The miniaturization of electronic devices has driven modern advances in computing, consumer electronics and sensor technology. Digital integrated circuits and the transistor are the most widely produced microelectronics. The metal-oxide-semiconductor field effect transistor (MOSFET) is widely used in modern devices. The core of this device is the metal-oxide-semiconductor (MOS) capacitor. While the physics behind the use of MOS capacitors are well understood, real world fabrication provides unique challenges that require experimental work and process characterization to repeatedly produce viable devices. For this project, we fabricated microscale MOS capacitors on an n-type silicon substrate. The dimensions and properties of the dielectric material used in a MOS capacitor determines its electrical properties. We characterized the effects of variations in the processing method for the deposition of the silicon oxide layer.  We varied parameters such as deposition rate, chamber temperature and silicon oxide thin film thickness.  Current-voltage (I-V) characteristics and electronic properties such as capacitance and breakdown voltage were characterized with electrical probing. We hope to quantitatively evaluate relationships between processing method and the electrical performance of MOS capacitors fabricated using tools and materials available in the Washington Nanofabrication Facility (WNF). Characterization of these relationships assists users and researchers in their circuit design and choice of processing parameters.

Flash memory, the predominant long-term storage device, is rapidly approaching its scaling limit, and among next-generation memory devices, phase change memory is a promising alternative. Phase change materials such as germanium telluride store information through electrical differences between crystalline ordered and amorphous disordered states. Since the speed and power consumption of a phase change device is directly related to the melting point of the material, we seek to find methods to reduce the melting point of germanium telluride. In particular, it has been hypothesized that small doping concentrations of bismuth could greatly reduce the melting point of germanium telluride. To investigate the effects of these dopants on germanium telluride's melting point and electrical properties, we utilize ab-initio modelling techniques to simulate these phase change devices on the nanoscale level. These simulations are performed using the modelling software Siesta where we can examine the impact of small changes in parameters such as doping concentration, temperature, and device geometry. Already, we have found that bismuth doping at high concentrations causes the germanium telluride to display metallic properties. We are still hoping to explore lower doping concentrations and the effects of bismuth on melting point. 

As magnetohydrodynamic (MHD) systems exhibit incredible nonlinearity, computational models have become increasingly common for parsing difficult engineering problems in plasma experiments. The dynamic mode decomposition (DMD) is a principled, data-driven, equation-free approach to extracting coherent temporal-spatial structures from raw data, with particular success in hydrodynamics and fluid-flow problems. We consider the DMD applied to MHD systems in the particular example of spheromak plasmas simulated in the Helicity Injected Torus experiment. We find that the DMD successfully resolves magnetic field structures in simulated magnetic field measurements that closely match expected physical properties. We further find that, under certain constraints, the DMD provides a powerful tool for further validation of simulation to experiment. Given the nonlinearity of magnetohydrodynamic systems, existing mechanisms for feedback controlling plasmas as instabilities occur are limited. Because the dynamic mode decomposition effectively serves as an approximate, linear reduced order model built from experimental and simulated magnetic field measurements on the fly, it offers considerable promise as a predictive feedback controller. This is significant as the detection of and control over plasma instabilities remains a major barrier to the use of spheromaks for fusion energy.

The security level of the interactions between human users and cyber systems is becoming more and more important in many areas. The traditional way of identifying a user is to utilize alphanumerical passwords, which are easy to realize but hard to memorize and vulnerable to various kinds of attack. To enhance security performance for user authentication, we take the advantage of the haptic interaction of new touchscreen devices that have recently been introduced. Previous studies show that how a person interacts with a haptic interface is unique and can be used for identification and authentication purposes. We applied this technique using the Microsoft Surface Pro 3, which has a pen, the N-trig digitizer. Having its API (ways for developers to talk to another site or services) made it easier to transfer data to the haptic password system. By developing an app on a writable windows tablet, we successfully captured ink data and pointer properties such as pressure, twist and orientation. We built a haptic password system and invited subjects to take specific tests. After data analysis, we found that this system is secure with good performance and forgery attack resistance. This system provides a complex but intuitive way of creating signature passwords. This haptic interaction technique can be applied to those pressure sensitive devices, such as tablets and signature touch screens. This could provide a much safer way of authorizing credit card transactions and lead to reduced transaction fraud.
 

In the United States alone, over 3 million people have glaucoma, yet only half of those people have been diagnosed. Elevated intraocular pressure is the principal symptom checked to detect glaucoma. Current medical technology lacks the capability to capture the fluctuation of eye pressure during the day, essential for an accurate diagnosis and management of the disease. Implanted during cataract surgery, this intraocular pressure sensor increases monitoring capability, allowing early detection and diagnosis of glaucoma before irreversible nerve damage occurs. Our device is a passive capacitive sensor; no chip or battery is needed to detect the changing of pressure within the eye. The signal is sent and transmitted to an external readout via RF. The sensor is fabricated using microelectronic techniques and is made of biocompatible materials, such as polyimide and gold. Two parallel gold rings plus the empty cavity created between them form the capacitive sensor. The device itself is covered and packaged in polyimide. In relation to device fabrication, my research involved the optimization of various processes related with the production and testing of the intraocular pressure sensors. I performed the optimization of polyimide deposition parameters, as well as the timing and process flow of baking and curing steps. I also performed the testing of varying gold thicknesses, deposited via several different processes. Two layers of polyimide are bonded together using heat and pressure to form the cavity. The heat and pressure used to fuse the two polymer layers morphed the shape of the cavity, resulting in inconsistent signals given off by the antenna. We now seek to develop a process flow that does not necessitate the bonding of two separate layers to form a cavity.

An electron bound to a donor impurity in a semiconductor exhibits properties of a quantum system that could be used as a spin quantum bit (qubit) for quantum computing. For such a quantum system to be considered a candidate for a spin qubit there must be minimal interaction with its surrounding environment, maximizing the time the system can store quantum information. Longitudinal electron spin relaxation time (T1), the characteristic time for perturbed electron spins to return to thermal equilibrium, serves as a fundamental upper limit for quantum information storage time of an electron spin qubit. Theoretical calculations and previous electron spin resonance measurements of donor bound electrons in bulk ZnO provide promising predictions for long storage times of ZnO donor electrons. To date there has been no direct optical measurements of T1 for donor bound electrons in ZnO. Here I present the experimental method and results of measuring T1 for electrons bound to gallium and indium donor impurities in ZnO nanowires. A measurement of T1 on the order of milliseconds for donor bound electrons in ZnO would merit further research to determine if these donor electrons can be realized as qubits with good properties. The nanowire structure of our ZnO samples also has the potential to meet another requirement for the donor bound electrons to be considered candidates for spin qubits; they must have qubit-specific measurement capabilities. By isolating a single nanowire, with the right sample dopant concentrations, we would be able to spatially resolve a single qubit.

This project aims to build a pulsed laser for lifetime measurements of nitrogen-vacancy (NV) centers in diamond. Such measurements are an important tool for characterizing the quantum information processing potential of NV-center-based photonic devices. Specifically, the radiative lifetime of an NV center is indicative of the proportion of photons being emitted at the useful zero phonon line (ZPL) wavelength. Optical emission at other wavelengths cannot be used for quantum information processing. As NV centers naturally emit only ~3% of photons at the ZPL wavelength, a major objective of our lab is to increase this quantity through photonic integration. Although lifetime measurements have been performed previously in our lab, the excitation source used had a pulse width of 4 ps, much shorter than needed and thus requiring a high optical intensity. This tended to damage the photonic devices under investigation. Therefore, a pulsed laser with a pulse width on the order of 10 ns is being constructed for future lifetime measurements. An important feature of this laser will be a pulse fall time significantly shorter than the NV center’s radiative lifetime. After the laser becomes operational, we will utilize it to implement lifetime measurements of NV centers coupled to photonic devices. This will lead to a better characterization of device performance, and ultimately to large scale NV-center-based quantum information processing.

Point-of-care diagnostic tests allow clinicians or care providers to make rapid, informed treatment decisions in settings where patient follow-up is unlikely or not an option (i.e. rural regions in developing countries). The Anisotropic Ratchet Conveyors (ARC) platform developed in our laboratory can transport small, individual droplets with a passive surface pattern and orthogonal vibration. In this work, a novel ARC system is characterized for future use in automated point-of-care diagnostic devices. Effects of rung radius on droplet transport were characterized with respect to the acceleration amplitude of the vibration required to initiate transport. The acceleration amplitude was significantly lower on the 1000 µm radius and 1500 µm radius rungs compared to the 500 µm radius rungs and slightly lower on the 1000 µm radius rungs compared to the 1500 µm radius rungs. Interestingly, the radius of the rungs was also observed to influence the natural frequency of the transported droplets. The results of this work will be applied to optimize droplet transport across functionalized detection regions on chip. Thus, ongoing work aims to functionalize the silicon dioxide rungs for immunosensing. In order to demonstrate proof-of-concept immunosensing in droplets, antibodies (e.g. antibodies targeting human immunoglobulin gamma - anti-IgG) will be bound to the silicon dioxide rungs using a linking molecule (e.g. silane-PEG-NHS or APTES) via micro contact printing, a common technique used to transfer proteins to a surface with elastomeric stamps. This process will be optimized to control the density of bound antibodies. Fluorescent microscopy will be used to verify and characterize the capture of fluorescently tagged target molecules (e.g. FITC-IgG) suspended in droplets transported on ARC tracks. Experiments will also be performed to characterize how antibody capture affects the velocity droplets transported on rungs of each radius group. The ability to capture target molecules suspended in droplets will present the potential of using ARC devices as a platform for rapid, efficient healthcare diagnostics.

In collaboration with Booz Allen Hamilton, our E E 497/498 Entrepreneurial Design Capstone project addresses the challenge of taking precisely georeferenced photographs of the underwater domain. These photographs will be mapped onto existing bathymetric datasets (from sources like NASA and NOAA) to provide visual textures of static and dynamic underwater points of interest. Our team’s role is the design and implementation of the imaging system hardware, which will complement visualization software under development at Booz Allen Hamilton’s Seattle office. Our design extends the OpenROV 2.8 platform, a small open-source, tethered, remotely-operated vehicle (ROV), with emphasis on developing a successful prototype using commercial off-the-shelf (COTS) parts at minimal cost. In total, our team has implemented (1) the extension of OpenROV into an underwater imaging platform, carrying a payload enabling hemispherical imaging below the ROV; (2) the design of a tethered surface buoy system, with access to Differential GPS data to act as a reference for geolocating the ROV; (3) the design of an ROV-mounted acoustic transmitter and buoy-mounted acoustic receiver array to determine underwater distance between the ROV and each array element; (4) the design of an efficient trilateration algorithm using nonlinear and linear least-squares optimization to accurately position the ROV relative to the buoy based on (3). Our research has significant implications for improving access to the underwater domain, including low-cost maritime sensing and deployment applications and military training and simulation applications.

3D integration of microelectronic structures are essential to addressing the demands of future high performance electronics. Current memory chips use wire bonding to connect memory cells but are reaching a limit due to assembly challenges. Through-silicon-vias (TSVs), which are vertical electrical connections through a silicon wafer, are a promising alternative to wire bonding and printed electrical interconnects. 3D integrated circuit architecture can be fabricated using TSVs to connect stacked memory cells to increase memory performance while reducing power consumption and chip area. The key to improved TSV technology relies on the conformity of the TaN diffusion barrier and dielectric layers (inside the TSV) prior to electroplating the Cu interconnects. In addition, longer and smaller TSVs form denser 3D microelectronic structures which could have major implications in maximizing data flow for memory cells. In this work, we focused primarily on fabricating ultra-high aspect ratio TSVs (20:1) through standard semiconductor fabrication techniques and used deep reactive-ion etching to create the TSVs. We addressed TSV fabrication process challenges, such as achieving conformal high aspect ratio metal interconnects, by utilizing a novel thin film deposition process called pulsed-CVD (chemical vapor deposition). Lastly, preliminary research on the annealing of the TaN barrier layer reveals a more uniform grain microstructure, which attributes to the TaN’s ability to resist ion diffusion.

The nitrogen-vacancy (NV) center is a defect in diamond which shows promise as a solid-state qubit for scalable quantum computation. However, inhomogeneity and instability in the frequency of NV optical emission inhibits efficient quantum entanglement generation between NV centers, thus preventing the realization of quantum algorithms. This work implements two methods—one passive, one active—to mitigate these effects. First, high-temperature annealing improves the uniformity of the carbon lattice, which stabilizes the electro-magnetic environment of NV centers. Specifically, detrimental defects occupying nearby lattice-points diffuse out of the crystal. We build and test a system capable of non-destructively annealing diamond to 1200°C. Second, the application of strong local electric fields across NV centers enables active frequency control via the Stark Effect. To apply these fields, we fabricate gold electrodes directly on the surface of diamond containing near-surface NV centers. To determine the effectiveness of this method, we monitor the frequency of NV emission with photoluminescence excitation spectroscopy while fluctuating the applied electric field. Future work will involve the integration of these two systems with previously-demonstrated photonic circuits to realize efficient multi-qubit entanglement generation.

Copper metallization is an essential step to provide low electrical resistance and reduced electromigration for interconnecting electronic devices. Recently, 2.5D and 3D chip integration has emerged as a desirable packaging technology for advanced electronics. 2.5D uses dedicated silicon-based circuit boards with precise, thick conductor lines while 3D uses vertical connections call Through Silicon Vias (TSVs). Both rely on the process of advanced electroplating where copper is deposited onto the surface of a wafer to fill high aspect ratio structures (either trenches or TSVs). In order to have level plating on a wafer and minimize pinching off of the features, specific levels of organic additives are necessary in the solution. These additives include accelerator, suppressor and leveler, all of which contribute to a successful electroplating process. The analysis of the levels of the additives are discussed as well as their depletion over time due to evaporation or usage. Producing uniform electroplated Cu is dependent on specific concentrations of organic additives in the Cu electroplating bath, namely the accelerator, suppressor and leveler. In this work, we performed copper bath analysis to characterize the amount of the additives needed to achieve an optimal plating bath composition to yield pristine Cu thin films. During the electroplating process, we addressed key challenges by adjusting variables such as increasing current in small intervals to prevent contact failure and taking into account the high aspect ratio structures and plating times to completely fill the trenches/TSVs.

A sizable obstacle to the success of wearable devices in supporting good health, independent living, and workplace safety is compliance: is the user wearing the device and wearing it as it was designed? According to the United States Department of Labor, thousands of people are blinded each year from work-related eye injuries; this costs the industry more than $300 million in lost production time, medical expenses, and worker compensation (“Safety and Health Topics | Eye and Face Protection,” n.d.). This loss easily could be prevented by the usage of protective eyewear; thus is it imperative to know whether workers are complying with safety standards. Occupational Safety and Health Administration (OSHA) requires that all protective equipment, including personal protective equipment for eyes, face, head and extremities be provided (“General requirements. - 1910.132,” n.d.). Therefore, any manual laborer worker is required comply with safety standards, including wearing protective eyewear. This research investigated inexpensive sensors to monitor compliance with regard to safety goggles and workplace safety. To ensure high fidelity and accurate data, it is paramount to understand user behavior in the wearing and use of these devices. Specifically, detailed knowledge regarding the subject’s usage of protective eyewear, including the orientation and location on the head directly to inform the quality of compliance to safety standards. However, the addition of compliance monitoring functionality is limited by size, power consumption, cost, and form factor constraints inherent to the wearable device of interest. These limitations restrict the types of sensors that can be used to achieve compliance monitoring. After literature review and feasibility study of compliance monitoring techniques and existing biomedical wearables, including assessing electrostatic field sensing, magnetic sensing, and various approaches to eye tracking, we have decided to pursue magnetic sensors as the most viable option for compliance sensing. We are optimistic that magnetic sensing may offer a low cost and low energy solution for the detection of direction and orientation of protective eyewear. Therefore, with magnetic sensing we will determine if the protective eyewear is being worn on the bridge of the nose, on top of the head, or completely off of the body to determine compliance with safety standards. A preliminary design has been modeled and simulated to be evaluated for feasibility.

Only a small portion of the market sells printers that are compatible with Google cloud print. In the case where a customer purchases a printer that is not compatible with Google cloud print, the printer can only be used through a wireless, wired, or USB connection. With the aim of allowing users to print from anywhere and at anytime, our project intends to design a low-cost and user friendly device that is capable of converting a regular printer into a cloud printer. The project is being developed on a small portable Linux machine (raspberry pi) using a 16GB SD card and a small Wi-Fi adapter. The raspberry pi is initially configured as an access point, enabling the user to connect wirelessly to the device. Once connected, the user then navigates to a web page hosted on the device accessing a user friendly interface to set up the printer. After the initial setup has been completed, the raspberry pi will then turn into a Wi-Fi client which allows itself to receive instructions through the Internet. The device will then interface with the printer through a USB cable. The users then can log in to their Google cloud print account on any device from any place at any time to print on the registered printer. This device, once developed, would provide a low cost and user friendly alternative to the existing expensive cloud-ready printers and allow users to take advantage of Google cloud print.

Real-time quantification of chemical compounds in urban and non-urban environments has far-reaching prospective benefits for our society. Such a possibility could provide instantaneous detection of dangerous chemicals, identification of areas that require air quality restoration, estimation of agricultural effectivity, etc. However, currently, dynamic and efficient platforms to perform quantification do not exist and only a few static sensors are deployed in sparse locations, e.g., a typical city has only a handful of such sensors. While several mobile devices have been constructed, they are typically large, expensive and imprecise. These factors make them practically unrealizable for gathering environmental data in real-time from a wide variety of locations. In our project, we address this fascinating and challenging problem by building an inexpensive, mobile multi-sensor device, that communicates with a mobile phone. We designed the device to incorporate an array of inexpensive sensors, many of which have cross sensitivities to particles measured by other sensors. We had performed testing on a variety of sensors to identify most efficient and precise sensors. To improve performance of the device and obtain readings, we have implemented a neural network, based on olfactory system in moths. We have assembled a prototype of a multi-sensor device according to our design. Our preliminary results show that the network was capable of discriminating and tracking more precise values of contaminants, that could even include mixtures of compounds. We are currently working on implementing a server-side application that will integrate information from multiple devices and will produce a real-time “heat-map” of contaminants and pollution levels across a geographic area.

The question of how to better represent harmonic and non-harmonic sounds with cochlear implants was addressed by reviewing journal articles on the subject and experimenting by simulation. By using electrodes to broadly stimulate the auditory nerves, cochlear implants can enable profoundly deaf individuals to understand a conversation in a quiet room. Because of this broad stimulation however, cochlear implant (CI) users have difficulty discerning pitch. The ability to discern pitch is not only nice for listening to music and necessary for understanding tonal languages like Chinese, but also has been shown to be important for aiding in the separation and understanding of speech in noisy environments. New processing strategies are being developed to represent pitch that rely on representing its fundamental frequency f_0 by amplitude modulating the carrier pulse trains sent to each electrode. This approach has improved users’ performance in timbre tests, melody recognition tests and Chinese tone recognition tasks but there is still much room for experiment and refinement. Since the harmonic components of pitched sound are evenly spaced in the frequency domain, two pitched sounds having fundamental frequencies related by a simple ratio will have overlapping harmonic components and be perceived as more consonant while notes with more complex relationships share fewer components and are perceived as more dissonant. Previous studies on normal hearing listeners as young as 9-months old have shown that it takes longer for subjects to notice and respond to a change in note sequences when the sequences switch from dissonant to consonant relationships than from consonant to dissonant ones. A better understanding of perceptual consonance and dissonance for CI users would aid in the development of better processing strategies and simulations. Experiments are proposed to investigate this perception by studying the reaction times of subjects to changes between electrode stimulation patterns.