Failure Is Percolation at Critical Defect Density

01The evidence

• Brittle fracture in materials with random defects has been analysed as a statistical problem in which defect population, not a single critical flaw, controls failure.OSTI, brittle fracture with random defects
• Microcrack nucleation models produce damage evolution that becomes collective at high density, which is the regime where a percolation description applies.JMPS, damage evolution

02The tension

Weibull statistics is the standard engineering tool: strength is set by the largest flaw, and size effects follow from extreme-value theory. It works well for small, clean samples and fails for large or damaged ones, where strength becomes insensitive to the largest defect and sensitive to density instead.

03The theory

Defects are distributed randomly. As density rises, the probability that a connected path of defects spans the sample rises sharply, and at a critical density such a path exists. Failure follows the path, not the largest single defect.

That makes strength a function of density through a power law with an exponent set by the percolation universality class, which is independent of the material. The size effect also changes: below the critical density, larger samples are weaker because they are more likely to contain a big flaw; above it, the density effect saturates.

The prediction is a crossover. Small samples should follow Weibull, large or heavily damaged ones should follow percolation, and the crossover point should be predictable from the defect density.

This matters for additive manufacturing, where defect density is high and variable, and for structural health monitoring, where damage accumulates before failure.

04What this is not

05Predictions

• P1Strength should scale with defect density through a power law whose exponent matches percolation universality.
• P2The size effect should saturate for samples above the critical density, contrary to Weibull.
• P3Acoustic emission before failure should show a connected cluster spanning the sample.

06What would refute it

Strength that tracks the largest defect at all densities, with a Weibull size effect that never saturates, would leave the percolation picture without a regime.

07What it would change

It would give additive manufacturing a density-based acceptance criterion, and replace an extreme-value framework with a collective one where it fails.