Plasma Strike

Plasma strike (ignition) is the moment a chamber's gas transitions from a non-conductive neutral state into a conductive plasma — a free electron, accelerated by the applied RF field, collides with a gas molecule hard enough to knock loose another electron, and that single event cascades into an avalanche of ionization within microseconds. Every plasma process in the fab, from etch to PECVD to plasma ashing, depends on this ignition event happening reliably and consistently, because a chamber that strikes late, strikes inconsistently, or fails to strike at all turns an otherwise-correct recipe into a wasted or defective wafer before any actual processing begins.

Striking requires more RF field strength than sustaining does, which is why generators often briefly overdrive power at ignition before settling to the recipe's steady-state setpoint. Breaking down a neutral gas into plasma takes a higher electric field than keeping an already-ionized gas conductive — once free electrons and ions exist, far less energy sustains the cascade than was needed to start it. Many process recipes exploit this directly with a deliberate ignition pulse: a brief spike of higher RF power specifically to guarantee breakdown, followed immediately by a drop to the lower power level the actual process step calls for.

UV pre-ionizers exist specifically because waiting for a random background electron to strike the first ionization event is too unreliable for production throughput. Cosmic rays and natural background radiation do produce free electrons in any gas eventually, but relying on that random event to start every single plasma step would make strike timing unpredictably variable, wafer to wafer. A small UV lamp or auxiliary ionization source deliberately seeds the chamber with free electrons right before RF is applied, making strike timing fast and repeatable instead of a probabilistic wait — which matters directly for wafer-to-wafer process consistency and overall tool throughput.

A chamber that strikes inconsistently or fails outright is almost always reporting a problem somewhere else in the RF or gas delivery chain, not a fault in "striking" as its own isolated failure. Low or unstable chamber pressure, insufficient RF field strength reaching the electrode (a degraded cable or an undertuned match network — see this site's RF power delivery troubleshooting entry), or a contaminated/worn electrode surface that raises the effective breakdown threshold can all manifest as the same symptom: the chamber simply won't strike, or strikes late and inconsistently. Troubleshooting a strike failure means checking those upstream systems rather than treating ignition itself as the broken component, since the ignition event has no moving parts of its own to fail.

Plasma strike is easy to take for granted because it happens in microseconds and rarely gets its own line in a process recipe, but every etch rate, deposition rate, and uniformity number downstream of it assumes the plasma actually ignited cleanly and on schedule — a chamber with strike problems produces process drift that looks like it belongs to a completely different part of the tool.