PhD Defense of LEON PEREZ Edgar

 This thesis project deals with the question of heterogeneous integration of interconnected nanowires on microelectronics chips in a view to MEMS and NEMS type devices. These devices aim to address the global problematic of “More than Moore”, that is the transformation of classical CMOS microelectronics processes to enable the development of new integrated micro and nanocomponents.

In particular, over the past few years, a variety of nanomaterial-based devices have arisen, revealing micro-actuators and micro-sensors with new functionalities and/or improved performances, e.g. in terms of resolution, sensitivity, selectivity. Here we will focus on a certain type of nanostructures, Zinc Oxide (ZnO) nanowires, which have mostly been used so far to design devices whose working principle exploits the piezoelectric effect, often judiciously combined with their semiconducting properties. Indeed, when submitted to a mechanical constraint or displacement, piezoelectric nanowires generate an electrical potential (piezopotential). If, in addition to this, nanowires are also semiconducting, the piezopotential can be exploited to control an external current as a function of the mechanical constraint imposed to the nanowire (piezotronic effect). The advantage of using one-dimensional nanostructures lies into the modularity of both their mechanical and piezoelectric properties, in comparison with the bulk material. Moreover, their integration is now possible thanks to growth processes compatible with microelectronic processes (CMOS/MEMS). All these considerations make it possible to design very high performance devices combining the very small dimension of their functional unit elements (hence a high integration density which implies a high spatial resolution) and their sensitivity to nanoscale phenomena.

In this project, we will adopt a very technology-oriented vision of the design of vertically-aligned ZnO-piezoelectric-nanowire matrix-type sensors. Relying on theoretical performance predictions and technological choices to solve device design and fabrication issues, this study aims to produce proof-of-concept prototypes of these high performance devices. First of all, the design process is elaborated based on finite element multiphysics models (FEM) of the piezoelectric response of a single bent nanowire, which we upgraded towards complete pixels, representative of an interconnected nanowire within a matrix. Following these considerations, we have imagined means of characterization of the piezoelectric response of a wire, then of a pixel. The implemented characterization experiment highlighted the complexity of carrying out a systematic, calibrated piezoelectric measurement, decorrelated from the environment of the pixel. Adequate technological solutions could then be implemented through the fabrication of elementary pixels suitable for characterization and whose piezoelectric response could be predictively modeled.

This technological part of the work encompassed several development stages, including the chemical growth of ZnO nanowires and the design of the electrode matrix contacting the nanowires individually. The former splits into two steps: first choosing a clean-room compatible seed layer which will favor growth on a Silicon chip; secondly developing a selective growth process enabling the localization of nanowires within a predefined matrix of electrodes. The second part of the fabrication work focused on defining and optimizing the technological stack with respect to all the above mentioned considerations, and implementing the technological processes yielding the final targeted matrix.

Keywords:
ZnO Nanowires, Piezotronics, Sensors, Multiphysics simulation, 3D integration.

Members of the  jury :
Mireille MOUIS (IMEP-LAHC, Grenoble) : Supervisor
Emmanuelle PAULIAC-VAUJOUR (CEA-Leti, Grenoble) : Co-supervisor
Olivier THOMAS (IM2NP, Marseille) : Rapporteur
Marc LETHIECQ (GREMAN, Tours) :  Examiner
Emmanuel DEFAY (LIST, Luxembourg) : Rapporteur