Integration thermoelektrischer Generatoren in metallische Gussteile
Final Report Abstract
The novel technological process developed within the scope of this project allows the fabrication of thermally and mechanically stable MEMS-structures, which were demonstrated in the example of a Seebeck-effect-based thermogenerator (TG). The TG can be effectively applied in hightemperature environments. It was tested up to 730°C and did not lose its functionality. Considerable power output of 515µW was achieved. We have demonstrated that a semiconductor device based on this technology can be directly embedded in molten metals such as aluminum, magnesium, brass and bronze. The thermomechanical stress (TMS) during casting and cool down was absorbed by a chip with an amorphous borosilicate glass (BSG)-substrate and a wavy layout of the interconnectors to allow stress compensation. Three different embedding processes were tested: dip soldering, high pressure die casting (HPDC) and metal-spray-deposition (MSD). Due to the combination of high TMS and long cool-down time, the traditional dip soldering can only be used when the melting temperature is low (e.g. tin –> 235°C). The two novel approaches were successfully implemented at relatively high temperatures: HPDC with aluminum (700°C) and MSD with bronze (500°C). Due to the low cool-down time, the first can be recommended when the main issue is the damage caused by the high temperature itself. On the other hand, lesser mechanical forces applied during the spraydeposition and reduced TMS serve to significantly improve the yield in comparison with HPDC. Aiming towards the application with high thermal gradient, it was necessary for the project realization to determine the thermopower of silicon with different dopant type and concentrations for the whole temperature range from the standard environment up to the metal melting point. But neither theoretical solution with adequate reliability, nor sufficient experimental data, nor adequate FEM software was available. In order to solve the problem, a semianalytical model (SAM) for the Seebeck coefficient (SC) was introduced within the scope of the project. This model is based on the temperature dependence of the electrochemical potential, which is determined not only through theoretical relations, given by the foundations of common semiconductor physics, but also with the help of experimentally defined temperature dependences of the intrinsic carrier density and carrier mobilities, which are well known for all common semiconductor materials. A comparison with experimental data for five different semiconductors proves the ability of the introduced approach to provide reliable predictions over a wide range of parameters (temperature, dopant type and concentration). Also, during the project’s realization, pilot side work was carried out: the novel design of a TG, based on a variable capacitor, whose plate separation is cyclically changed by a micromachined heat engine (HE), is introduced. Its prototype was fabricated using the silicon micromechanics technique. The prototype’s working membrane was brought to vibration, thus simulating the working cycle of the HE and affirming the introduced TG design. Potentially, the TG's performance can be improved, when other promising materials are integrated into the proposed technology. For example, porous instead of bulk silicon can be used and fused silica can be implemented instead of the BSG as the substrate material. The reduction of the contact resistivity (CR) at low temperature is required to broaden the application range of the device (possible approaches: metal choice or burn-through). Such exponential CR characteristics, as those shown by the WTi-Si contact, are often associated with some kind of p-njunction structures. Their exact nature will be addressed separately in future work. The goal of such research could not be the simple reduction of the CR, as it is determined in the TG case, but finding the possibility to precisely define the junction's characteristics, thus potentially opening the way to the novel high-temperature stable class of metal-silicon based semiconductor devices. The TGs embedding in other metal, glass or ceramic compounds during their manufacturing (e.g. the casting process) will be the subject of future research, which additionally has to include further development of the preparation stage with corresponding customizations to each application. Establishing reliable methods of the combination of micro- and macro-components in one working part would be truly useful in a wide range of applications where sophisticated MEMS-devices should be integrated into high-payload working parts. The further development of the SAM may include the discussion of the SC in degenerated semiconductors, as well as the implementation of the proposed semi-analytical approach in other areas of semiconductor physics (e.g. a detailed description of the thermal drift of the p-n-junction’s parameters). The construction of the whole functioning TG, based on the micromachined HE-design, will also be the subject of further research.
Publications
- “A thermoelectric energy harvester directly embedded into casted aluminum" IEEE Electron Device Lett. 33(2), 233 –235 (2012)
A. Ibragimov, H. Pleteit, C. Pille, and W. Lang
(See online at https://doi.org/10.1109/LED.2011.2174605)
