Terahertz technology is rapidly growing for applications in various fields such as medical sciences, remote sensing, material characterization, and security. This accelerated growth has motivated engineers to develop compact, portable, and cost-effective terahertz sources and detectors. Terahertz generation and detection can be achieved using photoconductive antennas (PCAs), which have unique advantages. Notably, they do not require a vacuum or cryogenic cooling to function. PCAs operate on the principle of photoconductivity, which allows for compact integration with a fiber optic laser. It is also possible to launch THz radiation to a waveguide, which can be used for making a robust THz spectroscopy system. Ultra-short laser pulses are available in both 800 nm and 1550 nm wavelengths. However, the 1550 nm window has distinctive advantages such as availability of fiber amplifiers and fiber based electro-optical components at a relatively lower cost. The goal of this research is to introduce cost-effective and state-of-the-art solutions to develop THz transceivers for use in terahertz time-domain spectroscopy (THz-TDS) at 1550 nm wavelength. In this thesis we explore three approaches for enhancing THz emission and reception using PCAs. First, an array of hexagonal shape plasmonic nano-structures was used to increase the optical field coupling to the minimum depth of the substrate. Next, nano-structures also helped with enhancing the local electric field inside a low-cost semi-insulating GaAs substrate. This technique resulted in a 60% enhancement of the THz emission compared to a commercial LT-GaAs based PCA with antireflection coating. Moreover, the plasmonic nano-structures efficiently remove heat from the gap area allowing for operation at higher bias voltages. Plasmonic structures on LT-GaAs were investigated, which use a mid-gap Arsenic defect state to absorb 1550 nm light. The plasmonic devices were found to outperform existing InGaAs substrate based THz devices by factor of two. Finally, optimization of the LT-GaAs growth and annealing conditions was investigated to maximize the THz signal at 1550 nm. Outcomes of this research pave the way for designing cost-effective THz transceivers for time domain Terahertz spectroscopy systems at 1550 nm wavelength. / Graduate

In this thesis, we describe the development of a time-domain terahertz (THz) spectrometer driven by two different laser sources: an Er-doped femtosecond fiber laser and a mode-locked Ti:Sapphire laser. The resulting THz electromagnetic radiation was generated and detected using photoconductive antenna detection methods in both systems. In these experiments we characterized the THz power output for both the fiber laser driven system and the Ti:Sapphire laser driven system. Emphasis is given throughout this thesis on understanding the working principles behind time-domain terahertz spectroscopy, applications of THz radiation and terahertz generation as well as terahertz detection methods. We calculated the THz power output using two different methods. By using the &ldquo / Hertzian Dipole&rdquo / method we estimated the generated THz power after the generation photoconductive antenna. Using this method, we showed that the v generated power is on the order of milliwatts, which is far larger than the expected power typical for these systems. The second, &ldquo / Open-Circuit Voltage&rdquo / method, allowed us to calculate the received power on the detection photoconductive antenna. Using this method we were able to show that the THz power generated and detected in these systems is on the order of microwatts. For the mode-locked fiber laser driven spectrometer we obtained on average a ~ 4 ps (0.25 THz) pulse length which corresponded to an average power in the range of 71.8 nW - 70.54 &amp / #956 / W on a dipole antenna with a 6 &amp / #956 / m dipole gap and 44 &amp / #956 / m dipole length. Using the mode-locked Ti:Sapphire laser driven spectrometer we observed a ~ 2 ps (0.5 THz) pulse length and average power in the range of 0.54 nW &ndash / 5.12 &amp / #956 / W on a different dipole antenna with a 5 &amp / #956 / m gap and 40 &amp / #956 / m dipole length. Since these values agree with expected values for these systems we believe the &ldquo / Open-Circuit Voltage&rdquo / method is appropriate when trying to calculate the THz power.

Time-domain terahertz spectroscopy is now a well-established technique. Of the many methods available for a terahertz source for terahertz spectroscopy, the most widely used may be the GaAs-based photoconductive antenna, as it provides relatively high power at terahertz frequencies, commercially available up to 150 µW, and a wide-bandwidth, approximately 70 GHz to 3.5 THz. One of the limitations for developing more accurate and sensitive terahertz interrogation techniques is the lack of higher power sources. Because of our research interests in terahertz spectroscopy, we investigated detailed design and fabrication parameters involved in the photoconductive antenna, which exploits the surface plasma oscillation to produce a wideband pulse. The investigation enabled us to develop a new photoconductive antenna that is capable of generating a high power terahertz beam, at least twenty times stronger than those currently available. Throughout this research, it was discovered that antenna electrodes with particular geometries could produce superradiance, also known as the Dicke effect. Chaotic electrodes with a predisposition to lead charge-carriers into chaotic trajectories, e.g. rippled geometry, were exploited to reduce undesirable heat effects by driving thermal-electrons away from the terahertz generation site, i.e. the location of the surface plasma, while concentrating the removed charge-carriers in separate locations slightly away from the surface plasma. Then, spontaneous emission of coherent terahertz radiation may occur when the terahertz pulse generated by the surface plasma stimulates the concentrated carriers. This spontaneous emission enhanced the total coherent terahertz beam strength, as it occurs almost simultaneously with the primary terahertz beam. In principle, the spontaneous emission power increases as N^2, with the number N of dipole moments resulted from the concentrated charge carriers. Hence, if the design parameters are optimized, it may be possible to increase the strength of coherent terahertz beam by more than one order of magnitude with a photoconductive antenna containing rippled electrodes. However, as the parameters are yet to be optimized, we have only demonstrated 10-20 % enhancement with our current photoconductive antennas. Photoconductive antennas were fabricated via photolithography and characterized by time-domain terahertz spectroscopy and pyroelectric detection. In addition to chaotic electrodes, a variety of other parameters were characterized, including GaAs substrate thickness, GaAs crystal lattice orientation, trench depth for electrodes, metal-semiconductor contact, and bias voltage across electrodes. Nearly all parameters were found to play a crucial role influencing terahertz beam emission and carrier dynamics. By exploiting wave chaos and other antenna parameters, we developed a new photoconductive antenna capable of continuous operation with terahertz power twenty times larger than that of the conventional photoconductive antennas, improving from 150 µW to 3 mW. With further optimizations of the parameters, we expect more dramatic improvement of the photoconductive antenna in the near future. / Physics

The frequency demands of radiating systems are moving into the terahertz band with potential applications that include sensing, imaging, and extremely broadband communication. One commonly used method for generating and detecting terahertz waves is to excite a voltage-biased photoconductive antenna with an extremely short laser pulse. The pulsed laser generates charge carriers in a photoconductive substrate which are swept onto the metallic antenna traces to produce an electric current that radiates or detects a terahertz band signal. Therefore, analysis of a photoconductive antenna requires simultaneous solutions of both semiconductor physics equations (including drift-diffusion and continuity relations) and Maxwell’s equations. A multi-physics analysis scheme based on the Discontinuous-Galerkin Finite-Element Time-Domain (DGFETD) is presented that couples the semiconductor drift-diffusion equations with the electromagnetic Maxwell’s equations. A simple port model is discussed that efficiently couples the two equation sets. Various photoconductive antennas were fabricated using TiAu metallization on a GaAs substrate and the fabrication process is detailed. Computed emission intensities are compared with measured data. Optimized antenna designs based on the analysis are presented for a variety of antenna configurations.

Le sujet de la thèse a porté sur la mise au point, la caractérisation et l'utilisation de matériaux semi-conducteurs, au sein desquels les porteurs libres ont un temps de vie extrêmement brefs (picoseconde ou sub-picoseconde), pour réaliser des antennes photoconductrices émettrices ou détectrices de rayonnement électromagnétique térahertz (THz). Contrairement au semi-conducteur LTG-GaAs (low temperature grown GaAs) à la technologie bien dominée et aux performances exceptionnelles lorsque photo-excité par des impulsions lasers de longueurs d'onde typiquement inférieures à 0,8 µm, le travail portait ici sur des matériaux permettant l'emploi de lasers dont les longueurs d'onde sont celles des télécommunications optiques, à savoir aux alentours de 1,5 µm. L'intérêt est de bénéficier de la technologie mature de ces lasers, et du coût relativement modique des composants pour les télécommunications optiques. Pour réaliser des antennes THz performantes et efficaces, le matériau semi-conducteur doit présenter plusieurs qualités : vie des porteurs libres très courte, grande mobilité des porteurs, haute résistivité hors éclairement, et bonne structure cristallographique pour éviter les claquages électriques. Pour obtenir une courte durée de vie, on introduit un grand nombre de pièges dans le semi-conducteur, qui capturent efficacement les électrons libres. Pour les matériaux de type InGaAs employés à 1,5 µm, le problème est que le niveau en énergie de ces pièges, par exemple pour les matériaux épitaxiés à basse température, est très proche de la bande de conduction du semi-conducteur. Cela est équivalent à un dopage n du matériau, ce qui en diminue fortement sa résistivité hors éclairement. Plusieurs solutions ont été apportées par différents laboratoires : compensation par dopage p pour les matériaux épitaxiés à basse température, bombardement ionique, implantation ionique, ou même structures à couches alternées où la photo-génération et la recombinaison des porteurs libres se produisent à des endroits différents. Le but du travail de thèse était de fabriquer des matériaux préparés suivant ces différentes techniques, de les caractériser et de comparer leurs performances pour l'optoélectronique THz. Les semi-conducteurs à étudier étaient de type InGaAs comme déjà publiés par la concurrence, l'originalité de thèse portant sur la comparaison de ces différents matériaux et si possible leur optimisation,. Au cours de ce travail de thèse, de nombreuses couches d'InGaAs ont été épitaxiées, en faisant varier les paramètres de dépôt, et des antennes THz ont été fabriquées. Les couches ont été caractérisées du point de vue cristallographique, ainsi que pour la conductivité électrique DC (mesures 4 pointes, mobilité Hall…), les propriétés d'absorption optique (spectroscopie visible et IR), la durée de vie des porteurs par mesure optique pompe-sonde. Pour les couches épitaxiées à basse température, l'influence d'un recuit thermique ainsi que du dopage en béryllium ont été étudiés. Dans le cas de couches bombardées ou implantées, plusieurs ions ont été utilisés, le brome, le fer et l'hydrogène. Les relations entre la cartographie des défauts structuraux et/ou des ions implantés et les propriétés électriques et de dynamique des porteurs ont été examinées en détail. Ces études permettent de comprendre le type de défauts qui piègent les porteurs dans ces matériaux, ainsi que leur formation lors du processus de fabrication et de traitement des couches. Finalement les meilleures couches fabriquées présentent des performances comparables à celles publiées par ailleurs. Les derniers travaux de thèse ont permis d'obtenir les premiers signaux de rayonnement THz générés par une antenne fabriquée avec l'InGaAs optimisé. / The subject of the thesis focused on the development, characterization and use of semiconductor materials, in which the free carriers have a very short lifetime (picosecond or sub-picosecond) to produce photoconductive antennas emitting and detecting electromagnetic terahertz (THz) radiation. Unlike semiconductor LTG-GaAs (low temperature grown GaAs) which is a well-dominated technology and present exceptional performances when photoexcited by typically less than 0.8 micron wavelength laser pulses, the work focused on here materials for the use of lasers whose wavelengths are those of the optical communication, namely around 1.5 microns. The interest is to benefit from the mature technology of these lasers, and relatively low cost components for optical telecommunications. To achieve effective and efficient THz antennas, the semiconductor material must have several qualities : lifetime of free carriers very short, high carrier mobility, high resistivity outside lighting, and good crystallographic structure to prevent electrical breakdown. For a short lifetime, a large number of traps are introduced into the semiconductor, which effectively capture the free electrons. For InGaAs materials used at 1.5 microns, the problem is that the energy level of the traps, for example, the epitaxial material at low temperature is very close to the conduction band of the semiconductor. This is equivalent to an n-doped material, what greatly reduces its resistivity outside illumination. Several solutions have been made by different laboratories : compensation for the p-doped epitaxial materials at low temperature, ion bombardment, ion implantation, or even alternating layer structures where photo-generation and recombination of free carriers occur in different places. The aim of the thesis was to produce materials prepared using these techniques to characterize and compare their performance to THz optoelectronics. The studied InGaAs-based semiconductors were as previously published by the competition, the originality of the thesis was on the comparison of these different materials and if possible their optimization. During this work, many of InGaAs layers were grown epitaxially by varying the deposition parameters, and THz antennas were fabricated. The layers were characterized from the crystallographic point of view, as well as the DC electrical conductivity (measures 4 points, Hall mobility ... ), the optical absorption properties (visible and IR spectroscopy ), the lifetime of carriers by optical pump-probe measurement. For low temperature epitaxial layers, the influences of thermal and doping beryllium annealing were studied. In the case of shelled or implanted layers, several ions were used, bromine, iron and hydrogen. The relationship between the mapping of structural defects of the implanted ions and electrical and carrier dynamics properties were discussed in detail. These studies allow us to understand the type of defects that trap carriers in these materials, as well as training in the process of manufacturing and processing layers. Finally the best layers are made comparable to those published elsewhere performance. The last study allowed to achieve the first signals of THz radiation generated by InGaAs-based optimized antenna.