edfas.org ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 27 NO. 1 12 Divacancies are reported to have ∆E of 0.42 eV and a sn in the 10-15 cm2 range for the single negatively charged acceptor state (0/-) V2(0/-) trap; [25,27] ∆E of 0.23 eV and a s n in the 10-15 to 10-16 cm2 range for V 2(0/+) trap, [25] and ∆E of 0.23 eV and a sn in the 10-15 to 10-16 cm2 range[25,26] for V 2(-/--) trap. Consequently, T4, T5 and T6 corresponds to V2(0/-), V2(0/+) and V2(-/--) traps, respectively. An activation energy of 0.17 eV may be linked both to traps related to the oxygen-vacancy complex (V-O) or to the interstitial–carbon–substitutional carbon complex CiCs.[25,28] The capture cross section s n is reported to be in the range of 10-14 cm2 for V-0,[28] while no value of s n for the CiCs complex is reported in the literature to our knowledge. As the CiCs complex is rather unlikely, and considering the estimated capture cross section of the trap, it is more reasonable to consider that T7 may be attributed to the oxygen-vacancy (V-O) pair. The nature of the T8 trap, characterized by a ∆E of 0.53 eV and a sn of around 10-14-10-15 cm-2 cannot be specified, due to lack of availability of bibliographical data. The main identified traps may be linked to divacancies (V2(0/-), V2(0/+) and V2(-/--)), to hydrogen (V2H and/or VOH), to phosphorus-vacancy complex (V-P), and the oxygenvacancy complex (V-O). The presence of divacancies could Fig. 6 Temperature evolution of the Sv,GR(fa,T)·fa , Sv,GR(fa,T) representing the GR noise level estimated from the experimental spectra at an imposed frequency fa of 60 Hz for the V2H trap identified in Fig. 5. fa is chosen to respect the conditions 2πfaτ0(Tmin) = fa/f0(Tmin) >>1 and 2πfaτ0(Tmax) = fa/f0(Tmax) << 1, f0(Tmin) and f0(Tmax) being the characteristic frequencies GR noise at the temperature limits (Tmin and Tmax) where this trap is active. From the maximum of the bellshaped behavior of SGR(f0, T)·fa the volume trap density may be estimated using Eq 6. The experimental B coefficient, expressed as Bexp = is consequently estimated at about 0.6, around a factor of 2 higher than the expected theoretical value evaluated for conventional planar transistors with one gate. Table 1 Different identified trap parameters Trap Activation energy, ∆E, eV Capture cross section, σ, cm2 Temperature range T, K Devices T1 0.45 3-6·10-17; 0.35 – 3.1 10-17 5·10-17; 5·10-17 0.2·10-17 1.6·10-17 250-300 240-280; 200-280 260-300 300-330 Standard UTBOX,[18,19] Rotated UTBOX [18;17] Si/SiGe superlattice I/O n-channel FinFET[15] GAA NW FET[20] T2 0.32 0.6·10-15 0.3·10-15; 0.4·10-15 1.3·10-15 210-250 210-270; 250-300 220-260 Standard UTBOX[17] Rotated UTBOX[17,18] Si/SiGe superlattice I/O n-channel FinFET[15] T3 0.44 1.3·10-14 - 3·10-14; 1·10-15; 5·10-14; 4·10-14 - 2·10-15; 2·10-15 0.76·10-14 210-300; 220-290 200-280; 300-310 240-260 Standard UTBOX[16-18] Rotated UTBOX[18,19] Si/SiGe superlattice I/O n-channel FinFET[15] T4 0.42 1.2·10-15; 4·10-15; 0.3-2.5·10-15 0.3-1·10-15; 0.4·10-15 3.8·10-15 1.4·10-15 0.67·10-15 200-250; 200-230 240-280; 285-295 220-260 300-325 300-340 Standard UTBOX[17-19] Rotated UTBOX[17,19] Si/SiGe superlattice I/O n-channel FinFET[15] GAA NS FET[21] GAA NW FET[20] T5 0.23 0.7·10-16 0.9·10-16 90-200 200-270 Standard UTBOX[16] Rotated UTBOX[17] T6 0.2 1·10-15 80-150 Standard UTBOX[16] T7 0.17 2·10-14 77-110 Standard UTBOX[19] T8 0.53 1.5·10-14 2.1·10-15 220-260 270-320 UTBOX[16] Si/SiGe superlattice I/O n-channel FinFET[15]
RkJQdWJsaXNoZXIy MTYyMzk3NQ==