Reactive sputtering is the sputtering of an elemental target in the presence of a gas that is purposely added to the system to react with the sputtered species to form a compound film. Although simple in theory, reactive sputtering can be difficult to carry out if one wants to obtain high deposition rates and good properties in the deposited films. The classic problem with reactive sputtering when the flow control of the reactive gas is used to regulate the inlet of the reactive gas into the chamber is that the surface of the target will suddenly change from the elemental state to a compound state when the flow of the gas is increased beyond a certain point. When this happens, the target becomes covered with the compound ("poisoned state"), and the sputtering rate drops precipitously while the composition and properties of the deposited film change. To avoid the problems associated with flow control of the reactive gas, it is possible to increase the pumping speed of the system, which will prevent avalanching from the metallic state to the poisoned state for most materials. However, this technique will not prevent the loss in deposition rate. To avoid the loss of rate from target poisoning and a degradation of film properties, it is necessary to control the partial pressure of the reactive gas. Partial pressure control is more involved than flow control of the reactive gas, but it does provide high deposition rates, good composition control, and excellent film properties. For partial pressure control to be effective, a feedback signal that varies as the partial pressure changes is required. Common feedback signals are the cathode voltage, a partial pressure signal from a mass spectrometer, or an optical emission signal from the sputtering plasma. Each has its advantages and disadvantages. Signal acquisition, processing, and response times are all very important for partial pressure control. If the update time of the feedback signal or if the response to a change in the control signal are too long, it may not be possible to maintain a stable reactive gas partial pressure. The feedback signal must be obtained and processed quickly, and the response to any change must also be quick. If this is done, it is possible to operate at any point along the transition from the metallic state to the poisoned state of the target, which means that it is now possible to achieve both high deposition rates and good film properties. Reactive sputtering of insulating films presents special problems because arcing can occur on the target surface, which leads to instabilities in the process. Fortunately, arcing can be eliminated by using a power supply that can suppress or prevent arcing. For deposition of insulating films, it is necessary to combine the correct type of power and with partial pressure control of the reactive gas to generate a stable, high-rate reactive sputtering process. One without the other will not produce the optimum results