plastic deformation and low energy absorption (Fig. 7.1-2).

Materials that fail by brittle fracture have low toughness.

(a) (b)

Fig. 7.1-2 (a) Flat brittle fracture surface with little deformation. (b) Stress-strain behaviour of a brittle material. A brittle material exhibits a small strain at

fracture, and absorbs little energy prior to fracture (small area under the curve).

• The separation of the material tends to occurs through the

grains (transgranularly) along specific crystallographic

planes (Fig. 7-1-3a). This process of splitting is called cleavage.

• Cleavage fracture surfaces are relatively flat and shiny, with

a faceted texture (Fig. 7-1-3b & c), as the reflected light changes

with the orientation of cleavage planes in different grains.

• Brittle fracture can also occur along grain boundaries

(intergranularly), whose cohesive strength become lower

when the local chemistry is changed by the segregation of

harmful impurities or brittle second phases (Fig. 7-1-4). 7-5

Fig. 7.1-3 (a) Transgranular/cleavage fracture

– crack propagation through the interior of grains. (b) Typical grainy appearance of

cleavage fracture. Note the lack of macroscopic plastic deformation

(c) Cleavage fracture surface with a faceted (many small, shiny planes) appearance.

Fig. 7.1-4 (a) Intergranular fracture – decohesion (separation) along grain boundaries.

(b) Intergranular fracture surface revealing clearly outlined grains.

7-6

• The yield strength, σy, of a brittle material is higher than its

fracture strength. When the local stress, σlocal, near the

crack tip reaches the fracture strength, the atomic bonds

will break (Fig. 7.1-5). The crack grows rapidly between a pair

of atomic planes, giving rise to an atomically flat surface by

cleavage. Cleavage crack propagation reaches the speed of

sound, which is why brittle materials fail with a bang.

Fig. 7.1-5 Crack propagation by cleavage occurs once the local stress, σlocal, near the crack

tip reaches the theoretical fracture strength. Brittle fracture is sudden and catastrophic.

• Cleavage fracture occurs in brittle ceramics, but is also

common in BCC and some HCP metals at low tempera-

tures (Sec. 7.1-4). True cleavage has not been observed in FCC

metals, due to their low shear stresses required for slip,

numerous slip systems, and the ability to cross-slip.

7-7

7.1.3 Ducti le Fracture

• Ductile fracture is characterized by extensive plastic

deformation prior to failure, which absorbs considerable

energy (Fig. 7.1-6b). Materials that fail by ductile fracture are

considered tough.

• A tensile specimen that has failed by ductile fracture in the

necked region typically exhibits a cup-and-cone

appearance (Fig. 7.1-6a, Fig. 7.1-7a).

(a) (b)

Fig. 7.1-6 (a) Typical cup-and-cone ductile fracture. (b) Stress-strain behaviour of a ductile material. A ductile material exhibits a large strain (a lot of plastic deformation) at fracture, and absorbs a large

amount of energy prior to fracture (large area under the curve).

• To the naked eye, the central region of the fracture surface

looks dull, with an irregular and fibrous appearance (Fig. 7.1-

7a). At high magnifications, “dimples” may be seen on the

fracture surface (Fig. 7.1-7b).

7-8

Fig. 7.1-7 (a) Typical cup-and-cone ductile fracture. (b) Dimples characteristic of ductile fracture. In some dimples, inclusions can be seen.

However, inclusions sometimes fall out of the dimples or they might be obscured by the surrounding matrix although they are still inside the dimples.

• The yield strength, σy, of a ductile material is less than its

fracture strength. Plastic flow will take place in the region

of the material ahead of the crack tip where the local

stress, σlocal, is greater than the σy (Fig. 7.1-8a). The size of this

plastic zone depends on the magnitude of σy: as σy

decreases, the plastic zone increases.

• The plastic flow ahead of the crack tip has the effect of

turning an initially sharp crack into a blunt crack, such that

the stress concentration at the crack tip decreases; i.e. σlocal

is reduced. At the same time, the plastic flow also strain-

hardens the material in the plastic zone (Fig. 7.1-8b). If the

crack becomes blunt enough, σlocal could drop below the

new higher yield strength at the crack tip, effectively

arresting crack growth unless a higher stress is applied. 7-9

• When σlocal is sufficiently high to match or exceed the yield

strength of the strain-hardened material at the crack tip,

crack propagation will continue via continual plastic

deformation ahead of the crack tip.

(a) (b)

Fig. 7.1-8 (a) Stress distribution ahead of the crack tip, showing a zone of plastic deformation where the local stress is higher than the yield strength; (b) plastic

deformation blunts crack tip, lowering local stress, while strain hardening in the plastic zone raises yield strength. Crack propagation will continue if the applied stress is high enough such that the local stress is greater than the strain-hardened yield strength.

• Well-annealed metals do not usually contain cracks, since

diffusion during annealing would have caused existing

cracks to close. However, most engineering alloys contain

second phase particles (inclusions or precipitates), at which

voids nucleate, typically through decohesion of matrix-

particle interface. Such voids grow and coalesce

(join/combine) to become a crack.

7-10

Fig. 7.1-9 Plastic zone ahead of crack tip where voids nucleate and grow.

• Within the plastic zone ahead of the crack tip, void

nucleation, growth and coalescence also occur (Fig. 7.1-9),

and the crack continues to grow to fracture by this

mechanism (Fig. 7.1-10).

Fig. 7.1-10 Stages in ductile fracture caused by void coalescence in the plastic zone.

• The “dimpled” appearance of a ductile fracture surface (Fig.

7.1-7b) shows the remnants of the voids, which were

separated by thin walls of material until these ruptured via

shear (slip) during crack growth. Sometimes, the particles

that nucleated the voids may even remain within the

dimples (Fig. 7.1-7b). 7-11

• The continual plastic deformation during ductile fracture

consumes a lot of energy. The larger the plastic zone

ahead of the crack tip, the more energy is absorbed, and

the tougher the material.

Brittle vs Ductile Fracture

• In a material under stress, there is a competition between

the processes of plastic deformation and fracture. If the

stress required to initiate plastic deformation by the

shearing of atomic planes (i.e. yield stress) is less than the

stress necessary to permanently separate atoms (i.e.

fracture stress), yielding occurs in preference to fracture

and vice versa.

• Ductile fracture is generally preferred because crack

growth is relatively slow and steady, and the presence of

plastic deformation gives warning that fracture is

imminent, allowing preventive measures to be taken.

• Brittle fracture, on the other hand, occurs suddenly and

catastrophically without any warning.

• Ductile fracture is also more desirable because of the

greater amount of energy absorbed prior to fracture.

7-12

7.1.4 Effects of Temperature & Strain Rate:

Ducti le-to-Brittle Transition

• Many materials exhibit a transition from ductile to brittle

modes of fracture as the temperature is lowered.

• Ductile fracture is accompanied by plastic deformation,

which involves dislocation motion. For materials in which

the critical shear stress is high (e.g. BCC metals [see Table 6.1-2,

pg 6-6]), thermal energy helps in overcoming the energy

barrier.

• At low temperatures, thermal activation for dislocation

motion is reduced. The result is that the yield strength

increases as temperature decreases (Fig.7.1-11), such that the

plastic zone at the crack tip shrinks until it becomes so

small that the fracture mechanism changes from void

coalescence to cleavage.

• Such materials also show a dependence on the strain rate

(i.e. how fast deformation is made to occur, which

depends on the loading rate), because thermal activation is

less effective at a faster rate of deformation. A higher strain

rate would raise the yield strength but lower the

elongation at fracture.

7-13

Fig. 7.1-11 Engineering stress-strain behaviour of iron at different temperatures. The strength and ductility of iron (BCC) show a strong temperature dependence.

• In general, slip systems in BCC metals and some HCP

metals become active only when there is sufficient thermal

energy for dislocation motion; such metals exhibit a

ductile-to-brittle transition (Fig. 7.1-12).

Fig. 7.1-12 The effects of temperature on the fracture behaviour of various materials in impact tests.

7-14

7.1.5 Impact Testing

• The ductile-to-brittle transition behaviour of materials may

be ascertained through impact testing. Impact testing may

also be used to compare the relative toughness of

materials.

• An impact test imposes a stringent set of conditions – a

pre-existing notch (crack) and high loading rates (fast

strain rates) – which tend to promote brittle fracture rather

than plastic flow.

• The Charpy test is one of the most widely used impact

tests (Fig. 7.1-13). A notched specimen is positioned at the

base of the machine. A heavy pendulum is released from a

known height, h, striking the specimen on its downward

swing, fracturing it. The pendulum continues its swing,

rising to a maximum height h’, which is lower than h. The

energy absorbed by the fracture is measured by the

difference between h and h’.

• A ductile material will absorb greater impact energy than a

brittle material. This ability of a material to withstand an

impact blow is often referred to as impact toughness.

7-15

Fig. 7.1-13 Schematic diagram of the Charpy impact tests.

• While impact tests are useful for qualitative assessment of

material toughness, the results obtained; namely, the

energy to fracture and the appearance of the fracture

surfaces, are not useful in design problems. Designing for

fracture requires the use of fracture toughness, a material

property derived from fracture mechanics.

7-16

7.2 FATIGUE

• If a component or structure is subjected to repeated stress

cycles, it may fail at stresses well below the ultimate tensile

strength, and often even below the yield strength of the

material.

• The processes leading to this type of failure are known as

fatigue, and it is estimated that fatigue accounts for

approximately 90% of all metallic failures.

• Fatigue fracture resembles brittle fracture in that failure is

sudden and catastrophic, with very little visible plastic

deformation, even in normally ductile metals.

7.2.1 Fatigue Mechanisms

• In fatigue, cracks grow slowly under stresses less than the

yield strength until the cracks become so large that the

remaining cross-sectional area can no longer support the

load, and sudden and catastrophic failure occurs.

• Fatigue failures generally start at the surface (where

bending or torsional stresses are highest). Fatigue cracks

may be pre-existing, or initiated through plastic

deformation at stress raisers (Figs. 7.2-1 & 7.2-2), including rough

surfaces left by tools or grinding. 7-17

Fig. 7.2-1 Fatigue crack initiation at stress raisers.

Fig. 7.2-2 Local plastic deformation under a cyclic stress can roughen a surface, in such a way that the “valleys’ concentrate stress and initiate fatigue cracks.

• Under a cyclic stress (Fig. 7.2-3), tensile stress produces a small

plastic zone at the crack tip, stretching open the crack tip

by an amount δ, creating new surface πδ/2 there (Fig. 7.2-4).

As the tensile stress is removed or reduced, the crack closes

and the new surface folds forward, extending the crack by

πδ/4. The process is repeated on the next tensile cycle, and

so the crack inches forward slowly and steadily.

7-18

Fig. 7.2-3 (a) Sinusoidal, and (b) random variable stress cycles.

(a)

(b) Fig. 7.2-4 (b) Tensile stress

causes plastic deformation at the crack tip, creating

new surface of length πδ/2, which, (c) when folded forwards as the tensile

(c) stress is relaxed, grows the fatigue crack by πδ/4.

(d)

• As the crack grows, the cross-sectional area that supports

the load is decreased and the stress in this section

increases, until the component fails catastrophically in the

next stress cycle due to overload (Fig. 7.2-5). 7-19

Fig. 7.2-5 The stages of fatigue failure: after crack initiation, the crack grows slowly

and steadily until the remaining cross section can no longer support the load, resulting in rapid and catastrophic failure.

• Fatigue fracture surfaces typically exhibit beachmarks (so-

called because they resemble ripples in the sand on a

beach), or clamshell markings (Fig. 7.2-6), which are visible to

the naked eye.

Fig. 7.2-6 Beachmarks or clamshell markings typical of fatigue fracture.

7-20

• Microscopic examination of the beachmarks reveal similar

markings on a finer scale, called striations (Fig. 7.2-7). The

spacing between the striations is a measure of the slow

crack advance per stress/strain cycle.

Fig. 7.2-7 Fatigue fracture striations (crack propagation is from left to right).

• Although fatigue is associated with slip, it is of such a

localized nature that very little overall plastic deformation

is produced during the development of a crack. It is this

lack of visible distortion that makes fatigue cracks difficult

to detect in service prior to final catastrophic failure.

7-21

7.2.2 Fatigue Testing • The resistance of a material to failure by fatigue may be

determined through a fatigue test, in which a specimen is

subjected to a stress cycling through simultaneous bending

and rotation (Fig. 7.2-8).

Fig. 7.2-8 Fatigue testing.

• The stress, in general, varies sinusoidally with time, and is

characterized by a stress amplitude, σa or S, that

alternates about a mean stress, σm (Fig 7.2-9).

σa =

€

σmax - σmin2

; σm =

€

σmax + σmin2

Fig. 7.2-9 A typical stress cycle.

7-22

• From the fatigue test, the stress amplitude, S, measured for

some constant mean stress, σm, is plotted against the

number of cycles to failure, N.

• For some alloys (e.g. steels and titanium alloys), the S-N curve becomes horizontal below a certain level of S (Fig. 7.2-

10). This is called the fatigue limit (or endurance limit). Below this stress amplitude, fatigue failure would not occur regardless of the number of cycles.

Fig. 7.2-10 Stress amplitude, S, versus the number of cycles to fatigue failure, N

for a material that displays a fatigue limit.

• Many nonferrous alloys (e.g. Al, Cu, Mg) do no have a well-defined fatigue limit. The S-N curve continues to slope

downward with increasing N values (Fig. 7.2-11). For such

materials, the fatigue strength is defined as the stress level at which failure would occur for some specified number of

cycles (e.g. 107 cycles). Fatigue life is the number of cycles

to cause failure at a specified stress level. 7-23

Fig. 7.2-11 Stress amplitude, S, versus the number of cycles to fatigue failure, N

for a material that does not display a fatigue limit.

• Apart from the dependence on stress amplitude, fatigue

life is also affected the mean stress, σm, about which the

stress cycle alternates. Increasing the mean stress level has

the effect of lowering the S-N curve to lower S values, thus

leading to a shorter fatigue life (Fig. 7.2-12).

Fig. 7.2-12 Effect of mean stress on S-N fatigue behaviour.

7-24

• The endurance limit / fatigue strength may be correlated

with strength (Fig. 7.2-13), since a stronger material has higher

resistance to slip at the crack tip during each tensile cycle.

Fig. 7.2-13 Relationship between endurance limit / fatigue strength and strength.

• Fatigue life may be improved by selecting a stronger

material, increasing surface hardness (e.g. carburizing in

steels [Fig. 7.2-14]), reducing stress raisers through proper

design and improved surface finish, as well as inducing

compressive surface stresses through heat treatment or

mechanical operations (e.g. shot peening [Fig. 7.2-15]).

7-25

Fig. 7.2-14 Carburized steel gear teeth showing gradual increase in hardness

from core to surface due to increasing carbon content in the steel.

Fig. 7.2-15 In shot peening, tiny steel, glass or ceramic balls shot at high speeds

onto metal surfaces creates plastic deformation and induces compressive stresses on the surface.

7-26

7.2.3 Non-Destructive Testing (NDT) • Since many alloys (e.g. Al, Cu, Mg) do not have endurance

limits, critical engineering components must be inspected

periodically for fatigue cracks, which must not be allowed

to grow to critical sizes for final fracture.

• Several techniques are available to detect surface and

internal flaws during manufacture and in-service (Table 7.2-1).

7-27

8-1

C H A P T E R 8

C O R R O S I O N

8.1 ELECTROCHEMICAL REACTIONS

8.2 FORMS OF CORROSION 8.2.1 Galvanic Corros ion

8.2.2 Di f ferent ia l Aerat ion Corros ion 8.2.3 Other Concentrat ion Cel l s

8.3 PROTECTION AGAINST CORROSION

8-2

• The performance of a material must be evaluated in the

context of its interaction with its service environment. Such

interactions may be mechanical (friction and wear), or

electrochemical (oxidation and corrosion).

• These interactions usually result in the degradation of the

material (i.e. deterioration in mechanical and/or physical

properties, or appearance), leading to huge economic

losses (5% or more of the GDP of industrialized countries),

as well as posing threats to human safety.

NITRICACID

PETROL

(a) Fresh water (b) Salt water

(c) Acids and alkalis (d) Organic solvents

(e) Oxidation (f) UV radiation

Fig. 8.1-1 Materials undergo electrochemical interactions with water, acids and alkalis, organic solvents, oxygen and radiation.

8-3

Corrosion occurs in an aqueous environment, and

together with wear, are the main causes of metal

degradation at ambient temperatures. Generally, corrosion

will occur in air with relative humidity (RH) greater than

60-70%, which forms an invisible thin film of moisture on

the surface of metal.

8.1 ELECTROCHEMICAL REACTIONS

• When immersed in an electrolyte, most metals have

varying degrees of tendency to dissolve into the electrolyte

as positively-charged metal ions, leaving behind the

negatively-charged electrons in the metal (Fig. 8.1-2a).

• The metal, A, now negatively-charged, will attract positive

ions in the electrolyte in its vicinity, such that the metal

ions that have just dissolved in the electrolyte tend to

revert back to solid metal atoms again.

• In the absence of external influences, an equilibrium is

reached in which metal atoms and ions are continually

entering and leaving the electrolyte at the metal surface.

Overall, there is no net change in the system, but the metal

has developed an electrochemical potential.

8-4

• The magnitude of this potential depends on the tendency

of the metal to dissolve into the electrolyte. The stronger

the tendency, the more negative the potential.

• When a different metal, C, is added to the same electrolyte

(Fig. 8.1-2b), it also develops its own potential, depending on

its tendency for metal dissolution.

• If the two metals were connected by a wire (Fig. 8.1-2c), the

excess electrons would try to minimize the potential

difference by distributing themselves around the system.

Electrons flow from the metal with more excess electrons

(more negative potential), A, to the metal with fewer

electrons (less negative potential), C.

Fig. 8.1-2 (a) Metal A at equilibrium with positive metal ions in solution, adjacent to the electrons left behind. (b) Metal C has less tendency to dissolve in the electrolyte, leaving less negative charge in the metal. (c) A wire conncection allows electrons to flow from the more negative to the less negative metal. As electrons flow, metal A becomes less

negative and metal C becomes more negative.

8-5

• Since metal A now becomes less negatively-charged, there is less attraction between the metal and the positive ions in the electrolyte. Some of the metal ions drift away, upsetting the equilibrium around the metal surface and allowing further dissolution of metal A, i.e. A corrodes.

• Metal A is known as the anode, where electrons are produced through the following reaction:

M → Mn++ ne– (anodic/oxidation reaction)

• On the other hand, metal C has gained electrons to become more negatively-charged and its positive ions in the electrolyte are thus more strongly attracted to the metal and may even convert back into solid metal atoms.

• The electrons from A are consumed through various reactions at C, called the cathode, depending on the electrolyte. Some cathodic/reduction reactions are:

Table 8.1-1 Some common cathode reactions for aqueous galvanic cells.

8-6

• In order for corrosion to occur in a system, a potential

difference must exist between different regions, which

must also be electrically connected, such that electrons

produced at the anode can be consumed at the cathode.

An electrolyte must also be present for the conduction of

ions. If any of these elements were missing, the corrosion

reaction would stop.

• The potentials of various metals under standard conditions

have been measured relative to a hydrogen electrode and

are listed in the electrochemical series (Table 8.1-2). A metal

that is lower on the electrochemical series has a greater

tendency to lose electrons and become metal ions in

solution (i.e. corrode).

Table 8.1-2 Standard electrochemical potentials.

8-7

• The potential of a metal depends on many factors, such as

temperature, type of electrolyte, concentration of metal

ions and other species (e.g. oxygen) in the electrolyte, etc.

The electrochemical series was generated under highly

idealized conditions; in most practical situations, the

potential of a metal differs from its value in this series.

• It is thus more useful to refer to the experimental galvanic

series, which is a simple, qualitative ranking of the relative

reactivities of a number of metals and commercial alloys in

a particular environment, e.g. seawater (Table 8.1-3).

Table 8.1-3 The galvanic series in aerated seawater at 25°C.

8-8

8.2 FORMS OF CORROSION

8.2.1 Galvanic Corrosion

• Galvanic corrosion occurs when two metals of different

compositions are electrically connected in an electrolyte.

• The alloy with the more negative potential becomes the

anode and corrodes, while the less negative metal acts as

the cathode. Generally, the greater the potential

difference, the more severe corrosion would be.

• Since the same total number of electrons must flow

through the anodic and cathodic regions during corrosion,

the ratio of anodic to cathodic surface area will affect the

rate of corrosion. Small anodes in contact with large

cathodes will corrode more severely than large anodes in

contact with small cathodes (Fig. 8.2-1).

Fig. 8.2-1 Effects of anode-to-cathode area:

(a) A large cathode (Cu) area consumes a large number of electrons during the cathodic reaction, which would have to be provided by the severe corrosion of the small anode (Fe).

(b) Corrosion is less severe when the ratio of the anode to cathode surface area is large (i.e. large anode–small cathode is favourable).

8-9

• Uniform corrosion is a special case of galvanic corrosion occurring at the microscopic level within the same metal. Even a nominally homogeneous metal surface contains tiny variations in composition and structure, giving rise to microscopic local anodes and cathodes. With time, the locations of the anodic and cathodic regions change randomly so that the result is a uniform loss of metal over the entire exposed surface.

• Alloys containing more than one phase (Sec. 9.1) would similarly undergo microscopic galvanic corrosion (Fig 8.2-2), since each phase has a different composition. Single-phase alloys are generally more resistant to corrosion than multi-phase ones.

Fig. 8.2-2 Microscopic galvanic corrosion in steel (a two-phase alloy).

Fig. 8.2-3 Intergranular corrosion due to compositional differences between grain boundaries and grain interior.

• Galvanic couples may develop at grain boundaries, leading to intergranular corrosion (Fig 8.2-3). Differences in composition and structure due to segregation of impurities to, or the formation of precipitates at grain boundaries could make these regions anodic to the bulk of the grains.

8-10

8.2.2 Differential Aeration Corrosion

• In many instances, oxygen is required at the cathode in

order for corrosion to occur [See Table 8.1-1, cathode reactions 3 and 4].

If there is a difference in oxygen concentration between

two regions, the area of low oxygen concentration will

become the anode, while the region of high oxygen

concentration will act as the cathode.

• Differential aeration is responsible for crevice corrosion

and pitting corrosion (Fig. 8.2-5). In these forms of localized

corrosion (“localized” as opposed to “general” or “uniform” corrosion, which

occurs everywhere on the metal surface), the oxygen in small crevices

and pits are consumed during uniform corrosion. Since

there is restricted exchange of electrolyte from within the

confined spaces of these locations, these regions become

lower in oxygen concentration than the rest of the material

that is exposed to a large volume of electrolyte. As a result,

corrosion occurs preferentially in the crevices and pits.

• Differential aeration is also the mechanism behind

waterline corrosion. Oxygen is able to reach metal near

the waterline than regions further away; thus, the metal

just inside the waterline will become cathodic while the

remaining immersed metal will become anodic (Figs. 8.2-6 & 7).

8-11

(a) (b)

Fig. 8.2-5 (a) Crevice corrosion, and (b) pitting corrosion, due to differential aeration.

Fig. 8.2-6 Waterline corrosion due to falling levels of dissolved oxygen

below the liquid surface.

Fig. 8.2-7 “Waterline” corrosion beneath a drop of water.

8-12

• An environment in which oxygen is severely limited will

tend to reduce corrosion significantly (Fig. 8.2-8), since

oxygen is required for cathodic reactions in all solutions

[Table 8.1-1, cathode reactions 3 and 4].

Fig. 8.2-8 Corrosion is more severe in (a) sandy soil than in (b) clay, due to the permeability of sand to oxygen.

• Corrosion of all forms also increases in severity with

temperature, because all electrochemical reactions are

thermally activated.

8-13

8.2.3 Other Concentration Cells

• A metal that has been cold worked contains a high density

of dislocations. Since dislocations are associated with

higher energy, these highly-stressed, cold-worked regions

in a metal will act as anodes to less-stressed cathodic areas

(Fig. 8.2-9).

Fig. 8.2-9 The regions of a nail that were stressed during fabrication or use

are anodic and will corrode locally.

Fig. 8.2-10 The higher energy of grain boundaries makes them anodic and

more susceptible to corrosion.

• Similarly, grain boundaries, with their higher energy, are

anodic to the bulk of the grains and tend to corrode more

severely (Fig. 8.2-10; see also Sec. 4.7 on Metallography). However, unlike

intergranular corrosion (Sec. 8.2-1), the rate of corrosion at

grain boundaries that do not contain segregated impurities

or precipitates is not significantly higher than the grain

interior, since the potential difference between the

boundaries and interior is only minimal.

8-14

8.3 PROTECTION AGAINST CORROSION

• Corrosion can be retarded or prevented if either the anodic

or cathodic reaction can be stifled.

• Many corrosion problems can be eliminated through

proper design (Fig. 8.3-1) and material selection to minimize

concentration or galvanic differences. Materials that

produce protective oxide layers, e.g. aluminium alloys

(Al2O3) or stainless steels (Cr2O3), may be used instead.

Fig. 8.3-1 Reducing crevice corrosion through improved design.

8-15

• Protective coatings prevent contact between the metal

and electrolyte (e.g. paints, chrome-plating). Furthermore,

if the coating is anodic to the underlying metal (e.g. Zn on

steel or galvanized steel), should the coating become

scratched, the metal will continue to be protected since

the coating will corrode instead (Fig. 8.3-2).

Fig. 8.3-2 Zinc is anodic to steel, such that should the Zn coating become disrupted, the

large anode (Zn) to small cathode (steel) area would ensure continued protection. However, steel is anodic to tin, which would result in a small anode-to-cathode area ratio if the coating

is scratched; this is undesirable since it would cause severe corrosion of the steel.

• Inhibitors may be used in closed systems. These are

chemicals added to the electrolyte that usually form

protective layers either on the surfaces of the anode or

cathode. Other inhibitors may retard the cathodic reaction,

thereby reducing rate of corrosion at the anode.

• Cathodic protection works by “forcing” the corroding

metal to become the cathode instead. One way is to use a

sacrificial anode by electrically connecting a more anodic

metal to the metal to be protected. The more anodic metal

corrodes in preference to the original metal (Fig. 8.3-3 & 4a).

The sacrificial anode would need to be replaced

periodically.

8-16

Fig. 8.3-3 Cathodic protection through sacrificial zinc anodes on the steel hulls of ships.

Fig. 8.3-4 Cathodic protection of underground pipelines using

(a) sacrificial zinc anode, and (b) impressed current.

• Another cathodic protection method similar to the

sacrificial anode method is the use of an impressed

current. Since an anode corrodes when electrons flow

away, the anode can be made cathodic by supplying a

larger counter-current to neutralize the corrosion current

(Fig 8.3-4b). The major advantage of this method is that it is

possible to use non-consumable anodes. Another

advantage is that the supply may be controlled to regulate

the current provided.