A physics-based silicon-germanium (SiGe) heterojunction bipolar transistor (HBT) aging model for mixed-mode stress based on the lucky-electron model and reaction-diffusion theory is developed for integration with compact models. An effective aging parameter extraction method is described, and the aging model parameters are fit for a modern SiGe HBT platform. The aging model is implemented as a wrapper in the Cadence Spectre circuit simulator. Device-level aging simulations are shown to be well-matched to measured degradation data. The aging model is further used to explore the effects of aging on a simple current mirror circuit, showing a decrease in mirror ratio with degradation.

No abstract

No abstract

No abstract

An empirical reliability model is proposed here that is able to predict parameter degradation for a SiGe Hetero-junction Bipolar Transistor (HBT) by scaling stress time laterally producing a universal curve that describes whole time evolution of degradation. The predictability of the degradation pattern is demonstrated in experiments at a forward active mode as well as the reverse VEB stress accounting for bias and current dependence of degradation. Furthermore, our model and methodology enables us to do an aging simulation at the circuit level.Index Terms -SiGe HBT, reliability model, aging simulation, safe operating area, hot carrier injection. + +0641&7%6+10Decades of reliability studies have tried to improve the stability of SiGe HBTs with respect to hot carrier injection [1], self-heating [2], electromigration [3], noise degradation [4], and an increase of mis-matching [5] due to the Bipolar / BiCMOS circuit being advanced at a bias much higher than BV CEO in high current and temperature.The first priority in implementing SiGe HBT reliability is to establish the precise safe operating area (SOA) which consists of the SOA, degradation, and catastrophic regions. This SOA provides the designer confidence on the stability of HBTs by being able to show the potential reliability risk at the use condition during circuit stimulation. Fig. 1. Typical safe operating area of a HBT consisted of SOA, degradation, and catastrophic failureHowever, the SOA shown in Fig. 1 assumes the worst case scenario in which a HBT is stressed at the high current and bias while being monitored at the low current and bias. The HBT with small bias and current swing between stress and monitoring appears functional even at the degradation region without any significant parameter degradations, in actuality the HBT is considerably damaged. Some papers [6] assess circuit level degradation using their own models but lack adequate explanation on methodology and its implementation. %QPUGSWGPVN[ KV KU GUUGPVKCN VQ DG CDNG VQ RTGFKEV VJG CEVWCN RCTCOGVGT FTKHV CV VJG URGEKHKE EQPFKVKQPU DCUGF QP FGXKEG TGNKCDKNKV[ RJ[UKEU HQT C FGUKIPGT VQ OCZKOK\G EKTEWKV RGTHQTOCPEG ++ 4'.+#$+.+6; /1&'.5 A. Forward Active Mode #P GORKTKECN TGNKCDKNKV[ OQFGN =? KU RTQRQUGF JGTG VQ FGVGTOKPG VJG FGITCFCVKQP QH *$6U CV C HQTYCTF CEVKXG OQFG YJKEJ JCU OKZGFOQFG UVTGUU EQODKPGF YKVJ EWTTGPV CPF DKCU GZRTGUUGF KP 'S [ ] ( ) n CB g E t V B J LN A Beta ⋅ ⋅ ⋅ ⋅ = Δ exp ) ( (%) YJGTG B I CPF P CTG DKCU EWTTGPV CPF VKOG HCEVQTU +P QTFGT VQ GZVTCEV VJG DKCU HCEVQT VJG EQPXGPVKQPCN CEEGNGTCVGF UVTGUU VGUVU CTG RGTHQTOGF CV C HKZGF EWTTGPV FGPUKV[ YKVJ XCTKQWU V CB XCNWGU HQNNQYGF D[ EWTXG HKVVKPI WUKPI QWT OQFGN CU UJQYP KP (KI %QORCTGF VQ C VTCFKVKQPCN CRRTQCEJ VQ GZVTCEV B HTQO C RNQV QH VKOGVQHCKN 66( CPF DKCU GZVTCRQNCVKPI 66( VQ Beta FGITCFCVKQP YKVJ FKHHGTGPV P (KI ENGCTN[ UJQYU FGITCFCVKQP VTGPFU YKVJ VJG UCOG P KP CP GPVKTG DKCU TGIKQP HTQO VJG UVTGUU CPF VJG WUG EQPFKVKQP+P IGPGTCN a high voltage (HV) HBT with low collector doping has a bias factor that is less than 2...