Tarek BOUZAR

Abstract: While significant progress has been made in the semiconductor technology and high speed integrated-circuit (IC) fabrication, the measurement infrastructure for characterizing such ICs remains largely undeveloped especially for on-wafer measurements. In particular, commercial probes for applications below 110 GHz use technologies that have been developed and patented about 15 to 20 years ago and are nearly expired. Moreover, these commercial technologies are very expensive since they typically rely on manual assembly of several components. Furthermore, this assembly method leads to large dispersion from product to product. Manual assembly also limits possible geometrical down scaling of the probe—something required to improve their performance. From an electrical performance point-of-view, current commercial probes are not sufficiently down-scaled—thus inducing a high coupling between the substrate and/or adjacent circuits, and the probes, or a high coupling from probe-to-probe. This leads to unreliable measurement results above 60 GHz. Hence, scaling down the probe is required; but becomes increasingly difficult to the point that an innovative probe design is the only option. Finally, measuring broadband up to 500 GHz requires one to measure frequency band per frequency band, inducing the contact issue, increasing man-power cost and making more difficult the analysis of the results. In this ANR project (named PRECISE), we propose a new approach for the design of broadband probes working from DC to at least 220 (and eventually 325) GHz with strongly reduced probe-to-substrate coupling. Electro-magnetic (EM) simulation, mechanical simulation (Year1), integrated balun, tunable loads design (Year 2), and finally physical and electrical characterization (Year2 and 3) will be the core tasks.