Coupled measurement-simulation procedure for very high power fast recovery – Soft behavior diode design and testing F. Bertoluzza a,b , P. Cova a, * , N. Delmonte a , P. Pampili b , M. Portesine b a University of Parma, Dipartimento di Ingegneria dell’Informazione, Viale G.P. Usberti, 181/A, Parma, Italy b POSEICO s.p.a., via Pillea, 42-44, Genova, Italy article info Article history: Received 2 July 2010 Accepted 19 July 2010 Available online 8 August 2010 abstract Reliability requirements for very high power devices are growing for their importance in industrial drives and renewable energy; testing those devices in operating condition is more and more difficult. A coupled measurement-simulation based design procedure is presented and applied to high power PiN diodes for application with fast IGBTs or IGCTs, in which high di/dt’s can result in too high current or voltage or energy peaks during turn-off, limiting the reliability of the circuit. Appropriately tuned simulation of the semiconductor device embedded in the test circuit allows to overcome measurement capability limits and to properly design the diode itself and a specific test circuit. Ó 2010 Elsevier Ltd. All rights reserved. 1. Introduction Nowadays more applications in industrial drives, renewable en- ergy as well as wind mill, heavy electrical traction are demanding for high power concentration from 2–3 MW to over tens of MW. In this case semiconductors with voltage blocking capability of 4.5 and 6 kV are commonly applied. This market is asking for very high level of reliability with appropriate testing procedure, in order to cover final application requirements. Design of very high power PiN diodes for applica- tions involving fast switching components, such as IGCTs and IGBTs, still represents a challenge for semiconductor industry [1,2]. These diodes must have very critical characteristics as: fast recovery, low recovery losses, soft behavior, low conduction losses, high voltage capability. Appropriate testing circuits must be developed to ensure cor- rect switching behavior and to evaluate the switching power losses, in cases of high di/dt applied. When currents and voltage in- volved are in the range of kA and kV, respectively, this target may be sometime out of testing capability or, anyway, it could require a big effort in laboratory testing layouts. Scope of this work is to set up a coupled measurement-simula- tion procedure to characterize the reverse recovery behavior of new diode structures with test conditions which approaches those of the real application. Due to high power handled levels, the fol- lowing test is referred to press-pack type diodes. 2. The diode under test The Device Under Test (DUT), on which we focused our study, was a power diode for free-wheeling or clamping application in circuits operating at high voltage and high current; in this case both the steady state behavior and the fast–soft recovery during turn-off are very important issues. We choose to use a press-pack device due to its performances in terms of maximum blocking voltage, current capability and high reliability in high power application; those kinds of devices have low conduction losses and can be cooled more efficiently than modules. The used diode is a 63 mm silicon disk device, diameter needed to ensure the current capability, with a starting intrinsic resistivity higher than 200 X cm and a thickness of silicon higher than 500 lm. Depending on the application, the last two values allow to obtain a blocking voltage ranging between 4.5 and 6 kV. The DUT is a PiN diode, with wide intrinsic region made by pro- ton irradiation to obtain a light phosphorus doping; cathode region is obtained by gallium diffusion while anode region is obtained by phosphorus diffusion. After the diffusion steps, an electron irradiation is performed to reach desired switching performance. The last irradiation step has good repeatability giving a good fabrication process yield (low num- ber of refused pieces, because of the possibility to rework them). 3. Measurement steps We developed a characterization procedure involving three measurement stages, hereinafter described. 0026-2714/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.microrel.2010.07.123 * Corresponding author. Tel.: +39 0521 905818; fax: +39 0521 905822. E-mail address: paolo.cova@unipr.it (P. Cova). Microelectronics Reliability 50 (2010) 1720–1724 Contents lists available at ScienceDirect Microelectronics Reliability journal homepage: www.elsevier.com/locate/microrel