Cosmic Ray Withstand Capability of RB-IGBT Utilizing different gate conditions Daniel Hofmann ISPS 2016, August 31 st September 2 nd
Topics - Overview 1. Motivation 2. IGBT Device: NPT vs. RB-IGBT 3. Effect of Cosmic Radiation 4. Experimental Setup 5. Measurement Results 6. Summary 2
Motivation Cosmic ray triggers a breakdown in power semiconductors High voltage applications are critical RB-IGBT device robustness to cosmic ray Determination of failure in time (FIT) Impact of gate conditions of RB-IGBT to FIT rate http://www.kn-online.de/ http://www.manager-magazin.de/ 3
Topics - Overview 1. Motivation 2. IGBT Device: NPT vs. RB-IGBT 3. Effect of Cosmic Radiation 4. Experimental Setup 5. Measurement Results 6. Summary 4
IGBT device: NPT vs. RB-IGBT Theoretically, IGBT structure has reverse blocking capability N + Impurity conc. Gate Emitter N - drift P P + collector Collector A A P-type doping P-type doping N-type doping A A Forward Biased Otsuki. M., Presentation: The Power Electronics South America, 11 th 13 th of September 2012, Sao Paulo Reverse Biased 5
diced side-wall IGBT device: NPT vs. RB-IGBT Conventional IGBT and RB-IGBT Conventional active area termination RB-IGBT active area termination Otsuki. M., Presentation: The Power Electronics South America, 11 th 13 th of September 2012, Sao Paulo Side-wall protection 6
diced side-wall IGBT device: NPT vs. RB-IGBT Blocking capability Emitter = GND Forward Blocking State + Cracks active area Depletion region extension termination Collector = positive bias _ Reverse Blocking State _ Emitter = + active area termination Collector = negative bias Leakage current generation + Destruction of device Otsuki. M., Presentation: The Power Electronics South America, 11 th 13 th of September 2012, Sao Paulo 7
IGBT device: NPT vs. RB-IGBT Blocking capability 1. Electrical deactivated side-wall Stop leakage 2. Junction termination +/ bias capable design Emitter = GND active area termination Collector = positive bias Otsuki. M., Presentation: The Power Electronics South America, 11 th 13 th of September 2012, Sao Paulo 8
IGBT device: NPT vs. RB-IGBT Conventional IGBT Uni-directional Leakage current NPT or FS structure Cross-section of chip when reverse voltage is applied: RB-IGBT Reverse blocking capability NPT structure Junction isolation region Bi-directional switch (2x RB) M. Takei et al., Fuji Electric Journal, Vol. 75, No. 8, 2002 9
IGBT device: NPT vs. RB-IGBT RB-IGBT has leakage current in reverse direction Reduce I ces by positive gate voltage 10
Topics - Overview 1. Motivation 2. IGBT Device: NPT vs. RB-IGBT 3. Effect of Cosmic Radiation 4. Experimental Setup 5. Measurement Results 6. Summary 11
Effect of Cosmic Radiation Primary radiation spallation with atmosphere Secondary radiation generated Neutrons interact with nuclei of the device Ziegler, J.F.: Terrestrial Cosmic Ray Intensities. In: IBM Journal of Research and Development 42 (1998), Nr. 1, S. 117 139 http:\\http://quantumcomputers.ch/qcblog/wp-content/uploads/2012/06/stdmodel2.jpg 12
Effect of Cosmic Radiation Electron-hole plasma induced Neutron Gate electrode Emitter electrode Gate oxide n + Filamentary current p p n - e - p+ Short Circuit n + plasma p + 13
Effect of Cosmic Radiation Collector-Emitter Filamentation example FWD - Streamer Multi-avalanche effect Weiss et. al., Techn. University Munich, Infineon 14
Topics - Overview 1. Motivation 2. IGBT Device: NPT vs. RB-IGBT 3. Effect of Cosmic Radiation 4. Experimental Setup 5. Measurement Results 6. Summary 15
Experimental Setup Set a target for the time to failure Let s assume 15years Within 15years only 1 system out of 100 systems should fail (1%) System is running for 12h per day 1FIT n n tot f t 1system 100 systems 12 h/d 365 d/y 15 y 1 6.57 10 h 152 1 10 h 6 9 152 FIT 16
Experimental Setup Assume 1 system contains 6 IGBT modules Half bridge module in 2-Level operation Hence, FIT-rate per power module 152 FIT 25,3 FIT 6 power modules module 17
Experimental Setup Neutron beam source Parallel connection of devices (RB-IGBT) source forward reverse 18
Experimental Setup RB IGBT Gate conditions IGBT with parasitic thyristor PNP transistor FWD mode 19
Experimental Setup RB IGBT Gate conditions Emitter Gate _ IGBT with parasitic thyristor p+ n+ n+ p+ p p n- Depletion region p+ + Collector 20
Experimental Setup RB IGBT Gate conditions Emitter Gate + PNP transistor p+ n+ n+ p+ p p n- Depletion region p+ _ Collector 21
Experimental Setup RB IGBT Gate conditions Emitter Gate VGE=+15V Channel + FWD mode p+ n+ n+ p+ p p n- Depletion region p+ _ Collector 22
Topics - Overview 1. Motivation 2. IGBT Device: NPT vs. RB-IGBT 3. Effect of Cosmic Radiation 4. Experimental Setup 5. Measurement Results 6. Summary 23
FIT rate per chip, 100% duty cycle Measurement Results FIT curves 50A/1200V (c) FWD mode (a) parasitic thyristor Usual appearance FWD: (b) PNP transistor not observed 24
Measurement Results Threshold of breakdown (a) parasitic thyristor (b) PNP transistor (c) FWD mode 25
Measurement Results Impact of E av(max) Left = foward direction ; Right = reverse biased Higher FIT in forward direction 26
Measurement Results Vge = +15V Higher threshold confirmed by several measurements No breakdown measured Vge = 0V Lower threshold Above threshold almost same FIT rate Forward Threshold at 800V (66%) 27
Topics - Overview 1. Motivation 2. IGBT Device: NPT vs. RB-IGBT 3. Effect of Cosmic Radiation 4. Experimental Setup 5. Measurement Results 6. Summary 28
Summary FIT rate of RB-IGBT under different gate conditions characterized Forward bias - Latch-up of parasitic thyristor (NPNP) Reverse bias FIT FWD mode and PNP transistor reveal similar FIT rate FWD mode has higher threshold 29