One of our customers described a situation where the current and the voltage in the plasma became unstable when depositing a highly conductive elemental metallic sputtering target with a magnetically enhanced cathode assembly which had been in continuous operation for over ten years. The current in the plasma had shot up and the voltage had fallen off. They also experienced severe arching in the plasma between the cathode assembly to ground potential causing the DC power supply to shut down and restart sporadically.
There may be two possibilities occurring, either separately or in conjunction with each other, causing this instability. If the magnets in the body of the magnetically enhanced cathode assembly had become overheated at some point they, may have reached their Currie temperature and become either partially, or totally, demagnetized. This could occur if there was any type of decreased water flow or cooling efficiency in the water cooling system facilitating the deposition system or blockage of some type in the water channels of the cathode body itself. This blockage may be a result of a chemical build up from minerals in the water (if it is untreated “city” water) or from reactive products associated with the caustic water and galvanic action of dissimilar metals used in the cathode construction. Loss in the magnetic field at the target surface may also be a result of magnet material erosion if the magnets are produced from rare earth elements, Many rare earth magnets corrode if they are in direct contact with the cooling water. Such magnets are usually potted with some form of organic protective coating or plated with an erosion resistance material. These materials can often break down over time thus allowing the magnets to erode and lose mass and magnetic strength. In general, it is necessary to have a minimum of 600 Gauss within the erosion profile of the target to sustain a plasma.
With a reduced, or a complete loss of the magnetic component in the MxB component of the power generating the plasma, the power level within the plasma will be severely reduced, since only the electrical component will be present on the target surface. This will critically alter the deposition rate or completely prevent any plasma from being ignited and sustained within the cathode assembly, even though the sputtering target itself is totally functional.
Alternatively, it may be that the plasma generated by the magnetically enhanced cathode assembly is functioning normally, but it is arching over to ground potential. This may be caused by stray deposition material building up on various parts of the system components (shutter assembly, dark space shield, chamber wall, etc.) and eventually flaking off and creating an electrically conductive path between the live cathode (negative DC current) to the dark space shield (ground potential) inside the chamber. Care should also be taken to make certain that the distance between the dark space shield and the cathode assembly is uniform and properly positioned. The gap should typically be around 6mm and uniform. The dark space shield should extend to, and be uniform to, the top of the top surface of the sputtering target. This may require repositioning if targets of different thicknesses are interchanged within a particular cathode assembly. If the inner diameter of the dark space shield (can) and the outer diameter of the sputtering target are not equally positioned (so that the gap between the two components is not uniformly equal) i.e. too close in some places and too far apart in others, this will prevent the ground potential shield from extinguishing the plasma in the localized area outside the erosion profile that is created by the magnetic field within the cathode assembly.
