On the fab floor, thermal excursions aren’t theoretical—they show up as line-width swings and yield hits. In RTP, the lamp is the knob you’re actually turning. If the heat profile drifts, the photoresist profile falls apart, and you’ve just burned through thermal budget for nothing. What matters, technically We spec halogen lamps for RTP with short-wave infrared so the silicon absorbs the heat directly and fast. That gives you sub-second ramp rates without overshoot. Across the active zone, wafer-level temperature uniformity stays within ±0.1°C—and that kind of control is what keeps critical dimension performance consistent. The quartz envelope and the filament geometry are engineered to hold output steady over 5,000+ hours, with intensity drift under 5%. Match the output to the thermal mass of the process: fast when you need spike anneal, controlled when you’re running soft bake and hard bake steps. Why it holds up in production You’re running Class 1–100 cleanrooms, and particle count is not a discussion. These lamps are built to avoid particle events during thermal cycling, so you’re not fighting the chamber or the wafer. Temperature repeatability from batch to batch cuts down on re-qualification and scrap. Tighter uniformity means fewer edge rejects. Stable output means you stop wasting energy on compensatory over-driving. And because the lamp responds quickly, you can tighten recipe times and move the cost-per-wafer in the right direction. Things to keep straight Installation is straightforward, but the lamp performance hinges on alignment—optical path and chamber reflectivity. After a lamp swap, give it a short warm-up and stabilization window before you expect everything back in spec. Double-check power and connector ratings against your RTP platform, and confirm the spectral match to your absorber stack. That’s the step that makes the thermal profile predictable instead of guesswork.