AVS 52nd International Symposium
    Advanced Surface Engineering Tuesday Sessions
       Session SE-TuM

Invited Paper SE-TuM3
The Origin of the Hysteresis Effect in Reactive Sputtering Processes

Tuesday, November 1, 2005, 9:00 am, Room 201

Session: Hard Coatings and Wear Mechanisms of Protective Coatings
Presenter: S. Berg, Uppsala University, Sweden
Authors: S. Berg, Uppsala University, Sweden
T. Nyberg, Uppsala University, Sweden
O. Kappertz, Uppsala University, Sweden
T. Kubart, Uppsala University, Sweden
D. Rosen, Uppsala University, Sweden
Correspondent: Click to Email
Sputtering is a well known and widely used thin film coating technique. The way the energetic incoming particle generates a collision cascade in the target that causes some of the surface atoms to be ejected is reasonable well understood. Since there is negligable consumption of the inert gas (normally argon) process control is not a problem. The sputtering process may easily be modified. By adding a reactive gas to a sputtering system it is possible to obtain a reactive sputtering process. This process, however, exhibits a significantly more complex processing behaviour than inert sputtering.The processing curves will form hysteresis loops where the processes have a tendency to make avalanche like abrupt transitions at the edges of the hysteresis width. Before the transition a metal rich film may be formed at high deposition rate while after the transition a compound film will be deposited at a low deposition rate. Without a sophisticated feedback control system no compositions in between these two extremes can be reached. Unfortunately the most attractive processing point is located at one of the critical edges of the hysteresis loop. This complicates industrial production carried out by these processes. The aim of this presentation is to describe the cause of the hysteresis loop and point out how different processing parameters will influence the hysteresis. The possibility to eliminate the hysteresis and obtain non-critical stable processing conditions will also be discussed. In addition comments will be made about common mis-information in reporting data from reactive sputtering processing experiments.