Systematic characterization for RF small-signal parameter extraction of 28 nm FDSOI MOSFETs up to 110 GHz

Abstract

A systematic RF characterization and small-signal parameter extraction for 28 nm fully depleted silicon on insulator (FDSOI) MOSFETs is presented in this paper. Two-step and hybrid de-embedding methodologies have been applied and compared for the accurate de-embedding of the on-chip RF test pads and access interconnect line elements. To avoid the accumulated error from the de-embedding for the small signal model of the FDSOI MOSFET, an error check criterion is adopted and utilized. Then, the complete extrinsic and intrinsic small-signal parameters of the device were extracted at multi-bias points based on the combination of direct extraction and linear regression approaches. For the extrinsic parameter extraction, the substrate-related parasitic elements and the extrinsic resistance components have been extracted. For the intrinsic small-signal parameter extraction, a complete model is presented and applied for direct extraction of the small-signal parameters. To verify the validity of the proposed modeling approach and the employed small-signal model, the S-parameters of a 28 nm FDSOI NMOS device with a 16 × 1 μm gate width were measured and characterized. From the comparison of the measured and modeled data, it is found that the proposed de-embedding and small-signal model provide an accurate characterization of RF parameter extraction for the FDSOI MOSFETs up to 110 GHz.