At Applied Physics, Linköping University, the main interest for sensor research has been the focused on MISFET (Metal-Insulator-Semiconductor Field-Effect Transistors) and related devices. The gas sensitivity of MISFET’s was first reported by Lundström et al. in 1975. Today a lot of work is being done in this field. There has also been a center of excellence established, S-SENCE, with participants from the university as well as from Swedish Industry, devoted to sensor research. The research is focused on basic studies about the functionality of MISFET's as well as applied work in which the focus is to implement the sensor technique at a variety of applications in different industries. One of the large problems with chemical sensors (as for most types of sensors) is the poor long-term stability. Chemical sensors are usually in direct contact with the ambient that is measured, and since a lot of the gases that are measured are corrosive and hostile, this affects the sensors. This problem is approached in two ways where one concerns data correction and drift reduction in the received data sets and the other tries instead to increase the stability of the active part of the sensor, by learning more about the materials in the active part and to improve them. For all types of chemical gas sensors the lack of long-term stability can be related to the fact that the sensors are interacting directly with the ambient that is measured. Therefore, the same type of problem arises for many types of chemical sensors. The active layer may change during gas exposure and thereby change the response patterns with time. The ultimate goal of this work is to obtain long-term stable sensors that can be processed in a controlled and reproducible way. The work presented in this thesis is a first step towards this goal Attempts have been made to understand how processing parameters affect the sensor responses, and how to find possible ways to improve the endurance of the film. By carefully controlling the deposition conditions and analyzing the resulting films with respect to morphology, crystal structure and interfacial properties, it is possible to correlate sensor characteristics to material properties. Through this knowledge, new information can be gained when it comes to understanding the mechanisms behind sensor responses. Testing new materials for the gate is interesting for many reasons. Some of these are to minimize known problems such as redistribution of the gate metal and reduction of the effect of the corrosive atmosphere that sensors are subjected to.