High-power devices require a thick silicon layer with high resistivity to achieve high breakdown voltages. In order to optimize the turn-on and turn-off behavior of these devices, precise adjustment of the carrier lifetime is necessary, together with knowledge of its dependence on process parameters and operation conditions ͑e.g., temperature and injection level͒. We have therefore analyzed the carrier lifetime of the starting material ͑float-zone silicon͒ as well as processed devices that have both been either doped with transition metals ͑Pt, Au, Fe͒ or irradiated with electrons, using two different measurement techniques based on photoconductivity decay and free carrier absorption measurements.The fabrication of high-power devices usually requires n-type silicon with high resistivity. For a blocking capability of 8 kV, for example, material with a resistivity of about 500 ⍀ cm and a wafer thickness of about 1.5 mm is necessary. 1 Furthermore, high-power devices should also be able to switch high currents of several kiloamperes, requiring large device diameters of up to 6 in. In contrast to integrated circuits, almost the total volume of the wafer is electrically active in the case of high-power devices. Therefore, the concentration of contaminating atoms in the starting material must be very low (10 11 atom/cm 3 ), typically two orders of magnitude smaller than the fairly small n-doping concentration. In addition, the lateral and vertical distribution of the doping atoms must be very uniform. Since the electrical performance of the devices, as, e.g., the switching behavior, can often be optimized by reducing the charge carrier lifetime in the electrically active region, the active area ͑or part of it͒ is doped with transition metals such as Pt or Au, or irradiated with electrons or/and light ions. Sometimes it is advantageous to combine these techniques; a typical example is reducing the carrier lifetime uniformly by diffusion of Pt atoms and locally in a certain depth below the wafer surface by means of ion irradiation. Since the operation of power devices involves a wide range of temperature ͑230-450 K͒ and current density (10 Ϫ5 to 400 A/cm 2 ), understanding the temperature and injection-level dependence of the carrier lifetime is important for device optimization. Although a number of investigations has already been performed on temperature and injection-level dependence of as-grown silicon, 2 of silicon doped with metal impurities ͑e.g., Ref. 3-7͒, and of irradiated silicon ͑e.g., Ref. 8-10͒, the characterization and understanding of lifetime dependence is still incomplete. 11 In this paper we present carrier lifetime measurements on silicon wafers and processed devices using two different experimental techniques: microwave photoconductive decay ͑-PCD͒ measurements and free carrier absorption ͑FCA͒ measurements. Both the temperature and injection-level dependence of the carrier lifetime can be measured with either of these contactless techniques. Since bulklifetime measurements on pure as-grown silicon o...

Valuable information about defect profiles and defect concentrations in high-power semiconductor devices can be obtained by analyzing electrical device characteristics. This is demonstrated by evaluating reverse current-voltage characteristics of p-n junctions, from which the vertical radiation-induced defect profiles of the dominant generation center can be extracted. Measured data from proton-irradiated high-power diodes find a reasonable interpretation when radiation-induced doping effects, as obtained from spreading resistance measurements, are taken into account. Further investigations focus on defects in electron-irradiated power metal oxide semiconductor transistors, which were analyzed by stationary and dynamical diagnostic methods in combination with device simulations. Equipped with a detailed understanding of the action of radiation-induced defects, we make use of it in order to tailor certain characteristic electrical properties of high-voltage devices by exploiting carrier-trapping effects as well as radiation-induced changes in resistivity. The main focus lies on the blocking voltage and the switching behavior.

Since power devices require an electrically active, thick n‐type silicon layer with high resistivity and a large area, their electrical characteristics are extremely sensitive to contamination. If transition metals diffuse into the wafers during the high temperature steps required for device fabrication, an uncontrolled increase in leakage current and on state voltage can be observed. Furthermore, current filamentation and instabilities of the electrical data can occur. As a consequence of the low doping level of the n‐base, the blocking voltage and the failure rate due to cosmic radiation are sensitive to contaminating atoms acting as donors or acceptors. To obtain information about the sources and the extent of the contamination, a contamination check was performed after each high temperature step. Moreover, the dependence of the carrier lifetime on temperature and injection level was analyzed for typical operation conditions of power devices. The results proved to be important in finding ways of keeping contamination and silicon defect densities as low as possible and to ensure that good electrical data with adequate stability could be obtained. © 2000 The Electrochemical Society. All rights reserved.