Ultra-fast, Antenna-coupled Mid-infrared Quantum-well Photodetectors
Author | : Quyang Lin |
Publisher | : |
Total Pages | : 0 |
Release | : 2021 |
ISBN-10 | : OCLC:1292926761 |
ISBN-13 | : |
Rating | : 4/5 (61 Downloads) |
Book excerpt: This Thesis is devoted to the conception, fabrication and experimental characterization of semiconductor-based ultra-fast photodetectors operating in the mid-infrared range (~3-12um). More specifically, the detectors that I have developed, generally known as multi-quantum-well infrared photodetectors (QWIPs), rely on intersubband (ISB) transitions in a GaAs-Al_0.2Ga_0.8As heterostructure, where an electron occupying the ground state of a quantum-well is photoexcited into an upper state, lying next to the energy continuum above the AlGaAs barriers.In my work I have exploited a specific device geometry that allows light-coupling at normal incidence, based on a two-dimensional array of electrically connected metallic patch-antennas. Each antenna is obtained by sandwiching the GaAs-AlGaAs multi-quantum-well heterostructure between a top contact metal layer and a bottom metallic ground plane, effectively forming a square metal-dielectric-metal microcavity, where the fundamental TM electromagnetic mode is resonant with the energy of the ISB transition. Finally, to allow for broadband microwave extraction, the antenna array is connected to a 50Ohm, monolithically integrated coplanar waveguide.In the first part of my work I have designed the antennas for optimum detection at 10um wavelength. This was done by running a set of simulations using a commercial electromagnetic solver based on the finite-difference time-domain (FDTD) method. Based on the results of the simulations I have fabricated a set of preliminary structures, without coplanar waveguide, to characterize the optical properties of the antenna array through Fourier transform micro-reflectance measurements. These measurements have allowed me to select the optimum patch array dimensions, namely the lateral size of the square-patch and the array periodicity.The second part of my work has been dedicated to the fabrication of the complete QWIP detector, including the monolithically integrated coplanar waveguide. In these detectors the size of the two-dimensional antenna array has been kept to a minimum, without compromising the radiation collection, in order to reduce as much as possible the device parasitic RC time constant and therefore maximize the detector speed. I have fabricated two generations of detectors relying on two slightly different active regions, respectively based on a bound-to-bound and a bound-to-continuum design. In the final part of my PhD I have also fabricated a third generation of devices, where the patch array, rather than to a coplanar waveguide, is connected to a spiral THz antenna. This device has not been characterized in this work and I present its relevance in the context of this Thesis in the perspectives.The last part of the Thesis is dedicated to the electro-optical characterization of the fabricated detectors. First, I have measured the dark current, the polarization dependence, and the dc photoresponse, that allowed me to determine the responsivity at 77K and 300K. Then I characterized the microwave frequency response of the detectors. To this end I have participated to the setup of an experimental apparatus based a high-speed (67GHz) cryogenic probe station. In this apparatus the beams of two quantum cascade lasers (QCLs) emitting at 10.3um wavelength, are simultaneously focused on the QWIP detector to generate a hererodyne signal at their difference frequency. By temperature/current tuning the emission wavelength of one QCL the heterodyne frequency can be swept continuously, thus allowing the measurement of the detector frequency response with the help of a spectrum analyzer. At room-temperature I obtain a flat frequency response up to 70GHz, solely limited by the bandwidth of the acquisition electronics. This is the broadest RF- bandwidth reported to date for a QWIP photodetector. To analyze the experimental data, I have modelled the electrical behavior of the QWIP using a small-signal equivalent circuit model.