Future Trends in Microelectronics: Up the Nano Creek

Origin of 1/f Noise in MOS Devices: Concluding a Noisy Debate

K. Akarvardar, S. Cristoloveanu, and P. Gentil

Institut de Microélectronique, Electromagnétisme et Photonique (IMEP)Minatec, 3 Parvis Louis Néel, BP257, 38016 Grenoble Cedex 1, France

1.   Introduction

The importance of 1/f noise in semiconductors is well documented, enthusiastically explored, and definitely feared. There are two distinct camps in the 1/f world. For McWhorter’s partisans, there is no doubt that the noise originates from fluctuations in the carrier number.¹ In MOS devices, for example, carrier trapping is related to the presence of slow traps in the gate oxide. Data on MOSFETs collected for years are rather convincing in this respect, except for Hooge and his disciples. For them, there is even less doubt that the 1/f noise in semiconductor devices proceeds from carrier mobility fluctuations Δμ.²

The aim of our paper is to show that the SOI four-gate transistor (G⁴-FET) is a competent referee for this rivalry. The G⁴-FET has four independent gates, which offer tremendous versatility of operation. In the G⁴-FET, the cross-sectional size and position of the conducting channel can be adjusted via the appropriate biasing of the four gates. This feature enables an in-situ comparison of the interface and the volume conduction properties.³ By comparing the bulk and the surface noise characteristics to the McWhorter and Hooge noise models, the domains of validity of these models can be determined. Also, the comparison ...