ED Sciences Physiques et de l'Ingénieur
Contribution to the design of miniature high frequency probes for precision microwave measurements on silicon wafers
by Tarek BOUZAR (Laboratoire de l'Intégration du Matériau au Système)
The defense will take place at 10h30 - Amphi J.P.DOM A0.85 351 Cr de la Libération, bâtiment A31, 33400, Talence
in front of the jury composed of
- Sébastien FREGONESE - Chargé de recherche - Université de Bordeaux - Directeur de these
- Thierry PARRA - Professeur - LAAS-CNRS - Rapporteur
- KAMEL HADDADI - Professeur - Université de Lille - Rapporteur
- Jean-Daniel ARNOULD - Maître de conférences - Grenoble INP - CoDirecteur de these
- Thomas ZIMMER - Professeur - Université de Bordeaux - Examinateur
- Joao AZEVEDO GONCALVES - Docteur - STMicroelectronics - Examinateur
Significant progress has been made in semiconductor technology and high-speed integrated circuit (IC) fabrication, yet the measurement infrastructure for characterizing these ICs remains largely underdeveloped, especially for on-wafer measurements. In particular, commercial probes for applica- tions below 110 GHz use technologies that were developed and patented roughly 15 to 20 years ago and are now almost obsolete. Furthermore, these commercial technologies are very costly, mainly because they rely on the manual assembly of several components. In addition, this assembly method causes large variations from one product to another. Manual assembly also limits the possibility of reducing the geo- metric scale of the probe, which is necessary to improve its performance. From an electrical performance standpoint, current commercial probes are not sufficiently miniaturized, causing strong coupling either between the substrate and/or adjacent circuits and the probes, or between probes themselves. This leads to unreliable measurement results above 60 GHz. Consequently, there is a need to reduce the probe size, although this becomes increasingly challenging, to the point where a truly innovative probe design is the only viable solution. Lastly, measuring in a wide frequency band up to 500 GHz requires measuring band by band, which raises issues with contact, increases labor costs, and complicates result analysis. In this ANR project (called PRECISE), we propose a new approach to designing wideband probes operating from DC to at least 110 GHz (a possibly up to 220 and 325 GHz) with significantly reduced probe-to-substrate coupling.