Fracture Mechanics Fundamentals

A.A. Griffith was working at the Royal Aircraft Establishment in Farnborough during World War One when he became interested in the discrepancy between theoretical and actual strength. The theoretical tensile strength of glass, calculated from the atomic bond energy, runs to about 10 GPa. Actual glass shatters at stresses closer to 100 MPa. The ratio is about one hundred to one. That gap was not explained by any material-property table in 1920.

Griffith's insight was that the gap is caused by cracks. An elliptical crack in a plate under tension acts as a stress concentrator. At the crack tip, the local stress can be orders of magnitude higher than the nominal applied stress. When that local stress exceeds the theoretical strength over the tiny volume at the tip, bonds break, the crack extends, and strain energy stored in the surrounding material drives it further.

His criterion is elegant: a crack of half-length a will grow when the release of elastic strain energy per unit area of crack extension equals the surface energy γ per unit area of new surface created. For a through-crack in an infinite plate under remote stress σ, the critical condition becomes σ = √(2Eγ/πa), where E is Young's modulus. Longer cracks need less stress to propagate. This single equation explained why scratched glass breaks far more easily than pristine glass, and why larger components fail at lower nominal stresses. This was the first rigorous, quantitative theory of fracture.

Griffith's energy approach worked for brittle solids. For metals, where plastic deformation at the crack tip dissipates far more energy than simple surface creation, it was less useful. George Irwin at the US Naval Research Laboratory tackled the problem in the 1950s by working directly with the stress field around a crack tip rather than the energy.

Irwin showed that for any crack geometry under any loading, the stress field near the tip has the same mathematical form. The stresses all vary as 1/√r relative to the crack tip, where r is the distance from the tip, and they are scaled by a single parameter, which he named the stress intensity factor K. The units of K are MPa√m. Load geometry, crack length, and component shape all enter through K, but once K is known, the stress field is completely determined.

Irwin then defined the material property that governs failure: the critical stress intensity factor K_IC (spoken as "K one see"), where the subscripts stand for Mode I loading and plane-strain conditions. When the applied K_I reaches K_IC, fast fracture begins. This is a material constant, measurable in a laboratory test, tabulated for most structural alloys, and directly comparable to the K calculated from a crack found in service. The forensic consequence is direct: measuring the critical crack length from the fracture surface and knowing K_IC for the material allows calculation of the stress that drove the final fracture. Compare that to the design stress and you know whether the material was inadequate or the load was excessive.

The distinction between plane stress and plane strain matters both for testing and for interpreting a real fracture surface. In a thin plate under load, the material at the crack tip is free to contract in the thickness direction as it yields. That lateral contraction is the hallmark of plane stress. The resulting fracture surface is visibly inclined at 45 degrees through the thickness, the classic shear lip.

In a thick section, the inner part of the crack front is surrounded by material that constrains that lateral contraction. The stress state is triaxial, a condition called plane strain. The material at the tip cannot flow sideways to relieve stress, so it fractures at a lower applied load. The central part of the fracture face is flat, at right angles to the applied stress, while shear lips form only at the free surfaces where plane-stress conditions prevail.

When a forensic engineer examines a fracture surface, the ratio of flat zone to shear lip gives a quick reading of the stress state during fracture. A fracture that is 100% flat zone broke under plane-strain conditions at relatively low stress. A fracture that is entirely shear lip, the so-called full-slant fracture common in thin sheet metal, failed by ductile shear rather than brittle cleavage. Most real failures fall between these extremes, and the proportions change if the crack grew slowly before the final overload, or if temperature dropped, or if the material aged.

LEFM is built on the assumption that material near the crack tip is elastic, with only a small contained plastic zone at the very tip. The stress analysis uses the K parameter derived from elastic theory, which is why the full name specifies linear and elastic. When the plastic zone radius, estimated by Irwin as r_p ≈ (1/2π)(K/σ_y)^2, is a small fraction of the crack length and of the remaining ligament ahead of the crack, LEFM gives accurate results.

• LEFM applies well: high-strength steels and aluminium alloys in structural applications, ceramics, glasses, and most situations where the component is large relative to the expected crack size.
• LEFM becomes unreliable: low-strength, high-ductility alloys (mild steel, austenitic stainless), polymer components, and situations where the applied stress approaches the yield stress of the material.

When LEFM breaks down, engineers and analysts move to elastic-plastic fracture mechanics (EPFM). The two main EPFM parameters are the J-integral, an energy contour integral developed by J.R. Rice in 1968 that is path-independent and remains valid when large plasticity is present, and the crack tip opening displacement (CTOD), which measures how far the crack faces have separated just behind the tip. Both are measured in standard tests (ASTM E1820, BS 7448) and serve the same role for ductile materials that K_IC serves for brittle ones.

Body-centred-cubic metals, which include most structural steels and many ferritic alloys, show a sharp transition in fracture behaviour as temperature drops. Above the ductile-to-brittle transition temperature (DBTT), the material fails with significant plastic deformation and high energy absorption. Below it, fracture occurs by cleavage, with very little plastic work and a flat, shiny surface. The transition can span as little as 20 degrees Celsius, which is why cold-water environments have caused so many structural disasters.

The Liberty ships of World War Two are the most-cited example of DBTT failures in service. A total of approximately 2,710 ships were built rapidly from low-quality steel that had a DBTT above the North Atlantic winter water temperature. When the ships entered cold water, their hulls transitioned from ductile to brittle. Cracks that a warmer hull would have blunted and arrested instead propagated catastrophically, splitting several ships in two at rest in harbour. The investigation that followed established Charpy impact testing, still the standard quality-control check for structural steel, which quantifies the energy absorbed during fracture at defined temperatures.

In failure investigations the fracture surface is the primary evidence. Fracture mechanics provides a bridge from what you can see, the crack size, to what you want to know: the stress at the moment of fracture. The key equation for a through-crack in a plate is K = Yσ√(πa), where Y is a geometry correction factor, σ is the remote stress, and a is the crack half-length. Setting K = K_IC and solving for σ gives the fracture stress. Setting K = K_IC and solving for a gives the critical crack size for a known design stress.

1. Identify the fracture origin
   Locate the point on the fracture surface from which crack-front lines (chevrons, river marks) radiate. This is where fast fracture began, and it defines the initial crack length a.
2. Measure the critical crack dimension
   Using a calibrated scale from photography or SEM imaging, measure the size of the pre-existing defect or the region of slow crack growth that existed before the final fast fracture step. This is the a term in the K equation.
3. Select the K_IC value
   Retrieve K_IC for the specific alloy, temper, and temperature from published databases (NIST, ASM), test certificates, or, when those are unavailable, from hardness-correlated estimates with appropriate uncertainty.
4. Back-calculate the fracture stress
   Solve σ = K_IC / (Y√(πa)). Compare to the design stress range. If the calculated fracture stress lies within the normal operating envelope, the critical crack should have been detected by inspection. If it lies above the maximum design load, an overload event is implicated.

This back-calculation is routinely used in aircraft accidents, pressure-vessel ruptures, and crane or lifting-gear failures. Its limitations include uncertainty in K_IC if the material had not been tested at the service temperature, uncertainty in the geometry factor Y for complex shapes, and uncertainty in the crack dimension if the fracture surface has corroded or been damaged after failure. Careful analysts document these uncertainties explicitly and present a range of fracture stresses rather than a single number.