The purpose of this work was to use Scanning Acoustic Microscopy (SAM) for identifying and characterising defects in power semiconductor devices1,2. Power semiconductor devices are a special class of active solid state switches. From an ideal point of view they do not differ from their low-power counterparts. However in order to carry thousands of ampere and to sustain thousands of volt specific technological requirements are necessary. A general characteristic common to different types of devices, like diodes, bipolar and MOS transistors, thyristors and triacs, is the large active area involved in the conduction of current and heat dissipation. A major problem in these devices is to ensure proper electrical and thermal contacts over large silicon surfaces, in order to guarantee an uniform density of current. This requirement is more and more important as the diameter of the active region increases to values in excess of 100 mm. Unlike microelectronics, where hundreds identical devices are manufactured on one single silicon wafer, a single power electronic component often uses the whole area available on the wafer. It is a well established fact that semiconductor devices are difficult to be scaled up to larger areas, as the increasing probability to have a crystal defect in the active area limits the production yields. Luckily power devices are less sensitive to crystal defects than integrated circuit are. However, manufacturing large area devices with a properly controlled levels of defects or unwanted impurities is often a formidable task. With the aim to improve the production yields, process control has been given more and more attention. Improving process control strategies requires the availability of new equipment and techniques for non destructive characterisation of the impurities and defects, either introduced on purpose or inadvertently, which can affect the electronic band structure of the semiconductor material and hence the electrical characteristics of the device. The aim of this research work is the development of a non-destructive test method for on-line monitoring of the manufacturing process for power semiconductor devices. The first part of the work was devoted to the calibration of the different SAM equipment available for the research. The core of the Project is the SAM investigation of samples containing a known amount and distribution of defects, as cross-identified by alternative analytical techniques. In the following a number of results achieved to date are presented. Next section reviews the essential characteristics of power semiconductor devices. Then the experimental methods used in the research are presented and discussed. Finally the results obtained to date are presented with the aim of demonstrating the feasibility of an industrial system for online monitoring of power semiconductor device manufacture.