Physical basis for nondestructive tests of MOS radiation hardness

Abstract

measurements have been extended to include temperatures in the range 80-300K. These results are used to expand upon a trapping model of the noise. The evidence suggests that the noise and the oxide-trap charge are both influenced by the preirradiation density of oxygen vacancies in the SiO 2 [2]. We have found that the 1/f noise and channel resistance of unirradiated nMOS transistors from a single lot with various gate-oxide splits closely correlate with the oxide-trap and interface trap charge, respectively, following irradiation. The 1/f noise is explained by a trapping model, while the variations in channel resistance are explained by scattering from interface-trap precursor defects. It appears that both noise and channel mobility measurements may be useful in defining nondestructive hardness assurance test methods for devices fabricated from a single technology. It may be difficult to use either for making cross-technology comparisons. Finally, during the course of this study it was found that process techniques that improve the radiation hardness of MOS devices at room temperature can greatly reduce the 1/f noise of MOS devices at cryogenic temperatures. * We also demonstrate a striking correlation between the preirradiation channel resistance of MOS transistors and their postirradiation interface-trap charge [3]. We have developed a model that relates the preirradiation 1/f-noise to the net radiation-induced oxide charge-trapping efficiency [2], and a model that relates the preirradiation channel mobility to radiation-induced interface trap generation efficiency [3]. We conclude that, even before a device is irradiated, carrier-defect interactions evidently reveal a great deal of information about the radiation hardness of MOS devices. Implications for hardness assurance testing are discussed.

Effects of aging on the 1∕f noise of metal-oxide-semiconductor field effect transistors

“…[8][9][10][11][12][13] Here, we estimate 1 / 0 Ϸ 10 12 to be consistent with previous work. For the device that received the N 2 post-gate-oxidation anneal ͑sample C, t ox =48 nm͒, we find N bt Ϸ 1.5ϫ 10 11 cm −2 in 1989 and N bt Ϸ 4.5ϫ 10 10 cm −2 in 2007.…”

Empirical relationship between the low-frequency noise spectral density and the transconductance of silicon-on-insulator n-channel metal-oxide-semiconductor transistors

“…This indicates that the relationship reported here is not restricted to one SOI CMOS technology or one class of devices and may be more general. However, for bulk CMOS devices, no such relationship has been observed, [15][16][17] so that this point needs further verification. Finally, it should be reported that not such a clear trend is observed, when S VG is plotted vs the corresponding threshold voltage ͑Fig.…”

Low frequency noise in GaN metal semiconductor and metal oxide semiconductor field effect transistors

“…9 ͑the 1/f noise originated from the surface is often dominant in Si MOSFETs͒. 10 For the noise sources located at the surface of regions 1 and 2 in Fig. 3, the relative spectral noise density of the short circuit drain current fluctuations can be presented in the following form:…”