Share this article
Continue reading with these related posts
Corrosion is not a single mechanism — it is a family of electrochemical attack processes, each with distinct visual signatures, distinct causes, and distinct prevention strategies. A pit that looks like surface roughness can initiate a fatigue crack at a third of the expected stress. A weld that looks structurally sound can be corroding from the inside along the heat-affected zone.
Fatigue failures do not announce themselves. A shaft carries its rated load for months, then fractures at a load it has handled thousands of times before. The failure looks sudden. The process was not — it began as a microscopic crack at a stress concentration, growing incrementally with each cycle until the remaining section could no longer carry the load.
The fracture surface of a failed part is a record of how the failure progressed. Beach marks, chevron patterns, shear lips, and ratchet marks each tell a specific story about load type, crack origin, and propagation rate. Learning to read these features is the most cost-effective skill in failure analysis.
A component that fractures in a single load event tells a different story than one that accumulated damage over thousands of cycles. The fracture surface of an overload failure is characterised by the absence of progressive damage features — no beach marks, no gradual crack front advancement, no systematic evidence of crack growth. Instead, the surface records the load direction, the material condition at the time of fracture, and the degree of plastic deformation that preceded it.
Reading these features correctly is important because overload failures have a different root cause distribution than fatigue failures. A fatigue failure almost always implicates the design — stress concentration, surface condition, operating load exceeding the endurance limit. An overload failure may implicate the design, but it may equally implicate an unexpected load, a material deficiency, an environmental condition that reduced toughness, or a combination of all three. The fracture surface is the starting point for sorting these possibilities.
The first question for any fractured component is whether the failure was fatigue (progressive, multi-cycle) or overload (sudden, single or few cycles). The answer is usually visible within the first minute of examination.
Fatigue fractures have beach marks and a propagation zone that is distinctly smoother than the final fracture zone. The crack grew incrementally, and the surface records that progression. The final fracture zone is typically a small fraction of the total fracture area in moderate-stress fatigue.
Overload fractures have none of this. The entire fracture surface is the result of rapid failure in a single event. There is no smooth propagation zone, no beach marks, and no systematic crack front geometry. The surface may show either ductile or brittle characteristics — but those characteristics are uniform across the fracture rather than zoned.
Some Fractures Combine Both Mechanisms
Many real-world fractures begin with fatigue crack growth and end in overload when the crack reaches critical size. In these mixed fractures, the fatigue zone (with beach marks) and the overload zone (rough, irregular) are both present on the same surface. The relative sizes of these zones indicate the stress history: a large fatigue zone and small overload zone means the crack grew to near-critical size under moderate stress before rapid fracture; a tiny fatigue zone and large overload zone means the crack was small when the overload event occurred, indicating either a single large overload or a material with low fracture toughness.
Ductile overload occurs when a ductile material — most structural steels, aluminium alloys in appropriate tempers, copper alloys — is loaded to fracture. Before fracture, the material deforms plastically at the failure site, absorbing energy through gross deformation of the cross-section. The fracture then proceeds by void nucleation, growth, and coalescence: micro-voids form at inclusions and second-phase particles, grow under the tensile stress, and link together to produce the fracture surface.
The characteristic features of ductile overload fracture are:
Necking or visible plastic deformation at the fracture site. The cross-section at the fracture is reduced relative to the nominal section — the material has thinned before fracturing. On structural sections, the deformation is visible as angular distortion, bending, or elongation of features near the fracture.
Shear lip around the perimeter of the fracture, typically at 45 degrees to the main fracture plane. The shear lip forms where the fracture transitions from a flat, opening-mode crack (Mode I) to a shear fracture at the free surface. A shear lip indicates that the material had sufficient ductility to deform plastically before fracturing at the surface. The width of the shear lip relative to the section thickness indicates the degree of constraint and the material's ductility.
Fibrous or dimpled texture on the fracture surface at moderate magnification. Under a stereo microscope at 20× to 40×, the surface shows a dimpled morphology from the void coalescence mechanism. Each dimple corresponds to a single void nucleation event. Larger dimples indicate larger void nucleation sites — large inclusions or second-phase particles — and can indicate a material with higher inclusion content than specified.
Grey or slightly irregular surface colour, without the bright, flat, reflective appearance of brittle fracture. The texture is rough and three-dimensional rather than the flat, crystalline appearance of brittle failure.
Brittle fracture occurs with minimal plastic deformation before fracture. In engineering metals, brittle fracture is not a fixed property of the material — it is a condition that occurs when the material's toughness is reduced below the level needed to sustain the local stress intensity at the crack tip. The conditions that cause brittle fracture in normally ductile materials include low temperature, high loading rate, high constraint (thick section, triaxial stress state), hydrogen embrittlement, temper embrittlement, and stress corrosion cracking.
The characteristic features of brittle fracture are:
Flat fracture surface perpendicular to the maximum tensile stress, with little or no associated plastic deformation. The surrounding structure is undistorted — the fracture occurred without the material yielding first.
Absence of shear lip or a very narrow shear lip at the surface. In a material that fractured brittly, there was insufficient ductility to form the characteristic 45-degree surface zone. Complete absence of shear lip on a material that should be ductile is a strong indicator of embrittlement.
Chevron or herringbone patterns on the fracture surface. These V-shaped markings on the fracture surface point back toward the crack initiation site. They form because the crack front is not perfectly flat — local variations cause the crack to grow at slightly different angles in adjacent zones, producing the characteristic V-shaped ridges that are visible at low magnification. Chevron patterns are the primary tool for locating the origin of a brittle fracture on large components where other origin features may be difficult to identify.
Bright, granular appearance, particularly in coarse-grained materials. In metals with large grains, the fracture path follows cleavage planes within individual grains, producing flat, reflective facets. The resulting surface has a bright, glittering appearance under raking light that is characteristic of cleavage fracture.
Brittle Fracture in Structural Steel Is a Low-Temperature Phenomenon
Structural steels are ductile at room temperature but can transition to brittle behaviour at low temperatures — the ductile-to-brittle transition temperature (DBTT). Below the DBTT, impact energy drops dramatically and fracture toughness falls to levels that allow catastrophic brittle fracture at stresses far below yield. The Liberty ship failures in World War II and several bridge failures have involved this mechanism. When specifying structural steel for cold environments, verify that the steel has a specified minimum Charpy impact energy at the design minimum temperature, not just a minimum yield strength.
Impact fractures are a subset of overload fracture characterised by very high loading rates. The high strain rate reduces the material's effective toughness — even materials with excellent quasi-static toughness can fracture in a more brittle manner under impact conditions. The fracture features depend on the material and the impact severity:
In ductile materials under moderate impact, fracture features are similar to quasi-static ductile overload, but with a narrower shear lip and less necking because the deformation occurred over a shorter time. In ductile materials under severe impact, the fracture may show shear bands — narrow zones of intense localised shear at approximately 45 degrees to the impact direction — which indicate adiabatic shear, a condition where plastic deformation is so rapid that heat cannot dissipate and the local temperature rise softens the material along a narrow shear zone.
In materials near their ductile-to-brittle transition temperature, an impact load can push the effective fracture behaviour into the brittle regime even when a quasi-static load at the same stress level would produce ductile fracture. Impact testing (Charpy, Izod) characterises this transition so that materials can be specified with adequate toughness at the operating temperature and loading rate.
Beyond identifying ductile versus brittle, the fracture surface carries quantitative information about the load level at failure. Two features are most useful.
Fracture orientation relative to applied stress. A flat fracture surface perpendicular to the loading axis indicates tensile overload. A fracture surface at 45 degrees to the loading axis indicates shear overload. In torsional failures, ductile overload produces a 45-degree helical fracture; brittle overload produces a flat transverse fracture perpendicular to the maximum principal stress.
Degree of plastic deformation before fracture. Extensive necking, large shear lip, and significant angular distortion of the surrounding structure indicate that the failure load was only moderately above yield — the material deformed substantially before fracturing. Minimal deformation and a flat fracture at or near the elastic limit of the section indicates that the failure load was much higher than yield, or that the material lacked the toughness to accommodate the applied stress.
The Fracture Load Can Be Back-Calculated
For a simple geometry, the load that caused a ductile overload fracture can be estimated by calculating the force needed to develop ultimate tensile stress (or shear stress) over the remaining uncracked section area. This back-calculation is a useful check: if the estimated fracture load is within the expected operating range, the failure may have been caused by an expected load on a section that was undersized or had a material deficiency. If the estimated fracture load is substantially above the operating range, an unexpected overload event is the likely cause — which shifts the investigation toward the operating history rather than the design.
Overload fractures provide direct evidence of material condition, which is not available from most other inspection methods without destructive testing.
Inclusions visible on the fracture surface — elongated stringers, large rounded inclusions, or clusters of inclusions — indicate a material with higher inclusion content than typical for the specification. In a tensile overload, inclusions act as void nucleation sites; high inclusion density produces a rougher, more fibrous fracture surface and can reduce toughness significantly below the specified minimum.
Grain size visible at low magnification indicates abnormal grain growth, typically from a heat treatment error (overheating during austenitising) or from forging at too high a temperature. Coarse-grained steel has lower toughness than fine-grained steel at the same strength level.
Hardness assessment from fracture texture. A very bright, flat fracture surface in a material specified as ductile suggests the material was harder than specified — possibly from incorrect heat treatment, decarburisation on the wrong side, or substitution of a higher-strength grade. Hard materials have higher yield strength but lower toughness, and can fracture at the same nominal stress level that a softer material would survive.
A 50 mm diameter drive shaft on an industrial conveyor fractured at the keyway during system startup after a three-day scheduled maintenance shutdown. The shaft was 1045 steel, specified for quench-and-temper heat treatment to 28–34 HRC. The fracture occurred at the first startup following maintenance.
Visual examination of the fracture surface: flat fracture surface perpendicular to the shaft axis, with a very narrow shear lip (less than 1 mm wide) around the perimeter. No beach marks, no progressive crack growth evidence. Chevron patterns on the fracture surface pointed clearly to an origin at the keyway root. The surface had a bright, slightly granular appearance under raking light.
The flat fracture, narrow shear lip, and bright granular appearance were inconsistent with a normal ductile overload of 1045 steel in the specified heat treatment condition. At 28–34 HRC, quenched-and-tempered 1045 should produce a fracture with a significant shear lip and visibly fibrous texture, not the flat, brittle-appearing surface observed.
Hardness measurement at the fracture site: 46 HRC — substantially higher than the specified 34 HRC maximum. The shaft had been through-hardened to a higher hardness than specified, reducing its toughness from approximately 80 J Charpy to approximately 20 J.
At startup, the conveyor experienced a brief jam before the overload protection activated, producing a torque spike estimated at approximately 2.5 times nominal. A shaft at the specified 28–34 HRC would have survived this spike with plastic deformation but no fracture. At 46 HRC, the reduced toughness allowed fracture at this stress level.
Root cause: heat treatment out-of-specification. The shaft had been re-heat-treated at some point in the maintenance cycle — the records showed a "re-treatment for corrosion" that was not authorised in the maintenance procedure. The re-treatment produced a hardness well above specification. Corrective action: mandatory hardness verification of shafts after any heat treatment, added to the maintenance quality control plan.
Specify material properties for the load case, not just the nominal condition. A high-strength material is not always the best choice for impact-loaded applications. For applications with significant shock or impact loading, specify toughness (Charpy impact energy at the operating temperature) alongside strength. A steel with 70 J Charpy at −20°C is more suitable for outdoor impact service than one with 150 MPa higher yield strength and 15 J Charpy at −20°C.
Verify that the material condition matches specification. Hardness testing and material traceability are the minimum; for critical components, Charpy testing from the heat provides direct evidence of toughness. Certificate review without verification testing is insufficient for safety-critical applications.
Design against stress concentrations at the section where overload will occur. A fracture initiated at a keyway, a sharp shoulder radius, or an undercut is not purely a material problem — the stress concentration reduced the effective fracture resistance at that location. Generous radii at geometric transitions reduce the stress intensity factor at that location and increase the overload fracture load.
Set overload protection at a level the design can survive. Mechanical fuses, torque limiters, and shear pins should be calibrated to release at a load below the fracture load of the critical component, not at the nominal operating load. An overload protection system that does not protect against the actual failure load provides false confidence.
Overload Failures Are Diagnostic Opportunities
Every overload failure carries direct information about the load that caused it, the material condition at the time, and the design margin that was exceeded. This information is rarely available from simulations or calculations alone. A fractured component that has been carefully examined, with load back-calculated and material condition characterised, provides better calibration data for the design than months of nominal operation. Treat overload failures as expensive but informative tests of the design envelope.
Overload and impact failures are driven by mechanical loading that exceeds the material's strength in a short time. The next post examines thermal failure — a different category where the loading is thermal rather than mechanical, and the failure mechanisms include creep, thermal fatigue, and phase transformation, all of which leave distinct signatures that require different examination techniques to identify.
