RF noise shielding method and modelling for nanoscale MOSFET

Jyh-Chyurn Guo, Yi Min Lin, Yi Hsiu Tsai

    Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

    1 Scopus citations

    Abstract

    RF noise shielding methods with different coverage areas (Pad and TML shielding) were implemented in two port test structures adopting 100-nm MOSFETs. Noise measurement reveals an effective suppression of NFmin but increase of NF50, simultaneously from the shielding methods. The suppression of NFmin is contributed from the reduction of Re(Yopt) while the noise resistance Rn is kept nearly the same. A lossy substrate model developed in our original work for a standard structure without shielding can be easily extended based on the layout and topology of the shielding schemes to predict the noise shielding effect and explain the mechanisms. The extended lossy substrate model indicates that the elimination of substrate loss represented by substrate RLC networks is the major mechanism contributing the reduction of NFmin. However, the increase of parasitic capacitance generated from the shielding structures is responsible for the degradation of fT and NF50. The results provide an important insight and guideline for low noise RF circuit design.

    Original languageEnglish
    Title of host publication2008 European Microwave Integrated Circuit Conference, EuMIC 2008
    Pages390-393
    Number of pages4
    DOIs
    StatePublished - 1 Dec 2008
    Event2008 European Microwave Integrated Circuit Conference, EuMIC 2008 - Amsterdam, Netherlands
    Duration: 27 Oct 200831 Oct 2008

    Publication series

    Name2008 European Microwave Integrated Circuit Conference, EuMIC 2008

    Conference

    Conference2008 European Microwave Integrated Circuit Conference, EuMIC 2008
    Country/TerritoryNetherlands
    CityAmsterdam
    Period27/10/0831/10/08

    Fingerprint

    Dive into the research topics of 'RF noise shielding method and modelling for nanoscale MOSFET'. Together they form a unique fingerprint.

    Cite this