This tutorial is designed to describe various impedance matching network problems, troubleshooting techniques and solutions. System level and work bench troubleshooting techniques will be discussed.
The automatic matching network does not properly tune.
Assumptions; this unit and the system its installed into was previously operational, the plasma does ignite, the match can be manually controlled via its controller (capacitors do move or the display indicates some activity).
Symptoms; the matching network continuously hunts for the low reflected power point but never locates it. The plasma ignites but may go out during the back and forth tuning.
Troubleshooting:
Can the match be operated in the manual mode and the plasma be properly tuned?
If a DC bias voltage is typically present during normal operation, is there now DC bias? Is it lower than normal? Does it drop to zero at any time during the tuning process?
What has changed since this component last worked properly?
Does the match operate at lower than normal power levels but not at the higher (typical) process power?
Diagnosis:
If the plasma can be matched manually but not automatically, the problem may be a mis-adjusted phase/magnitude detector or another part of the unit’s servo motor control circuit.
If the plasma cannot be tuned in either manual or auto modes, there may be a problem with the mechanical drive train to one or both capacitors. Check to make sure the linkage screws and clamps are tight. Make sure the capacitors are properly set – when the position meter displays 0% the capacitor should be set at zero – at 100% the capacitor should be close to or at 100% of its range. (Most problems are related to mechanical issues and not electrical.)
If there is no or low DC bias, check the plasma electrode for cleanliness and absence of conductive flakes or other types of shorts. Also inspect the water cooling lines to the electrode (magnetron cathode or substrate stage) – if the inner surfaces of the tuning is coated with rusty looking deposits, this can reduce the power getting to the plasma as well as causing a tuning failure.
If the DC bias suddenly drops to zero, this indicates a plasma short – caused by conductive particles or other intermittent grounding of the (in vacuum) power lead or electrode insulators.
If the plasma electrode’s power lead is mechanically movable (i.e., connected to a height adjustable substrate table) verify that the power conductors are making good contact.
Repairs:
Replace rusty cooling water tubing.
Re-align the matching networks capacitor drive mechanism.
Replace faulty power conductors, insulators and hardware.
Clean the vacuum process chamber and its internal components.
