The bread-and-butter job done right: stoichiometric gold δ-TiN, 2000–2500 HV, locked color and adhesion every run.
DEMO — generic parameters
The problem: Gold one run, muddy brown the next — and coatings flaking under load. Color and adhesion drift because the reactive point and the interface aren't controlled.
🧹 Substrate Preparation
CleaningUltrasonic: acetone 10 min → IPA 10 min → DI rinse → N₂ dry
Bake-out150°C, 30 min min under vacuum
Plasma EtchAr plasma, RF bias −300 to −500V, 10–15 min
🔧 Vacuum Conditions
Base Pressure≤ 5 × 10⁻⁶ Torr
Working Pressure2–4 mTorr
🔒 Locked
🎯 Main Deposition — the dial-in
Power350 W DC (49 W/in²)
N₂ Ramp ProtocolStart 0 sccm, +0.5 sccm / 30 s until discharge V drops 12%, lock
Substrate Bias−80V RF (stress + color control)
Ti Interlayer75 nm, 300 W DC, Ar only
Rate / Thickness3.8 Å/s → 2.5 µm
📊 Expected Result
ColorGold (stoichiometric δ-TiN)
Hardness2000–2500 HV
Adhesion>50 N critical load (Rockwell scratch)
⚠️ Failure Mode Warnings
Dark brown/gray film → N₂ insufficient, substoichiometric