| dc.description.abstract | For centuries, optical components were created using bulk components of various refractive indexes where light accumulates phase over a path much longer than its wavelength. With the advent of nanofabrication techniques, it has been demonstrated that manipulating light using metasurfaces was possible and could yield performances comparable if not better than bulk elements like lenses or prisms. As to this date, most optical metasurfaces have a functionality fixed at fabrication, the future of such metasurfaces is reconfigurability post-fabrication. Creating tunable metasurfaces would allow entire complex optical systems to be manufactured on-chip at a lower cost, with better reliability and energy efficiency. Light Detection and Ranging (LIDAR) systems are of great interest to enhance machine vision capabilities, notably for the automotive industry that is putting tremendous R&D efforts into autonomous vehicles. The current use of complex mechanical systems to achieve beam steering at acceptable rates ( ~1 million points per second) over a large number of resolvable points (640*16 for the Valeo Scala LIDAR) yields high production costs and reliability issues. Flash LIDARs scan the entire field of view all at once but require high-energy flashes to receive enough photons in the detectors and therefore need large VCSEL arrays that are expensive and energy-consuming. RADAR systems encountered similar problems in the mid-20th century as the massive antenna required to obtain a sharp beam could not be steered mechanically fast enough. The use of phased arrays enabled fast scanning rates over a wide far-field with no moving parts
by controlling the phase of a cluster of radio-waves emitters. Down-sizing phased arrays in the optical and NIR domain is no small task as the technologies used to control centimetric waves are not applicable to radiation at wavelengths of the order of a micron. There are many challenges associated with such a technological breakthrough. At that scale,nanofabrication techniques must be used which seriously impedes the range of materials and shapes of components. Furthermore, a device must be monolithic as assembling parts is out of the question at that scale. Tunable elements will also have to be included to control an output phase by user-controlled inputs, like a voltage. Several paths have been proposed and tested to achieve sub-wavelength reconfigurable phase control antennas, we will mostly focus on resonant antennas as this solution is simple, compact and energy efficient. A strong light-matter interaction at resonance within such antennas can yield a large phase shift with a relatively small change in wavelength for a given antenna or with a modification in the antenna properties for a given wavelength.
In order to reach full reconfigurability, a tunable element has to be included in each antenna, phase change materials are particularly well suited for this application as they offer a very high index modulation for low energy consumption. Vanadium dioxide (VO2) in particular showcases an insulator to metal transition (IMT) around 68? that supports a wide optical index change (~ unity), that can be triggered easily and where intermediate material properties can be obtained. This makes VO2 the perfect candidate for a tunable element in a resonant antenna phased array to achieve a compact, reliable and energy-efficient solution.
Resonant nanoantennas have already been investigated in the near-infrared (NIR) and optical domain and they suffer from inherent flaws like amplitude-phase correlation, non 2? phase shift and non-linearity between the input signal and output phase shift. These flaws are inherently due to the resonant nature of the devices and cannot be individually mitigated without worsening one or more of these three problems. We propose a novel control algorithm, binary control, for beam steering that enables us to ignore all these issues and can be ideally implemented with resonant nano-antennas, unlike the usual ideal control scheme. The binary control algorithm is clear and simple to implement, enables continuous beam steering over a wide field of view and can be applied without any loss of generality to any phased array, regardless of the wave nature or scale.
In chapter 2 of this Thesis, we will first present classical phased array theory, the current state of the art in the optical/NIR domain and how they can be improved based on RADAR research. We will then investigate Vanadium dioxide, its optical properties, behaviour around transition and how it can be manufactured, with an emphasis put on pulsed laser deposition (PLD). We then present a number of articles that have each pushed the boundaries of resonant optical/NIR metasurfaces in the past decade with theoretical and experimental investigations.
Chapter 3 will focus on theory, deriving a number of equations that will be useful later in the thesis and theoretically analysing binary control. We will see how well it performs compared to ideal control, the challenges associated with this novel control mode and how they are addressed. The robustness of binary control to manufacturing inaccuracies is also to be discussed as they are frequent in resonant nanostructures. The finite difference time domain (FDTD) method will also be shortly analysed as it is the main method used in this project to analyse the nanostructures, we will finish by a thorough description of the FDTD setups we used in this research project.
We will then focus in chapter 4 on inverse design, this design technique was used in this project to find and fine-tune a nanostructure with success. An overview of the subject will be provided to the reader along with a brief explanation of the common challenges associated with inverse design. The two algorithms we used are then presented along with their advantages and drawbacks. We motivate their choice in the final section as there are many methods to implement inverse design in practice.
In chapter 5, we will outline the design of a realistic resonant nanoantenna capable of dynamic phase control that ideally implements binary control in the NIR. We explain in detail the design process to achieve an individual thermally controlled phase shift for each antenna in a phased array and assess the performance of the structure. We also present variations on the design that have been obtained using the same design ideas but with different input design parameters like the operating wavelength, array period, or materials.
As we will see later, resonant nanostructures are capable of amplitude modulation, while this effect is problematic in phased arrays, it can be exploited separately in different devices. We will see in chapter 6 how VO2 driven nanostructures are capable of reflectance modulation in the optical/NIR domains, for narrowband or broadband applications or even colour generation.
We finish off this work with a conclusion where we outline future working directions that are certainly promising given the results shown in this thesis. | en |
