The metallurgy you actually need — Fe-N phase diagram in plain English

What this means on the shop floor

• A "compound-layer-free" claim is plausible at Kn ≈ 0.2 and below. Not at Kn = 3. If a vendor claims compound-layer-free at any Kn above the α/γ' boundary, ask for a microsection.
• A vendor running fixed ammonia flow with no H₂ measurement is delivering whatever the furnace produces. Dissociation rate drifts with load, retort condition, and time of day. Across a 40-hour cycle the actual Kn can vary 2-5×.
• Running an H13 die at 600°C to shorten the cycle is not a shortcut. It moves the substrate above its temper temperature, pushes the Fe-N system into the γ field, and produces a soft core under a coarse-nitride case. Surface hardness on the certificate reads as expected, the part fails in service well below its rated cycle count.
• Specify ε for sliding wear (stamping, mold cavities) and γ' for impact (cold heading, forge punch). Either call it out on the print or accept the vendor's default recipe.

Pushback questions for the vendor

1. What Kn setpoint are you running for my part, and what tolerance band do you hold it to?
2. How do you measure Kn: H₂ sensor on the exhaust, burette dissociation reading, or calculation from flowmeter setpoints?
3. Which phase are you targeting in the compound layer, and can you confirm with a witness coupon cross-section?
4. What is the control loop's response time when retort load shifts dissociation mid-cycle?

Common confusions

• ε and γ' get swapped routinely in vendor literature. Mnemonic: ε is high-nitrogen (harder, more brittle). γ' is low-nitrogen (more ductile).
• HRC on a nitrided surface is almost always wrong. The Rockwell C indenter penetrates 150-200 µm into the surface, deeper than most compound layers and a significant fraction of the diffusion zone. The reading is a weighted average of case and substrate. Insist on HV or HK microhardness at a defined load and depth: typically HV 0.05 (50 gf) for the compound layer and HV 0.5 (500 gf) for the diffusion zone, per ASTM E384.
• "Case depth" without a definition is ambiguous. DIN 50190-3 NHD is the depth where hardness equals core + 50 HV. SAE-style effective case depth is the depth to an absolute hardness, often HV 400. The two numbers can differ by 2× on the same part. Pin down the definition before agreeing to a number.
• HV-to-HRC conversions per ASTM E140 assume bulk homogeneous material. Applied to a 10 µm compound layer, the converted HRC number does not mean what it would on a uniform specimen.

Up next: gas nitriding.