Theory and experiment on the 1/f noise in p-channel metal-oxide- semiconductor field-effect transistors at low drain bias

Charles Surya, Thomas Y. Hsiang

Research output: Contribution to journalArticle

76 Citations (Scopus)

Abstract

Voltage noises with power spectra of the form 1/f were studied in a series of custom and commercial p-channel metal-oxide-semiconductor field-effect transistors. Detailed measurements of the time-correlation functions indicated that the noise originated from a stationary and Gaussian source. The spatial correlation function, measured in devices made with extra voltage probes placed in the conduction channels, showed no measurable amount of correlation down to a distance of 7 m, excluding the possibility of a diffusion mechanism for the noise. The results, combined with the experimental data on the dependence of the noise power spectra on the bias conditions, led us to establish a simple model based on a variation of the McWhorter model to account for the noise. Built into the model was an energy dependence of the trap concentration which in turn yielded a spatial dependence in the presence of a gate bias. This model explained quantitatively the experimentally observed change in the exponent of the noise spectrum as the gate bias was varied. It was then meaningful to compare the experimental and computed noise powers at a single, fixed frequency.

Original languageEnglish
Pages (from-to)4898-4905
Number of pages8
JournalPhysical Review B
Volume33
Issue number7
DOIs
Publication statusPublished - Jan 1 1986

Fingerprint

MOSFET devices
metal oxide semiconductors
field effect transistors
Power spectrum
noise spectra
Experiments
power spectra
Electric potential
electric potential
traps
exponents
conduction
probes
energy

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics

Cite this

Theory and experiment on the 1/f noise in p-channel metal-oxide- semiconductor field-effect transistors at low drain bias. / Surya, Charles; Hsiang, Thomas Y.

In: Physical Review B, Vol. 33, No. 7, 01.01.1986, p. 4898-4905.

Research output: Contribution to journalArticle

@article{54d5b69adc694fb0bb89a7f30cdb477e,
title = "Theory and experiment on the 1/f noise in p-channel metal-oxide- semiconductor field-effect transistors at low drain bias",
abstract = "Voltage noises with power spectra of the form 1/f were studied in a series of custom and commercial p-channel metal-oxide-semiconductor field-effect transistors. Detailed measurements of the time-correlation functions indicated that the noise originated from a stationary and Gaussian source. The spatial correlation function, measured in devices made with extra voltage probes placed in the conduction channels, showed no measurable amount of correlation down to a distance of 7 m, excluding the possibility of a diffusion mechanism for the noise. The results, combined with the experimental data on the dependence of the noise power spectra on the bias conditions, led us to establish a simple model based on a variation of the McWhorter model to account for the noise. Built into the model was an energy dependence of the trap concentration which in turn yielded a spatial dependence in the presence of a gate bias. This model explained quantitatively the experimentally observed change in the exponent of the noise spectrum as the gate bias was varied. It was then meaningful to compare the experimental and computed noise powers at a single, fixed frequency.",
author = "Charles Surya and Hsiang, {Thomas Y.}",
year = "1986",
month = "1",
day = "1",
doi = "10.1103/PhysRevB.33.4898",
language = "English",
volume = "33",
pages = "4898--4905",
journal = "Physical Review B",
issn = "1098-0121",
publisher = "American Physical Society",
number = "7",

}

TY - JOUR

T1 - Theory and experiment on the 1/f noise in p-channel metal-oxide- semiconductor field-effect transistors at low drain bias

AU - Surya, Charles

AU - Hsiang, Thomas Y.

PY - 1986/1/1

Y1 - 1986/1/1

N2 - Voltage noises with power spectra of the form 1/f were studied in a series of custom and commercial p-channel metal-oxide-semiconductor field-effect transistors. Detailed measurements of the time-correlation functions indicated that the noise originated from a stationary and Gaussian source. The spatial correlation function, measured in devices made with extra voltage probes placed in the conduction channels, showed no measurable amount of correlation down to a distance of 7 m, excluding the possibility of a diffusion mechanism for the noise. The results, combined with the experimental data on the dependence of the noise power spectra on the bias conditions, led us to establish a simple model based on a variation of the McWhorter model to account for the noise. Built into the model was an energy dependence of the trap concentration which in turn yielded a spatial dependence in the presence of a gate bias. This model explained quantitatively the experimentally observed change in the exponent of the noise spectrum as the gate bias was varied. It was then meaningful to compare the experimental and computed noise powers at a single, fixed frequency.

AB - Voltage noises with power spectra of the form 1/f were studied in a series of custom and commercial p-channel metal-oxide-semiconductor field-effect transistors. Detailed measurements of the time-correlation functions indicated that the noise originated from a stationary and Gaussian source. The spatial correlation function, measured in devices made with extra voltage probes placed in the conduction channels, showed no measurable amount of correlation down to a distance of 7 m, excluding the possibility of a diffusion mechanism for the noise. The results, combined with the experimental data on the dependence of the noise power spectra on the bias conditions, led us to establish a simple model based on a variation of the McWhorter model to account for the noise. Built into the model was an energy dependence of the trap concentration which in turn yielded a spatial dependence in the presence of a gate bias. This model explained quantitatively the experimentally observed change in the exponent of the noise spectrum as the gate bias was varied. It was then meaningful to compare the experimental and computed noise powers at a single, fixed frequency.

UR - http://www.scopus.com/inward/record.url?scp=0000299051&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0000299051&partnerID=8YFLogxK

U2 - 10.1103/PhysRevB.33.4898

DO - 10.1103/PhysRevB.33.4898

M3 - Article

AN - SCOPUS:0000299051

VL - 33

SP - 4898

EP - 4905

JO - Physical Review B

JF - Physical Review B

SN - 1098-0121

IS - 7

ER -