This thesis investigates whether a wafer-defined circular blade can realise a mechanically robust cutting edge at thin-film thickness scale and perform first material removal under rotation. The concept uses a silicon annulus carrying a deposited film. Selective substrate removal near the outer perimeter releases a circumferential overhang whose thickness defines the nominal kerf scale, while overhang length governs achievable cutting depth. The work is structured around four challenges: reaching the minimum cutting-edge speed with acceptable run-out, fabricating annular blade blanks, selectively underetching the rim while preserving mounting geometry, and integrating the system to demonstrate cutting.

A practical spindle operating envelope was established in which the blade reached the minimum target speed while operational displacement remained predominantly rotation-synchronous and below the realised overhang range used for survival and engagement trials. LS-Precess femtosecond-laser cutting enabled repeatable fabrication of 525 µm-thick silicon annular blanks, and a lid–base sacrificial masking concept enabled selective SF₆ underetching and controlled release of SiO₂ rims while preserving hub-bore integrity when the lid remained intact. In the integrated demonstration, released SiO₂ rims survived operation at and above target speed and produced a continuous kerf in rigid ABS under incremental approach at N = 17.7 kRPM, with measured trench depths of 11.72 µm to 14.71 µm and a central kerf width of ~46 µm.

The main limitations are that rigid cutting was attributable primarily to the rear-side ~10 µm rim, sustained controlled engagement of the front-side ~4 µm ultra-thin rim remains unproven, and circumferentially continuous ultra-thin a-SiC edges were not reliably achieved within the present process window.

In this thesis, a femtosecond pump-probe laser setup is used to study GHz acoustic waves in suspended phononic crystal (PnC) waveguides. These membranes consist of 200 nm thick Si3N4 coated with a 20 nm gold film and shamrock shaped PnCs. Various COMSOL simulations of these waveguides are conducted to investigate the expected behavior of the membranes. These simulations not only reveal the eigenmodes and phononic bandgaps of two different membranes, but also show the formation of a 300 MHz Lamb wave inside the waveguide. Experimental results show that this 300 MHz Lamb wave is confined inside the waveguide, when the frequency of the Lamb wave falls within the phononic bandgap of the membrane. For membranes with a bandgap that does not include the frequency of the Lamb wave, propagation through the PnC lattice is seen. Finally, an Acousto Optic Modulator (AOM) is successfully integrated in the pump-probe laser setup to increase the frequency resolution of the measurements. The experimental data obtained with this higher frequency resolution shows similar results of wave confinement and proves repeatability of the measurements.

Recent advancements in nanofabrication have enabled the creation of nanomechanical resonators (NMRs) with extreme aspect ratios, paving the way for high-performance resonators that can couple to light, quantum systems or other matters. A key requirement for such applications is achieving an exceptionally high mechanical quality factor, Q, which indicates minimal energy dissipation and strong isolation from environmental noise. Among various structural approaches, periodic phononic crystal (PnC)-based membrane resonators have demonstrated high Q values (10e8) with outstanding practicality, but their performance is fundamentally limited by the constraints of two-dimensional periodicity.

This thesis proposes an alternative design strategy based on 2D quasicrystal (QC) geometries, which are aperiodic yet possess rotational symmetries in certain variants. QC-based designs offer greater flexibility and potentially richer dynamics compared to conventional periodic PnCs. However, their geometric complexity and lack of established theoretical frameworks require advanced, computation-heavy design and optimization techniques. To address this, we introduce a data-driven design and optimization framework tailored for QC-based resonator designs. Our results demonstrate the promising potential of moving beyond periodic structures to aperiodic designs in the pursuit of ultra-high Q nanomechanical resonators.

Single-molecule detection is essential for investigating molecular interactions and dynamics but is often constrained by weak fluorescence signals and the potential for photodamage under high excitation intensities. Conventional enhancement approaches typically involve dried samples or introduce quenching and compatibility issues, making them unsuitable for in-solution measurements. This work presents a hybrid photonic platform that integrates a photonic crystal (PhC) slab with a hexagonal boron nitride (hBN) layer on top as a biocompatible substrate, offering a promising route for fluorescence enhancement under physiological conditions. Additionally, the atomically flat surface of hBN minimizes fluorophore trapping at edges or defects, enabling more uniform and reproducible single-molecule measurements across large areas.

Electromagnetic field simulations based on Rigorous Coupled-Wave Analysis (RCWA) revealed that the PhC structure can concentrate light at the hBN surface, leading to electric field intensity enhancements of up to 259-fold. To experimentally validate this, hBN flakes were successfully stamped onto the center of the PhC patches and characterized by optical microscopy and atomic force microscopy (AFM). The PhC patterns themselves were analyzed with both AFM and scanning electron microscopy (SEM). Single-stranded DNA (ssDNA) labeled with the fluorophore Atto647N was deposited and imaged by confocal fluorescence microscopy.

The experimental results demonstrated consistent fluorescence enhancement, with average increases of up to 13.15-fold across full flakes and localized enhancements up to 18.22-fold for individual fluorophores, likely positioned in high-field regions. Despite these successes, autofluorescence from the PhC introduced background signal and different flake thicknesses, complicating quantitative analysis.

This work confirms that the PhC–hBN platform can significantly enhance fluorescence in solution, providing a viable path forward for high-sensitivity single-molecule biosensing. It also highlights current challenges, such as autofluorescence and structural sensitivity, that can be subject of future research.

Microelectromechanical resonator sensors are crucial in the cutting-edge technologies used in our everyday communication, timekeeping and computing systems. Their extreme sensing capabilities make them ideal candidates for the innovation of future technologies. However, with our ever-growing desire for faster communication, more sensitive systems, and more advanced technologies comes the need for a new generation of resonator sensors. This next generation will have to be faster, more accurate, and just as cheap as their predecessors if they are to enable the rapid growth of our technological needs. In this thesis, we investigate recently fabricated state-of-the-art extreme aspect ratio membrane resonators. The characteristics of extreme aspect ratio membrane resonator sensors are researched, and
the effects of nonlinear forces on their operation are explored. Some of these nonlinear attractive forces, such as the Casimir effect, are common to the extreme dimensions of these resonators. Another common nonlinear attractive force in MEMS, the electrostatic force, and its effects on resonator operation and output are investigated as well. Analytical models are fashioned and a FEM model is produced and validated using experimental results, showing it reflects reality. FEM simulations show that for these extreme aspect ratio resonators, the nonlinear softening effect is solely responsible for the change in the eigenfrequency which proves to be able to boost the responsivity of these resonators by factors of hundreds to thousands. Models are investigated for both conductors and dielectric resonators with different geometries and different material parameters, which all show these results. Responsivities of 133.2 kHz/kPa and 1.6 kHz/nm are found, which exceed the state-of-the-art. The negative effects of nonlinear forces such as pull-in are considered, investigated, and models are produced which predict them to prevent device failure. Furthermore, the role of crucial resonator parameters is investigated to aid future research in leveraging this potential new technique of enhancing sensor capabilities.

Accurate constitutive modeling of hyperelastic materials remains a challenging task due to their inherently nonlinear and complex stress–strain behavior. Traditional phenomenological models often fall short in capturing this complexity, particularly in modern engineering materials with rich mechanical responses. In recent decades, data-driven modeling approaches have emerged as promising alternatives, offering flexibility in learning material behavior directly from data. Multi-Layer Perceptrons (MLPs), in particular, have become widely adopted due to their universal approximation capabilities. Despite their benefits, MLP-based approaches face significant limitations. Their "blackbox" nature limits interpretability and restricts insights into underlying material mechanics. Furthermore, although MLPs with fixed activation functions can approximate hyperelastic behavior in theory, their limited smoothness, such as in the case of ReLU, can restrict accurate representation of derivatives essential for modeling material responses. These shortcomings highlight the need for alternative frameworks that can represent material behavior more accurately and transparently. An emerging alternative is the Kolmogorov-Arnold Network (KAN), which offers improved interpretability and greater flexibility due to its architecture. By leveraging the Kolmogorov-Arnold representation theorem, KANs decompose complex functions into simpler, easy-to-understand components. WhileKANs have shown promise in various applications, including material modeling, their use in hyperelasticity remains limited due to challenges in ensuring physically consistent predictions. Current KAN-based frameworks cannot guarantee physically valid hyperelastic modeling. To address these challenges, this work introduces a novel Input-Convex Kolmogorov-Arnold Network (ICKAN) architecture tailored for hyperelastic constitutive modeling. The ICKAN model employs spline-based, learnable activation functions to capture material nonlinearities and explicitly incorporates convexity and monotonicity constraints to ensure adherence to physical principles. Validation using benchmark datasets demonstrates that ICKAN accurately predicts hyperelastic stress–strain behavior across a range of loading conditions. By enhancing interpretability and ensuring physically consistent predictions, the proposed ICKAN framework provides a robust and transparent solution, underscoring the broader potential of KANs in data-driven constitutive modeling.

Optical manipulation has recently become indispensable in various disciplines and optical levitation specif- ically offers many possibilities in highly sensitive motion detection. The counterintuitive "tractor beams" have therefore captivated research due to their possible new approach to levitation. The key principle in de- signing an optical pulling force is the enhancement of the forward momentum by the pulled object. The idea of using the light modulating capacities of a metalens to increase the forward momentum has been proposed in literature. Here, a simple theoretical analysis is done to give a simple formula for the optical pulling force (OPF) generated by a one-dimensional metalens. From there two parameters are determined to be limiting for the OPF generated by a metalens: the transmissivity and the numerical aperture of the lens. A large nu- merical aperture and a large transmissivity of the lens appear to be critical in obtaining an apparent optical pulling force. Unlike in traditional optical trapping, where the object must be confined at the focal spot and is therefore limited in size by diffraction and high local intensities, the metalens considered here is not placed at the focus itself. This configuration relaxes the geometric and diffraction-related constraints of conventional trapping and shifts the difficulty from maintaining a tightly confined focus to engineering a lens that can efficiently redirect momentum over a large angular range. A workflow to design one-dimensional slit-based metalenses is proposed taking into consideration the results from the theoretical analysis. The optimized metalens design achieves an accurate focal point, exhibiting a deviation of only 2.88% from the targeted focal length. Which is a competitive error compared to literature. The analytical formula predicts that optimized lenses can generate a linear force on the order of 10−12 N m/W, in good agreement with values reported in literature, thereby supporting the validity of the analytical model.

In modern-day research, magnetometry provides valuable information for a wide range of studies. Among all the different forms of magnetometers, the nitrogenvacancy (NV) lattice defect in diamond has emerged as a powerful magnetic field sensor thanks to the combination of sensitivity, spatial resolution and versatile capabilities. High-fidelity microwave control and optical readout of the NV spin over a wide range of conditions has enabled applications in condensed matter physics, chemistry, biology, geoscience and many more. In particular, its capability of visualizing magnetic phenomena with high spatial resolution has proven to be a powerful tool in both fundamental physics and applied sciences. Advances in NV magnetometry in the past decade have led to numerous breakthroughs, especially in revealing the nanoscale physics of condensed matter systems. However, the free-space optics generally used for optical interrogation of the NV spins are challenging to realize in cryogenic, intra-cellular, or other hard-to-reach environments. As such, realizing robust all-fiber-based NV probes with efficient optical readout could enable new measurements in low-temperature (quantum) or biological systems...

This thesis provides an overview of research focused on fabricating high-performance nanomechanical resonators from amorphous silicon carbide (a-SiC) and (super)conducting metallic niobium titanium nitride (NbTiN), and subsequently characterizing the superconducting NbTiN resonators using scanning tunneling microscopy (STM). The installation of on-chip nano mechanics with a minimally invasive STM detection technique enables the probing of subtle variations in the Casimir force between superconductors during their phase transition. This thesis consists of four parts.

With the aim of maximizing the coupling of the Casimir force to a large superconducting nanomembrane suspended over a sub-micron vacuum gap, we initially employed atomic layer deposition (ALD). Prior to using ALDto construct the high-aspect-ratio superconducting cavity, we investigated a novel amorphous silicon carbide (a-SiC) material in Chapter 2. Our study demonstrated that a-SiC exhibits high chemical inertness, remarkable ultimate tensile strength, and—most importantly—the capability to support nanomechanical resonators with high quality factors. Leveraging these excellent properties, we fabricated high-aspect-ratio a-SiC nanomembranes suspended over a sub-micron vacuum gap.

Subsequently, using the on-chip cavity formed by the strained a-SiC nanomembrane and the substrate with flat surface, we performed ALD to conformally coat all cavity surfaces with metallic NbTiN, thereby filling the vacuum gap atomically layer-by-layer. By optimizing the deposition conditions with the method described in Chapter 3, we achieved a high-aspect-ratio NbTiN cavity with a gap size of less than 100 nm. During this optimization, we also observed that metallic nanomechanical resonators fabricated via ALD can operate at room temperature with quality factors significantly higher than those of fully coated metallic resonators produced by other deposition techniques.

To measure the superconducting nanomembranes with minimal perturbation to their superconducting state, we installed the NbTiN nanomembrane fabricated by ALD into a scanning tunneling microscope (STM) and performed dynamic measurements in a cryogenic environment, as detailed in Chapter 4. In studying the tip–membrane interaction, we developed three measurement techniques to precisely determine the resonant frequency of the nanomembrane. One technique relies on the homodyne method, while the other two exploit the Van der Waals interaction between the STM tip and the nanomembrane.

Although the NbTiN nanomechanical resonators exhibit high performance, their superconducting properties are compromised by the absence of plasma bombardment on the inner cavity surfaces. Drawing on our experience with suspending large nanomembranes over small gap sizes in Chapter 3, we developed a new fabrication process in Chapter 5 to suspend a bare superconducting NbTiN nanomechanical membrane over another superconducting NbTiN film, with a sub-micron vacuum gap between them.

This configuration minimizes unwanted interactions that could otherwise affect the precision of measuring the variation in the Casimir force between superconductors during the phase transition. The high force sensitivity of the nanomembrane allows us to detect subtle force changes associated with the superconducting phase transition, including those arising from abrupt changes in the Casimir force.

The current frontier of human exploration is space. We have achieved sending rovers to planets and probes throughout the Solar System and beyond into interstellar space, yet the stars seemed to be out of reach. The current fastest man-made object would take thousands of years to travel to the nearest star, Proxima Centauri. However, Breakthrough Starshot presents an engineering challenge to reach it in a generation's time. Recent technological advancements make this possible through ultralight spacecraft equipped with mirror-like sails, known as Lightsails. Due to their structure, these sails could be propelled by lasers up to 20% light speed, or over 200 million km/h.
To achieve the required efficiency, the sails need to exhibit high reflectivity and low mass on the scale of 1g, while covering a surface area of ∼10m2. Some materials and membranes explored in literature have exhibited high reflectivity, but their mass did not comply with the necessary requirements. The most promising current design is a photonic crystal, which interacts with light in a way that maximizes reflectivity and removes material by incorporating cavities into the surface, effectively reducing overall mass.
Current research into photonic crystal Lightsails primarily focuses on the reflectivity of a small segment of a flat lattice. While this is essential, the literature suggests that the full-scale structures will probably behave like traditional sails on sailboats, tending to billow. This prompts us to consider not only the problem of reflectivity on curved surfaces but also mechanical deformations and stresses within a membrane that's 1000 times thinner than a human hair.
This thesis investigates the full-scale design of the Lightsail from both the structural and electromagnetic sides. Firstly, the material, microstructure, and macrostructure of the Lightsail are analyzed to determine stresses and deformations, as well as the optimal shape for stress distribution and efficiency. Then, through topology optimization of shell elements, the main load-carrying "backbone" of the sail is identified. Secondly, optimization of unit cell photonic crystal cavities is performed across the curvature of the sail derived from the previously obtained shape, maximizing the reflectivity. Finally, the thesis proposes a design of the Lightsail that integrates all findings, which is then used to obtain a Figure of Merit value for comparison with designs in existing literature.

This paper presents the design of a lab-scale silicon nitride photonic crystal lightsail, demonstrating the LightSailSim software package. LightSailSim, an open-source analysis pipeline, integrates a custom mechanical solver based on a particle system model with optical simulation results. The resulting design balances dynamic and structural stability with propulsive performance. It consists of an optimised photonic crystal that provides a suitable trade-off between stability and propulsive force. The sail is suspended at its edges by a circular ring made of silicon, providing sufficient boundary tension to prevent sail deformation-induced instability. The design is made such that it can be produced from a single silicon wafer and levitated using a single 400 W laser.

Artificial intelligence has a strong need for faster and more energy-efficient solutions, especially for computation performed at the sensor edge. On-chip photonic neural networks (PNNs) offer a promising solution for high speeds and energy efficiency. A less explored side of PNNs is their application to time-series data, which is often the case for real-world sensor applications. While PNNs promise high speeds and energy-efficient solutions, no good use cases have been proposed. This report will first review the state of the art of PNNs. It will be seen that to solve time-series tasks, photonic continuous-time recurrent neural networks (CTRNNs) are required. The dynamics of CTRNNs are thoroughly explored to leverage obtained insights to recommend novel practical applications. This is done through simple examples, and applied to a real machine learning task. It was seen that on a real classification task the network learned two distinct fixed points corresponding to the classes. A link between the time constant of the continuous-time neurons and the temporal dynamics of the task is also found. Two general directions for novel applications are then proposed. Firstly, photonic PNNs can be slowed down to match the task. Opto-electronic PNNs allow for more control of the time constant, and on-chip photonic filter neurons are suggested. Secondly, the high speeds of photonic neural networks can also be directly leveraged. Extremely fast convergence of photonic CTRNNs can be utilized for Modern Hopfield networks, pathfinding algorithms, and time-dependent optimization problems such as for example Model Predictive Control.

Electron energy-loss spectroscopy (EELS) is a powerful analytical technique used in transmission electron microscopy (TEM) to investigate the energy loss of electrons as they interact with a specimen. EELS provides valuable information about the electronic structure, composition, and bonding properties of materials at the nanoscale. Despite its tremendous potential, EELS techniques often face challenges related to spectral resolution and signal-to-noise ratio. In this context, the present project aims to significantly enhance the EELS spectra resolution by innovatively designing and fabricating a specialized Transmission Electron Microscope (TEM) holder. By leveraging advances in photonics and micro-resonator technology, this project seeks to revolutionize the quality and accuracy of EELS measurements, opening new avenues for high-resolution material characterization. The motivation behind this project lies in the critical need to overcome existing limitations in EELS techniques and push the boundaries of nanoscale material analysis. Traditional EELS setups, while offering valuable insights, often suffer from challenges associated with background noise, limited energy resolution, and compromised signal quality due to multiple scattering events. These factors hinder the ability to extract precise information about material properties, impeding progress in various scientific and technological fields. Motivated by these challenges, the central aim of this project is to design and develop a TEM holder that incorporates cutting-edge photonic micro-resonator technology. This combi- nation is expected to improve EELS abilities. The proposed holder will enable controlled interactions between the electron beam and the photonic micro-resonator, exploiting quantum optical phenom- ena to enhance the energy-loss signal and reduce unwanted noise. This approach aligns with recent advancements in Photon Induced Near-field Electron Microscopy (PINEM), quantum optics and electron-photon interactions, paving the way for unprecedented sensitivity and resolution in EELS measurements. The outcomes of this project hold immense promise for diverse scientific disciplines. In materials science, researchers will gain deeper insights into the electronic behavior and properties of nanoma- terials, facilitating the design and optimization of advanced materials with tailored functionalities. Additionally, the improved EELS resolution will impact fields such as catalysis, nanoelectronics, and biological imaging, where precise characterization at the nanoscale is essential. To achieve our objectives, we will address the following key research questions: • What are the optimal design parameters for the micro-resonator, facilitating efficient electronphoton interactions and maximal photon generation per electron? • What strategies can be employed to mitigate potential sources of photon loss within the integrated TEM holder, such as losses in coupling, transmission, and detection, thereby maximizing the efficiency of photon collection and measurement? • How do the characteristics of the micro-resonator, such as its size, shape, and material properties, impact the generation and propagation of cavity photons, and how can these parameters be tailored for optimal EELS performance? We expect that the results of this project will revolutionize the field of material characterization by enabling unprecedented high-resolution EELS measurements, poised to impact diverse scientific and technological domains.

This research explores the integration of Neural Architecture Search with Optical Neural Networks to optimize the efficiency and performance in traditional visual image classification tasks. The study introduces a new approach that applies Neural Architecture Search, a technique traditionally used to optimize the performance of Artificial Neural Networks, to the field of Optical Neural Networks.

Research has shown concepts of phononic integrated circuits (PnICs) consisting of multiple acoustic waveguides and beam splitters. Traveling acoustic waves experience energy loss when moving. Minimizing these losses is critical for large and complex PnICs. By fabricating on-chip PnICs, it is possible to operate under vacuum and at low temperatures resulting in lower acoustic losses.
Studies have shown on-chip data transport using straight acoustic waveguides fabricated inside suspended membrane phononic crystals. The use of high stress Silicon Nitride (SiN) has enabled
the manufacturing of high aspect ratio suspended structures. This is achieved by stress keeping the membrane under tension resulting in a flat surface.
Current acoustic beam splitter designs feature liquid and air for the wave supporting material. Making them not suitable for on-chip devices. Therefore, on-chip PnICs require a new type of splitter
design.

In this study, finite element methods are used to investigate if suspended SiN membranes offer a solution for on-chip acoustic beam splitting. The wave confinement of square and hexagonal lattice splitters are also compared to find the best performing design. Phononic bandstructures are constructed by performing eigenfrequency studies on various 2D phononic crystal unit cells. These bandstructures reach full GHz bandgaps with relative bandgap sizes of 57.3%.
Symmetric beam splitters are then made by introducing line defects into the crystal structures. Multiple wave excitation locations, frequencies and waveguide widths are investigated to establish the highest wave confining splitter design. The final design consists of a y-shaped splitter inside a suspended SiN hexagonal lattice phononic crystal and achieves a wave confinement of 85.03% at an excitation frequency of 3.25GHz.
The results show that suspended SiN membranes featuring hexagonal lattice phononic crystals offer a promising solution for on-chip beam splitting.

Due to the demand for renewable energies, gasses likeH2 need to be detected with sensitive, accurate and fast detectors. The US Department of Energy has made a list of requirements that H2 sensors need to fullfill of which the minimum detectable concentration (0.1%) and the reaction time (one second) are challenging for current hydrogen sensors, particularly for low-power and low-cost mass-produced sensors. This thesis covers the design and simulation of a photonic crystal nanobeam cavity sensor, using the Pt-catalyzed WO3 as a H2 sensitive material. The thicknesses of both layers show a trade-off between the minimally detectable concentration and the reaction times, resulting in a cavity with sensitivity of almost 40 nm/RIU and a Q factor varying between 5e5 and 5e4 depending on how much catalyst is required to meet the one second reaction time performance target. Based on the conservative assumptions regarding the reaction times of the sensor, the 0.1% performance target is almost achieved. This means that if in practice, any of the layer thicknesses show to be more favourable than assumed in this thesis, together with the fact that for optical sensors there is no risk of sparks, the photonic crystal nanobeam cavity as a H2 sensor shows to be a good fit for a high performance, low-cost, mass-produced H2 sensor that meets the DoE performance targets.

Nanophotonics is the study of structures’ interaction with light with features at or below the nanometer scale. It has gained the interest of many researchers, as it can be used to control the flow of light very effectively in the design of, e.g., solar cells, highly efficient biosensors or lasers. The design of such devices can be non-intuitive and complex and therefore computational tools like topology optimization techniques have been used to improve their designs. However, the topology optimization methods used in the literature often use a density-based representation of the geometry, which often leads to jagged edges. It has been shown in the literature that jagged edges can deteriorate the accuracy of simulation results. Using a level-set method in combination with an enriched finite element method offers a smoother boundary representation than the often used density-based methods. This work aims to develop an analysis and level set optimization for 2D electromagnetic scattering and eigenvalue problems using an enriched finite element method. Furthermore, we showcase that even for a non-conforming discretization, the enriched finite element method achieves the same convergence properties as the standard finite element method with fitted meshes. Finally, we perform topology optimization on the design of both a 2D meta lens and 2D reflector, maximizing their ability to focus light onto a point, using a level set method to define the geometry in combination with the enriched method used in the analysis.

Suspended microchannel resonators are a kind of mechanical resonator with an embedded microchannel inside. This is done as for some samples it is convenient or even necessary to put the sample into a fluid. The sample can then travel through the channel and its mass can be determined. The channel also simplifies placement of the samples compared to regular nanomechanical resonators. By having this channel embedded within the resonator, it is then possible to place the resonator as a whole in a vacuum and so eliminate medium losses. The main purpose of these suspended microchannel resonators is mass spectrometry which has already been performed on viruses and biological cells and so could potentially be very useful for the medical field as well. Especially if suspended microchannel resonators can be used to measure individual proteins which has not been done before with these resonators. It would also be interesting to use these kinds of resonators in quantum mechanical experiments. This means that the performance of these suspended microchannel resonators needs to be improved.

In order to accomplish this, a new kind of suspended microchannel resonator is introduced. The determination of these designs is strictly FEM based. The new resonator design, which is a doubly clamped beam, is made from pre-stressed silicon nitride which enables high Q-factors caused by an effect known as dissipation dilution. Dissipation dilution means diluting the energy losses of the system by increasing the stored energy which is increased by the initial stress in the silicon nitride. To increase this effect, the resonator is tapered towards the center. This is called strain engineering. Silicon nitride has been used in suspended microchannel resonators before but it did not result in higher Q-factors. Possibly due to clamping losses. The designs presented here offer a solution to that in the form of soft-clamping. This reduces the curvature near the boundaries to a minimum which means that the clamping losses are also reduced or eliminated. The combined effect of these features results in very high Q-factors as has already been reported for nanomechanical resonators. Through machine learning optimization it is possible to achieve 6E12 Hz in Qf product required for quantum mechanical experiments at room temperature and more than one order of magnitude improvement in mass sensitivity.