Etch Rate Troubleshooting
Etch rate troubleshooting is the process of diagnosing why a chamber's etch rate has drifted from its qualified baseline — too fast, too slow, or uneven across the wafer — by working through the same variables that determine etch rate in the first place: RF power delivery, gas flow and chemistry, chamber pressure, temperature, and consumable condition. Etch rate uniformity (how consistent the rate is across the wafer) and overall etch rate (how fast it is on average) are two different symptoms with overlapping but distinct root causes, which is why a competent troubleshooting approach treats them as related but separate questions rather than one undifferentiated "etch is wrong" problem.
A global rate shift (fast or slow everywhere) and a uniformity shift (fine in the center, off at the edge, or vice versa) point at genuinely different subsystems, which is the first branch point in any methodical diagnosis. A uniform, chamber-wide rate change most often traces back to something that affects the whole wafer equally — RF power level, overall gas flow, or chamber pressure. A uniformity problem that's specifically worse at the edge, by contrast, points much more strongly at the process kit (see this site's chamber process kit entry) or at chamber wall conditioning, since those affect the field and gas distribution unevenly across the wafer rather than scaling the whole process up or down together.
MFC calibration drift is one of the most common and most overlooked causes of a global rate shift, because the flow controller can report its commanded setpoint correctly while actually delivering a slightly different flow. A mass flow controller that's drifted out of calibration doesn't throw an alarm — it keeps reporting the setpoint value the recipe commanded while the true gas flow quietly differs from that value, which shifts etch chemistry and rate without any obvious fault anywhere in the system. Periodic MFC verification against a calibrated reference, not just trusting the controller's own readback, is part of why chamber-calibration-setup's sensor calibration step exists as a distinct checkpoint rather than being assumed correct indefinitely.
Chamber wall conditioning state is a real, if less intuitive, contributor to etch rate drift, because the walls themselves become part of the chemistry over time. As a chamber processes wafers, reaction byproducts deposit on interior walls, and that accumulated film can itself react with process gases or alter the chamber's effective surface chemistry — which is why many etch processes include a periodic wall-clean or seasoning step, and why a chamber's etch rate can drift measurably between a freshly cleaned state and one approaching its next scheduled clean, independent of anything wrong with RF, gas, or the process kit.
Etch rate troubleshooting is ultimately a process of elimination through this same short list of variables every time — RF delivery, gas and chemistry, pressure, and consumable/wall condition — and reading whether the drift is global or edge-localized, which this site's RF power delivery troubleshooting and chamber process kit entries each cover in more depth, is usually the fastest way to narrow that list before touching anything on the chamber.