CoatingIQ
← Course index

Step 2 of 5

38%

Lesson 10·QC and failure

Failure modes — white-layer spall, network cracking, distortion, embrittlement

Five ways a nitrided tool fails in service, how to read each one from a microsection, and when 'the nitride failed' is honestly 'the wrong recipe was specified.'

11 min readLesson 10 of 12

Step 2 of 5Compound-layer network cracking at sharp corners and edges

Visual: hairline cracks that follow part geometry, almost always concentrated at internal radii and sharp edges. On a forging die the pattern traces the flash land and the corner-of-impression features. On a stamping die it traces the punch entry edge.

Microsection: cracks initiate in the compound layer at the corner, propagate into the diffusion zone, and arrest somewhere in the depth-decreasing hardness gradient. The fracture face is brittle, with no plastic deformation evident around the crack tip.

Cause: stress concentration plus a brittle compound layer. The compound layer cannot follow the strain at a sharp internal radius. The substrate underneath is fine; the case is the failure path.

Rule of thumb on corner radius for nitrided tooling: internal radius should be at least 10× compound layer thickness. An 8 µm compound layer wants an 80 µm (0.003") radius at the inside corner. Tighter than that and the compound layer cracks during the first few thermal or load cycles regardless of cycle quality. For high-impact features, the Journal of Materials Engineering and Performance paper on optimized compound-layer design for highly loaded gears makes the same point quantitatively: corner radius drives crack initiation more than absolute compound-layer thickness does.

Three fixes, in order of effectiveness. Radius the corner before nitriding. Reduce compound-layer thickness on features that cannot be radiused. Polish or grind the compound layer off the corner after nitriding (only valid if the spec allows compound-layer removal and the diffusion zone still meets the case depth requirement after the stock comes off).

Quick check

A blanking die has a 0.001" (25 µm) sharp internal radius at the punch entry corner and is being quoted for a 12 µm ε compound layer FNC cycle. Will the corner crack, and at what radius does the geometry pass the 10× rule?