Failure Analysis Guide

How to recognize it (fracture surface, microstructure, SEM)

• The fracture shows large plastic deformation such as necking or bending. A round bar pulled in tension gives a cup-and-cone fracture with shear lips slanted at about 45° at the edge.
• In the scanning electron microscope (SEM), the surface is covered with tiny cups called dimples. Tension produces round, equiaxed dimples; shear produces dimples elongated in one direction.

Tests to confirm it

• Measure the dimensions and hardness of the broken part and compare the applied load with the part’s strength (tensile strength or 0.2% proof stress).
• Run tensile and hardness tests on material from the same lot to confirm it met the specified strength, ruling out mixed-up material or faulty heat treatment.
• Use stress analysis to check whether a load beyond the design assumptions was applied.

Common mix-ups

• The final fracture zone of a fatigue failure that has grown large is also ductile and dimpled. Do not conclude overload from the last zone alone; look for beach marks or striations around the origin.

Prevention

Material

Choose a material and heat treatment with the required strength and ductility.

Design

Provide a margin against the largest expected load and avoid stress concentrations such as abrupt changes in section.

Environment and use

Review how the part is used, and add protective devices that stop overloads.

Manufacturing and maintenance

Prevent low strength from mixed-up material or faulty heat treatment through incoming and in-process inspection.

Clues that score for this mode

Dimples (tiny cups) all over the surface in the SEM +3Large (necking, bending or stretching) +3Impact or a sudden large load +2Immediately, during manufacture or at the first large load +2Slow bulging or stretching over a long time +1Steady load (self-weight, internal pressure, bolt preload…) +1Ordinary carbon or low-alloy steel +1Almost none (the pieces fit together closely) −3Little external load (residual stress from welding, bending or press fits) −3Ceramic or glass −3Beach marks (shell-like bands spreading from the origin) −2Flat, shiny facets (cleavage) or river patterns −2Striations (fine parallel lines) in the SEM −2Intergranular surface showing the shapes of grains −2Material was lost before it broke (wall thinning or wear) −2Chevrons (V-shaped marks) or radial marks pointing to the origin −1No crack, but scratches, worn surfaces or transferred material −1Slight −1Cyclic load (vibration, rotation, changing pressure or load) −1Low (below about 0 °C) −1High-strength steel (quenched-and-tempered high-strength bolts, springs…) −1Rubber or elastomer (seals, O-rings…) −1Years or more −1

Related terms: 0.2% offset yield strength Work hardening

Practice questions:

1. Which pairing of a feature seen when observing a fracture surface in a scanning electron microscope with the type of fracture is correct?
2. In a ductile metal such as steel, the energy needed for a crack to advance (the critical energy release rate G_c) is orders of magnitude larger than 2γs (a few J/m²) estimated from the surface energy. What is the main reason?
3. For materials such as aluminum alloys that show no distinct yield point, which correctly describes how the 0.2% offset yield strength is determined?

How to recognize it (fracture surface, microstructure, SEM)

• The fracture surface is flat and roughly perpendicular to the load. There is no obvious deformation, and the pieces fit back together well.
• Fractures of plates and walls show chevrons (V-shaped marks) or radial marks; following their direction leads back to the origin.
• In the SEM you see flat cleavage facets, where the crystal split along specific planes, and river patterns. Fracture can also follow grain boundaries, for example in temper embrittlement, where impurities segregate to the boundaries.
• In ceramics and glass, a smooth area called the fracture mirror forms around the origin, and its size can be used to estimate the fracture stress (ASTM C1678).

Tests to confirm it

• Use Charpy impact tests to find the absorbed energy at the service temperature and the ductile-to-brittle transition temperature.
• Measure the fracture toughness K_IC and calculate K = Yσ√(πa) from the crack size found and the stress to check whether the fracture condition was reached.
• Examine the microstructure and chemistry for factors that lower toughness: coarse grains, impurities segregated to grain boundaries, or hardening in a weld heat-affected zone.

Common mix-ups

• Even when the final fracture is brittle, the origin may be a small crack grown by fatigue or stress corrosion that simply reached critical size (as in the Silver Bridge collapse). Examine the origin at high magnification.
• Ferritic (body-centered cubic) steels become brittle abruptly at low temperature, but aluminum, copper and nickel alloys and austenitic stainless steels (face-centered cubic) show no clear transition.

Prevention

Material

Choose steel with enough toughness (Charpy energy, fracture toughness) at the service temperature, and refine its grain size.

Design

Avoid notches and abrupt changes in section, and design so that the whole member yields before a crack can cause fracture (yield before fracture).

Environment and use

Avoid impact loading below the ductile-to-brittle transition temperature.

Manufacturing and maintenance

Inspect welds for defects and hardening, and control welding conditions to avoid low-toughness heat-affected zones.

Clues that score for this mode

Chevrons (V-shaped marks) or radial marks pointing to the origin +3Flat, shiny facets (cleavage) or river patterns +3Impact or a sudden large load +3Low (below about 0 °C) +3Ceramic or glass +3Almost none (the pieces fit together closely) +2Ordinary carbon or low-alloy steel +2Immediately, during manufacture or at the first large load +2Intergranular surface showing the shapes of grains +1High-strength steel (quenched-and-tempered high-strength bolts, springs…) +1Large (necking, bending or stretching) −3Austenitic stainless steel −3Aluminum alloy −3Copper alloy (brass…) −3Dimples (tiny cups) all over the surface in the SEM −2Striations (fine parallel lines) in the SEM −2Material was lost before it broke (wall thinning or wear) −2High (above about 0.4 of the absolute melting point; roughly 450 °C or more for steel) −2Heat-resistant alloy (nickel-base superalloy…) −2Rubber or elastomer (seals, O-rings…) −2Beach marks (shell-like bands spreading from the origin) −1No crack, but scratches, worn surfaces or transferred material −1Little external load (residual stress from welding, bending or press fits) −1Moderate (about 60–300 °C; hot-water or steam piping) −1Plastic −1

Related terms: Fracture toughness Stress intensity factor Griffith theory Yield before fracture Heat-affected zone

Practice questions:

1. Which material shows a distinct ductile-to-brittle transition at low temperatures?
2. A steel plate with a fracture toughness K_IC = 50 MPa·√m carries a tensile stress of 200 MPa. Using K = Yσ√(πa) with a geometry factor Y = 1.0, which value is closest to the critical crack length a at which fracture occurs?
3. When the fracture toughness of plates of the same material but different thickness is measured, what happens to the apparent fracture toughness K_c as the thickness increases?
4. The theoretical strength of a perfect crystal is estimated at roughly one tenth of the Young's modulus, yet the actual fracture strength of brittle materials such as glass is often only a few tens to a few hundreds of times smaller than that. What is the main reason?

Case studies: Silver Bridge Collapse (1967): Brittle Fracture from an Eyebar Crack

How to recognize it (fracture surface, microstructure, SEM)

• To the eye, beach marks (shell-like bands) spread out from the origin. They record stops in operation and changes in load.
• When there are several origins, ratchet marks (steps along the edge) appear between them.
• The region where the crack grew is smooth, has no shear lip and may be discolored. The final fracture zone is rough.
• In the SEM you see striations, each marking roughly one cycle of crack advance. They are clear in aluminum alloys and may be indistinct in high-strength steels.
• Origins are often at places where stress concentrates: surface scratches and notches, weld toes and inclusions.

Tests to confirm it

• Record the location and size of the origin, beach marks and final fracture zone under a stereo microscope, and look for striations in the SEM.
• Estimate the crack growth rate and the stress intensity range ΔK from striation spacing, and work back to the stress the part saw.
• Measure surface roughness, hardness and residual stress (by X-ray diffraction) at the origin.
• Estimate the size and number of stress cycles by measuring loads or analyzing vibration.

Common mix-ups

• Missing striations do not rule out fatigue. They are hard to see where the crack grew slowly, and rubbing between the fracture faces or corrosion can erase them.
• Stress corrosion cracking can look similar to fatigue. Decide by whether the load was cyclic and whether a specific corrosive environment was present.
• Do not judge overload or brittle fracture from dimples or cleavage in the final fracture zone alone.

Prevention

Material

Choose materials with high fatigue strength, such as clean steels with few inclusions.

Design

Reduce stress concentrations from notches and abrupt changes in section, and lower the stress amplitude and tensile mean stress (assessed with a modified Goodman diagram).

Environment and use

Corrosion lowers fatigue strength, so protect against it.

Manufacturing and maintenance

Introduce compressive residual stress at the surface by shot peening, carburizing or nitriding, reduce surface damage, and improve the shape of weld toes.

Clues that score for this mode

Beach marks (shell-like bands spreading from the origin) +3Striations (fine parallel lines) in the SEM +3Cyclic load (vibration, rotation, changing pressure or load) +3Ratchet marks (steps along the edge; several origins) +2Almost none (the pieces fit together closely) +2Slight +1Repeated heating and cooling (start-ups and shutdowns, quenching) +1High-strength steel (quenched-and-tempered high-strength bolts, springs…) +1Ordinary carbon or low-alloy steel +1Aluminum alloy +1Hours to months +1Years or more +1Steady load (self-weight, internal pressure, bolt preload…) −3Material was lost before it broke (wall thinning or wear) −2Impact or a sudden large load −2Little external load (residual stress from welding, bending or press fits) −2Intergranular surface showing the shapes of grains −1Fine, branching cracks in cross section −1Large (necking, bending or stretching) −1Ceramic or glass −1Immediately, during manufacture or at the first large load −1

Related terms: Fatigue limit Fatigue striations Paris law Miner's rule Modified Goodman diagram Stress concentration factor

Practice questions:

1. The fracture surface of a part that failed by fatigue shows shell-like patterns visible to the naked eye (beach marks) and fine lines visible in an electron microscope (striations). Which statement about them is correct?
2. Which statement correctly describes the initiation of fatigue cracks when a stress amplitude below the yield stress is repeatedly applied to a metal with a smooth surface?
3. Which surface treatment can be expected to improve the fatigue strength of a steel part?
4. Which statement correctly describes the curve of fatigue crack growth rate da/dN plotted against the stress intensity factor range ΔK on log-log axes?
5. Which statement about fatigue S-N curves is correct?

Case studies: Aloha Airlines Flight 243 (1988): Fatigue and Corrosion in Lap Joints

How to recognize it (fracture surface, microstructure, SEM)

• Beach marks and ratchet marks appear as in fatigue, with pits or corrosion products at the origin.
• Many small cracks can start from pits and join into a larger crack.
• The fracture surface may be covered with corrosion products that hide the striations, or show brittle striations: lines across flat facets.

Tests to confirm it

• Analyze the corrosion products at the origin and on the fracture surface (by EDS, for example) for constituents of the environment such as chloride.
• Run fatigue tests in the actual environment. In seawater and similar environments, life can drop sharply at lower cycling frequencies, so test at realistic frequencies.

Common mix-ups

• A fatigue limit no longer appears in a corrosive environment, so a part designed below its fatigue limit in air can still fail.
• Stress corrosion cracking can grow under a steady load, whereas corrosion fatigue needs cyclic loading. Use the type of load to tell them apart.

Prevention

Material

Choose a material with good corrosion resistance in that environment, or use a protective coating.

Design

Avoid crevices and pockets where water collects, and lower the stress amplitude.

Environment and use

Control corrosion with paint, plating, cathodic protection or water chemistry control.

Manufacturing and maintenance

Introduce compressive surface residual stress by shot peening, and inspect to keep coatings free of damage.

Clues that score for this mode

Cyclic load (vibration, rotation, changing pressure or load) +3Beach marks (shell-like bands spreading from the origin) +2Ratchet marks (steps along the edge; several origins) +2Striations (fine parallel lines) in the SEM +2Pits or corrosion products at the origin +2Water, humidity or condensation +2Chlorides (seawater, salt, de-icing salts) +2Years or more +2Cyclic load × chlorides +2Many surface cracks in a network, filled with oxide +1Slight +1Almost none (the pieces fit together closely) +1Ordinary carbon or low-alloy steel +1Aluminum alloy +1Cyclic load × water or humidity +1Steady load (self-weight, internal pressure, bolt preload…) −3Plastic −3Rubber or elastomer (seals, O-rings…) −3Dry air −2Ceramic or glass −2Immediately, during manufacture or at the first large load −2

Related terms: Fatigue limit Pitting corrosion Fatigue striations

Practice questions:

1. When SUS304 (JIS; AISI 304) is used in chloride-rich seawater, which correctly pairs the localized corrosion likely to occur with a countermeasure?
2. Which statement about fatigue S-N curves is correct?
3. The fracture surface of a part that failed by fatigue shows shell-like patterns visible to the naked eye (beach marks) and fine lines visible in an electron microscope (striations). Which statement about them is correct?

Case studies: Aloha Airlines Flight 243 (1988): Fatigue and Corrosion in Lap Joints

How to recognize it (fracture surface, microstructure, SEM)

• Fine cracks grow with branches. From outside there is little sign of corrosion and only a small crack shows at the surface, so it is hard to find.
• Chloride cracking of austenitic stainless steel usually runs through the grains (transgranular), while season cracking of brass in ammonia and cracking of sensitized stainless steel usually follow grain boundaries (intergranular); the crack path is a clue.
• There is no large deformation at the fracture, and the cracks or the fracture surface may be discolored by corrosion products.

Tests to confirm it

• Polish and etch a cross section and look under an optical microscope for branching and for a transgranular or intergranular path.
• Check susceptibility in the environment with U-bend, constant-load or slow strain rate tests (SSRT). Austenitic stainless steels are often tested in boiling 42% magnesium chloride.
• Measure tensile residual stress from welding or bending by X-ray diffraction.

Common mix-ups

• It can look similar to fatigue. Tell them apart by whether the load was cyclic and whether a specific environment was present.
• Intergranular fractures from hydrogen embrittlement, liquid metal embrittlement and creep look similar.
• When cracking is driven by hydrogen produced by corrosion, cathodic protection can make it worse, the opposite of cracking that grows by dissolving metal.

Prevention

Material

Switch to higher-nickel grades or duplex stainless steels, or use a crack-resistant heat treatment such as the overaged T73 temper of aluminum alloy 7075.

Design

Reduce tensile residual stress and stress concentrations, and avoid crevices and pockets where chloride can concentrate.

Environment and use

Reduce chloride and dissolved oxygen, lower the temperature, and use inhibitors. Chloride cracking of austenitic stainless steel is generally considered likely above about 60 °C.

Manufacturing and maintenance

Lower tensile residual stress with stress-relief annealing or shot peening. Stress-relief annealing is widely used for carbon steel but is often ineffective for austenitic stainless steel.

Clues that score for this mode

Fine, branching cracks in cross section +3Little external load (residual stress from welding, bending or press fits) +3Strong alkalis or ammonia +3Chlorides × austenitic stainless steel +3Ammonia or alkali × copper alloy (season cracking) +3Intergranular surface showing the shapes of grains +2Almost none (the pieces fit together closely) +2Steady load (self-weight, internal pressure, bolt preload…) +2Moderate (about 60–300 °C; hot-water or steam piping) +2Chlorides (seawater, salt, de-icing salts) +2Austenitic stainless steel +2Copper alloy (brass…) +2Strong alkali × carbon steel (caustic embrittlement) +2Pits or corrosion products at the origin +1Water, humidity or condensation +1Hydrogen sources (pickling, electroplating, welding, cathodic protection, hydrogen gas) +1High-strength steel (quenched-and-tempered high-strength bolts, springs…) +1Aluminum alloy +1Hours to months +1Years or more +1Plastic −3Rubber or elastomer (seals, O-rings…) −3Large (necking, bending or stretching) −2Dry air −2Dimples (tiny cups) all over the surface in the SEM −1Cyclic load (vibration, rotation, changing pressure or load) −1Impact or a sudden large load −1Low (below about 0 °C) −1High (above about 0.4 of the absolute melting point; roughly 450 °C or more for steel) −1Ceramic or glass −1Immediately, during manufacture or at the first large load −1

Related terms: Stress corrosion cracking Sensitization Pitting corrosion Stainless steel Brass Passivity

Practice questions:

1. Which statement about stress corrosion cracking is correct?
2. When SUS304 (JIS; AISI 304) is used in chloride-rich seawater, which correctly pairs the localized corrosion likely to occur with a countermeasure?
3. When austenitic stainless steels such as SUS304 are welded, intergranular corrosion can occur in the heat-affected zone a short distance from the weld line (weld decay). Which combination of cause and countermeasure is correct?
4. What is the main reason stainless steel resists rusting in air and water?

Case studies: Silver Bridge Collapse (1967): Brittle Fracture from an Eyebar Crack

How to recognize it (fracture surface, microstructure, SEM)

• There is almost no deformation at the fracture, and in high-strength steel the fracture often follows grain boundaries.
• It breaks suddenly after some time under a steady load or bolt preload.
• In welds, cold cracking (hydrogen cracking) can appear in the heat-affected zone some time after welding.

Tests to confirm it

• Measure hardness or strength. Steels with a yield strength below about 600 MPa are considered unlikely to suffer, while above about 1000 MPa it is likely to be a major problem.
• Investigate the manufacturing history (pickling, electroplating, welding) and hydrogen sources in service (corrosion, cathodic protection).
• Compare charged and uncharged specimens in sustained-load or slow strain rate tests.

Common mix-ups

• It overlaps with stress corrosion cracking, but hydrogen embrittlement can occur without a corrosive environment, from hydrogen introduced by plating or welding.
• Cracking after hot-dip galvanizing raises the suspicion of both liquid metal embrittlement and hydrogen, so check for both.

Prevention

Material

Do not raise strength (hardness) more than needed. Austenitic steels take up hydrogen more slowly by diffusion, so they take much longer to become embrittled.

Design

Review high-strength bolt preloads and stress concentrations to reduce sustained tensile stress.

Environment and use

Control the potential of cathodic protection so that overprotection does not generate hydrogen.

Manufacturing and maintenance

After pickling or electroplating, bake at about 200 °C for a few hours to drive out hydrogen. When welding, slow the cooling with preheating and post-heating to let hydrogen escape.

Clues that score for this mode

Hydrogen sources (pickling, electroplating, welding, cathodic protection, hydrogen gas) +3High-strength steel (quenched-and-tempered high-strength bolts, springs…) +3Hydrogen source × high-strength steel +3Intergranular surface showing the shapes of grains +2Almost none (the pieces fit together closely) +2Steady load (self-weight, internal pressure, bolt preload…) +2Little external load (residual stress from welding, bending or press fits) +2Hours to months +2Flat, shiny facets (cleavage) or river patterns +1Fine, branching cracks in cross section +1Around room temperature +1Water, humidity or condensation +1Immediately, during manufacture or at the first large load +1Plastic −3Rubber or elastomer (seals, O-rings…) −3Ceramic or glass −3Large (necking, bending or stretching) −2Dimples (tiny cups) all over the surface in the SEM −1Cyclic load (vibration, rotation, changing pressure or load) −1Austenitic stainless steel −1

Related terms: Heat-affected zone Carbon equivalent Stress corrosion cracking

Practice questions:

1. In welding high-strength steel, preheating is used to prevent cold cracking (hydrogen cracking), which appears in the heat-affected zone and elsewhere within minutes to days after welding. Which statement correctly describes the effect of preheating?
2. The carbon equivalent used as a guide to the weldability of steel is Ceq = C + Mn/6 + (Cr + Mo + V)/5 + (Ni + Cu)/15 (mass %). Which value is closest to Ceq for a steel with 0.20% C, 1.20% Mn, 0.30% Cr, 0.10% Mo, 0% V, 0.30% Ni and 0.15% Cu?
3. Which statement about stress corrosion cracking is correct?

How to recognize it (fracture surface, microstructure, SEM)

• Cross sections near the fracture show many grain-boundary voids (cavities) and wedge cracks growing from grain triple points.
• Grain boundaries are where creep deformation and damage progress fastest, so cracks often follow them.
• Because creep is deformation that increases with time, it may show as bulging of a tube or stretching of a member before rupture.

Tests to confirm it

• Run creep rupture tests at the service temperature and stress, and estimate life by trading temperature against time with the Larson–Miller parameter.
• Use operating records to find the actual metal temperature (was there overheating?), and infer the temperature history from microstructural changes such as precipitate coarsening.

Common mix-ups

• An intergranular fracture alone is hard to distinguish from hydrogen embrittlement or stress corrosion cracking, so treat temperature and time as the most important clues.
• If the temperature also cycles, consider thermal fatigue and combined creep and fatigue (thermomechanical fatigue).
• Plastics have low melting points, so they can creep, stretch and eventually crack under long-term load even at room temperature. This guide mainly covers high-temperature creep of metals, but suspect creep if a plastic part stretched slowly over a long time.

Prevention

Material

Choose materials with high high-temperature strength such as nickel-base superalloys. Because grain boundaries are a weak point in creep, directionally solidified columnar grains or single crystals can remove boundaries perpendicular to the stress.

Design

Lower the service temperature and stress, and allow a life margin using the Larson–Miller parameter or similar.

Environment and use

Prevent overheating, and lower metal temperatures with thermal barrier coatings and the like.

Manufacturing and maintenance

Carry out solution and aging heat treatments correctly to obtain the intended precipitate structure.

Clues that score for this mode

Many grain-boundary voids or wedge cracks near the fracture +3Slow bulging or stretching over a long time +3High (above about 0.4 of the absolute melting point; roughly 450 °C or more for steel) +3Heat-resistant alloy (nickel-base superalloy…) +3Plastic × slow stretching over a long time (plastics creep even at room temperature) +3Intergranular surface showing the shapes of grains +2Steady load (self-weight, internal pressure, bolt preload…) +2Years or more +2Slight +1Repeated heating and cooling (start-ups and shutdowns, quenching) +1Hot gases (combustion gases, hot air) +1Low (below about 0 °C) −3Impact or a sudden large load −2Around room temperature −2Immediately, during manufacture or at the first large load −2Cyclic load (vibration, rotation, changing pressure or load) −1Moderate (about 60–300 °C; hot-water or steam piping) −1Rubber or elastomer (seals, O-rings…) −1

Related terms: Creep Power-law creep Diffusion creep Larson–Miller parameter Nickel-based superalloy Deformation mechanism map

Practice questions:

1. Which statement about the three stages of a creep curve (strain versus time) measured at constant load and temperature is correct?
2. At room temperature, refining the grains raises strength (the Hall–Petch relation), but components loaded for long times at high temperature often use coarse grains or single crystals. Which is the correct reason?
3. A heat-resistant alloy ruptured by creep in 1000 h at 1000 K under a certain stress. Assuming the Larson–Miller parameter P = T(C + log₁₀t_r) (T is the absolute temperature, t_r is the rupture time in h, and C = 20) is constant at the same stress, which temperature is closest to that giving a rupture life of 10⁵ h at the same stress?
4. Why are directionally solidified columnar-grained or single-crystal nickel-based superalloys used for jet engine turbine blades?
5. Take about 0.4 times the melting point (in absolute temperature) as the rough temperature above which creep can no longer be ignored. For aluminum (melting point 660 °C), which value is closest to this temperature?

How to recognize it (fracture surface, microstructure, SEM)

• Many cracks form on the surface that is quenched, giving a crazed, networked look.
• The cracks are wide, start at the surface and are often filled with oxide from high-temperature exposure. They are dagger-shaped and run through the grains.
• Cracks often follow weld toes and places where the section thickness changes.

Tests to confirm it

• Use operating records to count start-ups and shutdowns and the rates of temperature change, and estimate the thermal stress from differences in expansion and temperature.
• Because the cracks usually connect to the surface, look for them by magnetic particle or penetrant testing.

Common mix-ups

• Many shallow cracks are easily mistaken for corrosion cracking. Check whether they concentrate on the quenched surface and run perpendicular to the thermal stress.

Prevention

Material

Reduce the difference in thermal expansion between joined materials, and choose materials with low expansion and good thermal conductivity.

Design

Use flexible supports that allow expansion and contraction, and make welds and changes in section smooth.

Environment and use

Control heating and cooling rates during start-ups and shutdowns, and keep cold water off hot components.

Manufacturing and maintenance

Grind weld toes smooth to reduce stress concentration.

Clues that score for this mode

Many surface cracks in a network, filled with oxide +3Repeated heating and cooling (start-ups and shutdowns, quenching) +3High (above about 0.4 of the absolute melting point; roughly 450 °C or more for steel) +2Heat-resistant alloy (nickel-base superalloy…) +2Beach marks (shell-like bands spreading from the origin) +1Ratchet marks (steps along the edge; several origins) +1Striations (fine parallel lines) in the SEM +1Almost none (the pieces fit together closely) +1Hot gases (combustion gases, hot air) +1Years or more +1Plastic −2Rubber or elastomer (seals, O-rings…) −2Steady load (self-weight, internal pressure, bolt preload…) −1Low (below about 0 °C) −1

Related terms: Thermal stress Coefficient of linear thermal expansion Thermal barrier coating

Practice questions:

1. An aluminum film (Young's modulus 70 GPa, Poisson's ratio 0.35, linear thermal expansion coefficient 23×10⁻⁶ K⁻¹) on a silicon substrate (linear thermal expansion coefficient 2.6×10⁻⁶ K⁻¹) is cooled from a stress-free 125 °C to 25 °C. Assuming the substrate is sufficiently thick, what are the sign and magnitude of the thermal stress in the film, σ = EΔαΔT/(1 − ν)?
2. What is the main cause of solids expanding when heated (thermal expansion)?

How to recognize it (fracture surface, microstructure, SEM)

• The part cracks with almost no deformation while in contact with molten metal. In hot-dip galvanized steel, intergranular cracks form beneath the zinc layer.
• Grain boundaries on the fracture surface are covered with zinc or another embrittling metal.
• Well-known couples include aluminum with gallium or mercury, steel with zinc, copper or tin, and copper alloys with mercury.

Tests to confirm it

• Analyze the fracture surface and the inside of cracks with SEM and EDS or EPMA to find the embrittling metal.
• Check in cross section whether the cracks follow grain boundaries and how far the metal has penetrated.
• Review plating and welding conditions, forming conditions, residual stress and the steel’s chemical composition.

Common mix-ups

• Because the fracture is intergranular, it is easily mistaken for stress corrosion cracking or hydrogen embrittlement. Check the history of contact with molten metal.
• Cracking after plating can also be caused by hydrogen absorbed by the steel, so investigate both.

Prevention

Material

When choosing steel, watch for elements such as copper and tin that segregate to grain boundaries and promote cracking.

Design

Use shapes that do not concentrate tensile residual stress from cold bending or welding.

Environment and use

Keep molten metals such as solder or mercury away from the part in service.

Manufacturing and maintenance

Control forming and welding before plating, and the conditions of the plating operation itself.

Clues that score for this mode

Another metal (zinc, solder…) on the fracture surface or grain boundaries +3Molten metal (hot-dip galvanizing, solder, braze, mercury) +3Intergranular surface showing the shapes of grains +2Almost none (the pieces fit together closely) +2Little external load (residual stress from welding, bending or press fits) +2Immediately, during manufacture or at the first large load +2Steady load (self-weight, internal pressure, bolt preload…) +1High (above about 0.4 of the absolute melting point; roughly 450 °C or more for steel) +1High-strength steel (quenched-and-tempered high-strength bolts, springs…) +1Aluminum alloy +1Copper alloy (brass…) +1Molten metal × during manufacture or early service +1Plastic −3Rubber or elastomer (seals, O-rings…) −3Ceramic or glass −3Large (necking, bending or stretching) −2Years or more −2

Related terms: SEM (scanning electron microscope) EDS (energy-dispersive X-ray spectroscopy)

Practice questions:

1. Which correctly compares EDS (energy-dispersive spectroscopy) and WDS (wavelength-dispersive spectroscopy) for analyzing characteristic X-rays in SEM and EPMA?
2. Which correctly describes secondary electron and backscattered electron images in SEM?

How to recognize it (fracture surface, microstructure, SEM)

• When hard particles or asperities cut the surface (abrasive wear), scratches line up in the sliding direction.
• When metals bond and the junctions shear (adhesive wear), material is transferred to the counterface or torn from the surface.
• The part does not crack; its dimensions shrink.

Tests to confirm it

• Examine the worn surfaces for the direction of scratches and the shape and composition of wear debris, and compare the surface hardness with that of the counterface and any contaminating particles.
• Measure wear volume in pin-on-disk or similar tests and compare materials and conditions with Archard’s equation V = KWL/H.

Common mix-ups

• A section thinned by wear may finally crack, so check not just the crack but whether the dimensions had already shrunk.
• Wall thinning of piping in flowing fluid (erosion–corrosion) is a different mechanism with no contacting counterpart, and needs different countermeasures.

Prevention

Material

Make the surface at least as hard as the abrasive particles (carburizing, nitriding, hard coatings…).

Design

Lower the contact pressure and sliding speed.

Environment and use

Move lubrication toward the hydrodynamic regime (see the Stribeck curve) and keep sand and other contaminants out.

Manufacturing and maintenance

Control surface roughness and finish so that asperities are less likely to touch through the oil film.

Clues that score for this mode

No crack, but scratches, worn surfaces or transferred material +3Material was lost before it broke (wall thinning or wear) +3Sliding, rubbing or rolling against another surface +3Inner surface scalloped like fish scales or orange peel, and thinned +1Fast flow, droplets or particle-laden flow +1Rubber or elastomer (seals, O-rings…) +1Ceramic or glass +1Years or more +1Immediately, during manufacture or at the first large load −1

Related terms: Archard wear equation Adhesive wear Abrasive wear Stribeck curve Carburizing Nitriding

Practice questions:

1. Which of the following descriptions of wear mechanisms corresponds to adhesive wear?
2. A construction-machine part is rubbed by soil containing quartz sand grains (hardness about 1000 HV). Which is the correct way to choose a material that reduces abrasive wear?
3. In a pin-on-disk test, a pin of hardness H = 2.0 GPa is loaded with a normal force of 50 N and slid over a distance of 1.0 km. Using the Archard wear equation V = KWL/H with a dimensionless wear coefficient K = 1.0×10⁻⁴, which correctly pairs the wear volume of the pin with the change in wear volume when the load is doubled?
4. Which statement correctly describes the Stribeck curve, which plots the friction coefficient of a plain bearing against the bearing characteristic number ηN/P formed from the lubricant viscosity η, the rotational speed N and the bearing pressure P?

How to recognize it (fracture surface, microstructure, SEM)

• The inner surface is scalloped like fish scales or orange peel and thinned locally.
• Thinning occurs where flow is disturbed, such as downstream of orifices, elbows and tees.
• The area around the rupture is very thin, and the thinned wall tears open under internal pressure.

Tests to confirm it

• Measure wall thickness with an ultrasonic gauge and estimate the thinning rate and remaining life from the change since the previous measurement.
• Analyze the chromium content of the steel (even a little chromium slows thinning).
• Review water chemistry (pH, dissolved oxygen) and records of temperature and flow. In single-phase water flow, thinning is reported to be fastest at about 140–160 °C.

Common mix-ups

• Thinning is local and, unlike general corrosion, cannot be predicted from an average corrosion rate.
• Liquid droplet impingement erosion from droplets in steam also thins walls, but it occurs in different places and needs different countermeasures.

Prevention

Material

Replace with chromium-containing low-alloy steel (Cr–Mo steel with at least 1.25% Cr) or stainless steel.

Design

Reduce complex geometry that disturbs the flow.

Environment and use

Keep the pH mildly alkaline and control dissolved oxygen.

Manufacturing and maintenance

Register every location to be inspected, with no omissions, and measure the thickness first where flow is disturbed.

Clues that score for this mode

Inner surface scalloped like fish scales or orange peel, and thinned +3Material was lost before it broke (wall thinning or wear) +3Fast flow, droplets or particle-laden flow +3Fast flow × carbon steel +3Moderate (about 60–300 °C; hot-water or steam piping) +2Years or more +2No crack, but scratches, worn surfaces or transferred material +1Water, humidity or condensation +1Ordinary carbon or low-alloy steel +1Austenitic stainless steel −3Plastic −3Rubber or elastomer (seals, O-rings…) −3Ceramic or glass −3Immediately, during manufacture or at the first large load −3Dry air −2High (above about 0.4 of the absolute melting point; roughly 450 °C or more for steel) −1Aluminum alloy −1Copper alloy (brass…) −1Heat-resistant alloy (nickel-base superalloy…) −1

Related terms: Passivity Faraday's law of electrolysis

Practice questions:

1. When the corrosion current density of iron is uniformly 1.0 µA/cm² over the whole surface, which value is closest to the rate at which the surface recedes? Assume iron dissolves as Fe²⁺, with a molar mass of 55.85 g/mol, a density of 7.87 g/cm³, a Faraday constant of 96485 C/mol, and one year as 3.15×10⁷ s.
2. What is the main reason stainless steel resists rusting in air and water?

Case studies: Mihama Unit 3 Pipe Rupture (2004): Thinning Downstream of an Orifice

How to recognize it (fracture surface, microstructure, SEM)

• A rubber seal may show traces of leakage through the gap or heat damage (charring, erosion) even without cracking.
• A plastic part may break with almost no deformation below its glass transition temperature, where its molecules can barely move (polypropylene, for example, has a glass transition at about −20 to 0 °C).

Tests to confirm it

• Measure how fast the rubber returns to its original shape after compression (resilience) at the service temperature.
• Measure the glass transition temperature and compare it with the lowest service temperature.
• Test over the full service range, including the lowest temperature, in conditions as close to the real part as possible.

Common mix-ups

• A seal failure leaves no crack or fracture surface, so it may not be recognized as a materials problem.
• Polymer properties depend on time and temperature: lowering the temperature has the same effect as measuring over a shorter time (faster deformation). Following fast movements becomes even harder at low temperature.

Prevention

Material

Choose a material whose glass transition temperature is well below the lowest service temperature.

Design

Make the design insensitive to variations in temperature and dimensions, so the seal can follow a moving gap.

Environment and use

Keep the part warm with heaters, or set temperature limits for launch or operation.

Manufacturing and maintenance

Qualify the part with tests close to real use across the whole service range, including low temperature.

Clues that score for this mode

Low (below about 0 °C) +3Rubber or elastomer (seals, O-rings…) +3Plastic +2Rubber or elastomer × low temperature +2Plastic × low temperature +2Fine crazes or whitening in a plastic +1Almost none (the pieces fit together closely) +1High (above about 0.4 of the absolute melting point; roughly 450 °C or more for steel) −3High-strength steel (quenched-and-tempered high-strength bolts, springs…) −3Ordinary carbon or low-alloy steel −3Austenitic stainless steel −3Aluminum alloy −3Copper alloy (brass…) −3Heat-resistant alloy (nickel-base superalloy…) −3Ceramic or glass −3Moderate (about 60–300 °C; hot-water or steam piping) −2Large (necking, bending or stretching) −1Around room temperature −1

Related terms: Glass transition temperature Viscoelasticity

Practice questions:

1. Which statement about the glass transition temperature Tg and melting point Tm of a crystalline (semicrystalline) polymer is correct?
2. When a plasticizer is added to poly(vinyl chloride) (PVC), what happens to the glass transition temperature Tg and the properties of the material?
3. Which correctly describes the idea of time–temperature superposition in polymer viscoelasticity?

Case studies: The Challenger Accident (1986): Cold O-Rings That Recovered Too Slowly

How to recognize it (fracture surface, microstructure, SEM)

• Several fine crazes or cracks start where stress is high, from bending, clamping or press fits, and one of them grows into the failure.
• The area near the crack may turn whitish (stress whitening).
• The plastic breaks in a brittle way with little deformation.

Tests to confirm it

• Bend and hold a specimen, apply the liquid, and record the time or strain at which cracks or crazes appear (standards such as ASTM D1693 exist for polyethylene).
• Identify the liquids, detergents and lubricants on the part to establish its service environment.

Common mix-ups

• It can occur without the plastic swelling or dissolving, so suspect it even when the liquid looks harmless.
• A liquid at a harmless concentration can cause cracking once it concentrates, for example by evaporation.

Prevention

Material

Choose a plastic that resists cracking in the liquids it will contact.

Design

Lower the tensile stress from clamping, press fits and bending.

Environment and use

Keep the part away from liquids that cause cracking, such as solvents and surfactants.

Manufacturing and maintenance

Reduce molding residual stress.

Clues that score for this mode

Fine crazes or whitening in a plastic +3Oils, solvents or detergents (surfactants) +3Plastic +3Steady load (self-weight, internal pressure, bolt preload…) +2Little external load (residual stress from welding, bending or press fits) +2Plastic × oils, solvents or detergents +2Almost none (the pieces fit together closely) +1Around room temperature +1Rubber or elastomer (seals, O-rings…) +1Hours to months +1High-strength steel (quenched-and-tempered high-strength bolts, springs…) −3Ordinary carbon or low-alloy steel −3Austenitic stainless steel −3Aluminum alloy −3Copper alloy (brass…) −3Heat-resistant alloy (nickel-base superalloy…) −3Ceramic or glass −3Large (necking, bending or stretching) −2High (above about 0.4 of the absolute melting point; roughly 450 °C or more for steel) −1

Related terms: Crystallinity Viscoelasticity

Practice questions:

1. For the same type of polymer, what generally happens as the crystallinity increases?