EDM after heat treat — recast, stress, and the nitriding handoff

The handoff failure is structural, not certificate-level. Every certificate in the chain reads correctly. The heat-treat cert says 48 HRC. The EDM operator's measurement says Ra 1.6 µm on the cavity face. The nitride cert says case depth 0.32 mm, surface 1050 HV, compound layer 6 µm γ' dominant. The die spalls at 4,200 hits.

The failure mechanism is that nitriding diffuses nitrogen into whatever surface is presented to the retort. If the surface is sound tempered substrate, the case forms cleanly and the structural integrity follows the case depth. If the surface is a 20 µm recast layer with vertical microcracks, the nitrogen diffuses into the cracks first because the crack interiors expose more surface area per unit volume than the planar bulk, and the diffusion zone forms around and through the existing fracture network. The case looks correct on a hardness traverse because the nitrogen did diffuse to depth, but the substrate it diffused into was already broken. The compound layer that forms on the recast surface is anchored to a cracked layer that fractures under contact load and takes the compound layer with it.

The nitride vendor cannot detect this from their side. Their inputs are time, temperature, atmosphere, and the part as received. Their certificate reports what their cycle produced. The integrity of the substrate is the EDM operator's deliverable and the tool builder's verification, not the nitride vendor's responsibility. The handoff document that prevents the failure is a recast-depth measurement on a witness coupon from the same EDM cuts, signed off before the part ships to nitride.