Jury:

• Emmanuel DUBOIS, Director of Research, University of Lille (Thesis supervisor)
• Thomas SKOTNICKI, Professor, Cezamat Warsaw (Thesis co-supervisor)
• Jean-François ROBILLARD, Senior Lecturer, IEMN Lille & ISEN Lille (Examiner)
• Stephane MONFRAY, Research Engineer, STMicroelectronics Crolles (Examiner)
• Olivier BOURGEOIS, Director of Research, Institut Néel Grenoble (Rapporteur)
• Sylvie HEBERT, Research Director, CRISMAT Caen (Rapporteur)
• Edith KUSSENER, Senior Lecturer, IM2NP & ISEN Toulon (Examiner)
• Guillaume SAVELLI, Research Engineer, CEA Grenoble (Examiner)

Summary:

The rise of the Internet of Things (IoT) and autonomous, communicating sensors appears to be delayed due to the lack of a reliable, secure and low-cost energy source [Nordrum 2016]. Thermoelectric energy harvesters offer these key advantages. Silicon has the advantages of being highly abundant, less environmentally polluting and having the technological facilities and processes to enable mass production of thermoelectric energy harvesters at low cost compared to conventional materials (bismuth tellurium alloys). However, silicon is a poor thermoelectric material due to its high thermal conductivity (150Wm^(-1) K^(-1)) [Haras et al. 2015]. The ability to reduce thermal conductivity while preserving electrical conductivity and Seebeck coefficient is the key to improving silicon as an efficient thermoelectric material. To this end, efforts are directed towards the phononic part of heat transport, which is the dominant contribution in semiconductors [Jin 2014].

The research carried out during this thesis focused on the integration of nanostructured silicon membranes of phononic arrays [Haras 2016; Lacatena et al. 2014; Tang et al. 2010; Yu et al. 2010] in thermoelectric energy recovery demonstrators and their characterisation with respect to the state of the art. The results of these studies demonstrated the feasibility of a silicon-based thermoelectric energy recuperator with sufficient performance to supply energy to autonomous sensor nodes [Vullers et al. 2009] and performance comparable to that of a state-of-the-art recuperator based on bismuth tellurium [Bottner 2005], depending on the cooling conditions of the latter. In addition to energy recovery, this thesis has demonstrated the possibility of developing silicon-based thermoelectric coolers, paving the way for the possible integration of thermoelectric coolers into silicon-based microelectronic devices.

Abstract:

The blooming of the internet of things (IoT) and wireless sensors nodes seems to be delayed owing to the lack of reliable, safe and low-cost source [Nordrum 2016]. Thermoelectric harvesters feature those key advantages. Silicon presents the advantages to be most abundant, environmental less harmful and to benefit from facilities and technological processes for low cost thermoelectric harvesters mass production compared to the conventional materials (bismuth telluride alloys). However, silicon is a poor thermoelectric material due to its high thermal conductivity (150Wm^(-1) K^(-1)) [Haras et al. 2015]. The possibility to reduce the thermal conductivity while preserving electrical conductivity and Seebeck coefficient is the key to upgrade silicon as an efficient thermoelectric material. To that end, efforts are oriented towards the phononic part of heat transport, which is the dominant contribution in semiconductors [Jin 2014]. The researches carried out during this thesis dealt with the integration phonon engineered silicon membranes [Haras 2016; Lacatena et al. 2014; Tang et al. 2010; Yu et al. 2010] into thermoelectric harvester demonstrators and their characterizations with respect to the state of the art.

The studies' results demonstrated the feasibility of a silicon based thermoelectric harvester exhibiting performances sufficient for autonomous sensor nodes' power supplying [Vullers et al. 2009] and comparable performances with the bismuth telluride state of the art harvester [Bottner 2005] according to the harvesters' cooling conditions. Moreover, this thesis demonstrated in addition to the energy harvesting, the possibility of developing silicon based thermoelectric coolers, opening the way to possible integration of thermoelectric coolers in silicon based micro-electronic devices.