Resumen de Enhancing the radiated power in the Terahertz band

Belén Andrés García

• The needs of the Technology, Science and Communications to have available the Terahertz band has grown dramatically during the last years. This band is still a challenge from the technological point of view. Considering two approximations, photonic and microwave, the first is not capable of delivering enough power to have a reliable communication or an efficient system. Regarding microwave, as the existing devices and systems should be scaled in frequency, the available manufacturing techniques and design processes are not capable of providing enough precision for the miniaturized devices on these frequencies. Despite these issues, the efficiency on the devices, either in the sources itself or in the antennas, regarding radiated power or coupling of electromagnetic fields, is not good enough, and most of the power is lost either on the lasers illuminating the devices, in the bias or simply in the coupling and matching. It is important to know, also, that classical antenna theory, and the well known antenna topologies, need to be redesigned to allow the limitations of the technology in manufacturing process and also in the measurement of devices, for example, the design in high permittivity substrates or semiconductors. Photomixers are, from the optics point of view, one of the most common way to generate a Terahertz signal. Their properties, such as easiness in manufacturing, tunnability, high integrability, operation at room temperature, and reduced cost make them affordable devices to implement in new systems, both with commercial purposes and also in the research field. Gathering together all the circumstances, the optimization of the radiation elements that serves as a matching between free space and the feed, is a key point to develop efficient devices as well as reduced in cost. It is important to develop radiating elements that can be processed and measured in the same wafe of the THz sources, reducing this way the complexity of the system. This means the radiating elements should be preferably in a technology that allows the manufacturing at the same procedure than the source. The main goal of the proposed Thesis is the development and design of new devices, that allow the optimization of radiaton power and coupling between the THz source and the radiating element. This goal has been achieved with different antenna topologies, either in planar structures and in 3D devices. The first block of the Thesis is based on the use of photomixing devices, framed in a research stage at FAU (Erlangen), for the optimization of the radiating element adapted to n-i-pn-i-p photomixers. From this optimization, a reduction in the size of the element has been achieved, allowing the implementation of the next step, which are new 3D horns etched in the semiconductor material where the PM is processed. The second block of the present work is the development of a modal theory for TSA antennas together with some applications that allows the validation of the theory as well as the implementation over semiconductor or high permittivity substrates. The baseline of waveguide theory was employed for the analysis with PMC conductors and air discontinuities. Three different applications have been designed, manufactured and measured, with good results in the band over study. The three of them are based on new geometries on the substrate where the antenna is printed, including EBG, substratesuperstrate configuration and thick wedges. The third block is based on numerical methods, with the objective of analysis acceleration as well as the reduction of computational resources in the calculations and in simulations. First of all, QO techniques are employed, with the aid of the BME coefficients of the feed. A complex optical system including lens and mirrors can be analyzed in a matter of seconds, with a reduced error. To continue with resources optimization and also the analysis of reflectors, a combination of MoM+PO has been developed based on Krylov subspaces, imposing orthogonality on each step, together with the use of Macro Basis Functions; the feed is analyzed with MoM, and the impedances matrix is obtained, while the matrix from the reflector is computed with a PO approach. The complexity of the problem is reduced by a Nf/P factor with Nf the number of unknowns in the reflector and P the number of MBFs. To continue with MoM, and the analysis of large arrays, a method based on the modification of the impedance matrix taking into account the effect on impedance of the surrounding elements is developed. -----------------------------------------------