The technology of RF communications systems has experienced a phenomenal progress in the last decades. Gallium nitride (GaN) has been identified as the semiconductor that will take over after silicon (Si) to meet the needs of the increasingly market requests. This is explained by the physical parameters of this new material, which highly exceed those of Si. The objective of this project is the development of an enhancement and depletion modes HEMT-GaN based technologies suitable for the manufacturing of MMIC circuits operating at high frequency. This project is divided in 3 research axes: 1) GaN based MMIC fabrication process, including epitaxy, 2) device characterization and modeling, 3) circuit design and characterization. The consortium gathers 3 research laboratories and 1 industrial company covering a very wide range of skills and applied knowledge. The different partners are complementary and expert in the field of GaN activities, which is a major key that will lead to the success of this project. In Sky-GaN project, the development of short gate length GaN-HEMT process will be used to fabricate a prototype Power Amplifier (PA) MMIC circuits to validate these new developed technologies. Indeed, the optimized micro-fabrication process that will be developed will allow the production of new custom PA circuit with higher performance in frequency band covering the E-band [71-76] GHz and [81-86] GHz for both technologies. These fabricated prototypes will also be used as demonstration samples to support the development of new business opportunities for the industrial partner, creating new job opportunities for young researchers and highly qualifier professional. On the other hand, for the academic laboratories participating to Sky-dream project, the research and development work will produce important scientific impact and economic value, which is in complete agreement with the mission of CNRS.

Ce projet propose le développement d’un TBH (Transistor Bipolaire à Hétérojonction) sur phosphure d’indium et à base antimoniée, composant analogique très haute fréquence (vers le THz), selon 2 axes complémentaires : le premier concerne l’amélioration significative des performances du TBH développé dans le projet RNRT MELBA, pour des applications hyperfréquence (Fmax >0,4 THz), grâce à l’optimisation de la structure et une réduction progressive des dimensions (700-500 nm) ; cette adaptation sera validée par la réalisation de circuits pertinents à l’état de l’art (source opérant au delà des 0,1 THz, avec un niveau de puissance significatif, 100-200 mW) ; Le deuxième axe regroupe d'une part la conception d'hétérostructures novatrices permettant la mise en oeuvre d'un transport électronique ultra-rapide dans l'ensemble du transistor et d'autre part la nanofabrication de ces transistors (WE=300 nm, J > 1 MA/cm2) pour établir un nouvel état de l'art (0,5THz). L’utilisation de l’antimoine dans la base, dont la faisabilité et l’intérêt ont été montrés dans le projet RNRT/MELBA, permettra de bénéficier d’une part des forts dopages possibles qui permettent une optimisation verticale de la structure, d’autre part des caractéristiques physiques propres permettant une technologie appropriée aux dimensions submicroniques.

The OCELOT project aims at developing the optical linear sampling of very high rate (> 100Gbps) optical signals with advanced modulation formats, especially in phase (x-PSK), in order to extract amplitude and phase information and display it in a constellation diagram. The main purpose of OCELOT is the transfer of this technology from a research industrial laboratory to a SME by realizing the prototype of a future product, which is planned to be launched on the very competitive constellation oscilloscope market. The two major obstacles identified in the project refer on one hand to the pulse laser source used for linear sampling, on the other hand to the acquisition and processing of the sampled and digitized signals, with display refreshment rate (> 1Hz) matching end users requirements about real time measurement and analysis. The laser source must provide short ( -3dBm) but also low pulse rate (40MHz-300MHz), in order to fit the economical requirement of simple implementation and reduced cost. This source, containing an amplifying quantum dot chip, will be developed in a specific task, with close collaboration of several partners of OCELOT project. The low economical cost is a hard requirement and will be considered as important as the pure technical purposes. After validation, the laser will be transferred to the SME APEX for its integration inside the constellation oscilloscope prototype. The digital acquisition board (CAN) will be designed with digital signal processing integrated circuits (DSP, FPGA) enabling the end display of the constellation diagram. A robust and fast algorithm must be designed, developed and implemented in FPGA, and transferred to APEX with the CAN for its integration in the prototype. Then, the prototype, integrating the laser pulse sources, the linear sampling modules, and the digital acquisition and processing boards, will be realized by APEX and validated with the very high rate and diverse format modulation sources developed by partner UR1-ENSSAT. The project will be promoted by presenting the prototype in important scientific and technical international conference exhibitions, in order to probe the market and attract future customers.

Le projet vise à réaliser des microsystèmes optiques pour l’analyse spectroscopique dans le proche infrarouge [entre 1 et 2.5 µm] dans le système de matériaux InP / InGaAs. Il réunit des compétences en conception de système MOEMS, épitaxie, microfabrication et validation des applications et associe deux laboratoires publics, le LEOM à l’Ecole Centrale de Lyon (coordonnateur) et le CEMAGREF, et deux industriels, ALCATEL-THALES III-V LAB et la PME DATALINK Il s’agit d’associer, dans un microspectromètre intégré, une photodiode résonnante à puits quantiques d’InGaAs contraints avec une fonction “ accordabilité en longueur d’onde ”. On cherchera aussi à intégrer ce composant en matrices linéaires afin de constituer des imageurs hyperspectraux. On utilisera les matériaux de la filière InP / InGaAs épitaxiés par MOCVD et les techniques de micro-usinage développées au LEOM pour réaliser des micro-cavités Fabry-Pérot accordables électriquement basées sur une alternance de couches d’air et d’InP. Ce microsystème, grâce à son faible encombrement et son coût réduit, ouvrira la voie à des applications industrielles extrêmement porteuses comme : - les analyses biologiques non invasives (monitoring du diabète en particulier); - les analyses industrielles en ligne (agro-alimentaire, pétrochimie, chimie,…); - l'imagerie hyperspectrale (utilisable en agriculture, dans l'environnement,... etc) Des prototypes de systèmes d’analyse spectroscopique exploitant ces microsystèmes seront conçus et réalisés pour des applications environnementales concernant des produits agroalimentaires.

ULTIMATE is a three-year proposal for an ambitious research project focused on upper limit technology investigations mandatory to attain THz electronics. InP-based transistors are the world’s fastest three-terminal semiconductor devices. Regardless of whether bipolar or field-effect transistors are considered, current gain cutoff frequencies tend to stagnate in the range fT = 500-600 GHz (0.5-0.6 THz). Progress in device design so far relied on semi-qualitative electronic energy band structures interpolated from often incompletely known materials. While standard simulators can predict cutoff frequencies exceeding 1 THz, experiments fall far short from this goal, presumably because strong electric fields within nano-devices drive electrons to higher energy states (L, X valleys) not accounted for in traditional TCAD tools or even Monte Carlo simulators which neglect quantum mechanical transmission effects at barriers. In the ULTIMATE project, for the first time, we will tackle the design of ultrahigh-speed transistors on a fundamental atomistic level, with Density-Functional Theory (DFT) for accurate band structure calculations and quantum transport simulations to break through the present bandwidth bottleneck and finally experimentally achieve 750-1000 GHz cutoff frequencies. The partners of ULTIMATE project are two academic laboratories (IMS-Bordeaux and ETHZ), one industrial laboratory (Alcatel Thales III-V Lab) and a start-up (Xmod technologies). Each of them brings some specific contributions based on an excellent expertise and a high-level research in the fields of InP HBT technology, electrical characterization, Parameter extraction methodology, compact modeling and multi-scale simulation. Such gathering of complementary skills is particularly relevant to provide guideline for technology achievement in order to perform state of the art technology and measurement up to THz. The way around the present fT cutoff frequency bottleneck requires securing and exploiting an intimate knowledge of the electronic structure of the involved materials. To our knowledge, no others have attempted to implement transistors from an atomistic level starting point. The work is unique because it synergistically builds on the leading device processing capabilities of III-V Lab and ETHZ mm-wave device fabrication together with the IMS-Bordeaux up-to date 500GHz S parameters measurements equipment and on the pioneering materials/device simulation expertise the ETHZ team. The entire project chain, from physical modelling to epitaxial growth, device fabrication and characterization is thus entirely within the France-Switzerland area. Density-Functional Theory and quantum transport computations will be performed by ETHZ (Prof. Luisier team) for InP/InGaAs and InP/GaAsSb HBTs. Their fabrication will be identically realized using III-V Lab and ETHZ (Prof. Bolognesi team) fabrication processes respectively. Subsequently, their electrical characteristics will be verified against experimental data measured by IMS and extracted by XMOD. In particular, the band structure of strained heavily carbon-doped GaAsSb and InGaAs layers will be calculated to determine if higher lying valleys (L-minima) become populated and how they affect device dynamic performances. The transistor design will then be adjusted to optimize the injection efficiency of electrons from the GaAsSb or InGaAs base layer into the InP collector and enhance electron transport through both the base and the collector layers to reach cutoff frequencies in the range of 750-1000 GHz. Atomistic simulations will provide new insights on the inner workings of ultrahigh-speed heterostructure transistors and enable the design of a new generation of THz devices. Success is expected in extending the life of transistor electronics to higher bandwidths, a critical outcome because no credible alternative to transistor technology has so far been developed.