On the fab floor, thermal drift in rapid thermal processing eats into the thermal budget, and photoresist profiles start to collapse. Wafer-level uniformity isn’t some academic metric—it’s the line between a profitable run and a bucket of scrap. What actually matters under the hood We spec halogen lamps for RTP with short-wave output and quartz envelopes because they get the wafer hot, fast, and repeatable. You can hold wafer-level uniformity at ±0.1°C, and the temperature stays stable even when the line is running 24/7. The filament geometry and reflector path are tuned so the thermal profile stays consistent lot after lot. That’s how soft bake and hard bake follow the recipe, instead of chasing tool drift. Why it holds up in production In Class 1–100 cleanrooms, particle generation is the quiet yield killer. These lamps run without shedding particles, so you’re not contaminating lithography and downstream layers. Photoresist baking becomes predictable, which cuts rework and tightens cycle time. Energy use is trimmed without compromising ramp rates, so cost per wafer comes down while the chamber stays stable. The payoff is fewer excursions, tighter CD control, and uptime you can actually bank on. What you need to get right Installation comes down to precise lamp alignment and calibration to the chamber’s emissivity and thermal mass. Matching the base, connector, and voltage to the host RTP tool is mandatory—mismatches create hot spots and burn out filaments early. Schedule PMs based on actual lamp hours, not the calendar. Treat the lamp like the calibrated component it is, and the process repeatability will follow.