Step 2 of 5
38%
Surface treatment sequencing: nitride, PVD, polish, in what order
The build order that takes an H13 forge die from rough block to coated, polished, ready-to-run, and why getting any step out of sequence is a failure mode rather than a slow ramp to nominal life.
Step 2 of 5Why PVD on non-nitrided H13 spalls
H13 at 48 HRC measures around HV 550 at the surface. A 3-5 µm PVD coating at HV 2500-3500 sits on top of that substrate. Hot-forge contact stress runs 800-1500 MPa peak, well above the substrate's yield strength at temperature. The substrate yields plastically under load. The coating, which is rigid and brittle, cannot follow the yielding substrate. It cracks at the interface, and once cracked it spalls in sheets. This is eggshell failure, named for the way a hardened shell over a soft interior fractures under point load.
Nitriding fixes the eggshell by building a deep hard diffusion zone underneath the coating. A 0.30 mm diffusion zone at HV 750-900 supports the PVD coating through the full contact stress profile. The load path becomes PVD at HV 3000, then diffusion zone at HV 800, then bulk H13 at HV 550, then unhardened core. No soft layer in the stack means no plastic yield under the rigid coating and no spalling. This is what duplex treatment means in practice. PVD without nitride is not a faster duplex. It is a different and worse die.
The exception is PM grades through-hardened above 58 HRC. Caldie, ASP 2080, and similar PM tool steels are hard enough through-section to support PVD without an intermediate nitride. For H13 at typical forge hardness, the exception does not apply.
Quick check
A vendor proposes PVD direct on H13 at 48 HRC to save the nitride step on an aluminum forge die. Why does this fail in the first few hundred hits even if the coating is deposited correctly?