<title/>In coating applications, magnetron sputtering is frequently the technology of choice. In industrial continuous in-line sputtering processes, such as web coating and glass coating, rotatable magnetrons have, to a large extent, supplanted the more classic rectangular sputter magnetrons and constitute the heart of many coating systems with target lengths up to 4 m. However, for many people in the coating business, rotatable magnetrons are quite unknown. When comparing the two types, it looks as if rotatable magnetrons offer significant advantages over planars, as long as the required cylindrical targets are available. Rotabable magnetrons are still a relatively new technology, and have not been fully investigated on a lab scale due to the lack of availability of small systems. Experiments with an in house developed downscaled rotatable magnetron have allowed some of their typical characteristics to be identified. Especially in reactive sputtering, rotatable magnetrons exhibit unusual properties. In this article, it is shown that this peculiar behaviour can be understood as the result of redeposition of previously sputtered material back on the target outside the current discharge plasma zone. This redeposition is, to a large extend, also responsible for the poisoning characteristics as a function of rotation speed of the cathode cylinder. It is also shown that in reactive sputtering, rotatable magnetrons generate two beams of high energy and negatively charged particles which seriously influence the morphology of the growing layer. In between both beams, there is a zone of silence where only sputtered low energy material is deposited and in which fragile materials, such as perovskites can be grown.
Read full abstract