Bismuth nanoparticle (Bi NPs) have wide-ranging applications in both biomedical imaging and nanotechnology. These applications are hampered by the absence of a sized-controlled, uniform synthesis using conventional laboratory techniques. The current synthesis of such particles is arduous and requires specialized techniques and equipment, specifically air-free techniques, which are not readily available to all. By using air stable and cost effective reactants, we hope to develop and characterize a synthesis that is reliable, aerobic, free of lithium based organic bases, and potentially offers size control between 30 nm and 70 nm. This synthesis would produce high quality Bi NPs that are available to a wider range of researchers.

This research aims to understand how Commute Trip Reduction (CTR) transit subsidy programs, when controlling for various built environment variables and the structure of the transit network, impact employee transit utilization commuting to the worksites of large employers in the Central Puget Sound region. This transit utilization is measured using the ORCA fare card records over two nine week periods in 2015 and 2016 of the major transit agencies in the Central Puget Sound. Manipulating monetary costs is a known method of transportation demand management. Earlier research on preliminary data has suggested that these transit subsidies do have a significant impact on transit utilization. However, results in the wider literature suggest that transit utilization was operationalized in a way—defined on the level of an individual card without accounting for the existence of people who never take transit—that may have altered the significance of the control variables. Indeed, some of the results were counterintuitive. In this research I attempt to avoid this by focusing solely on trips of employees of large employers to and from their worksites. This allows me to incorporate data on how many employees are not utilizing transit. Specifically, I create a regression tree model that predicts the number of trips taken to and from a worksite on an individual card. The features include subsidy values associated with each card, the closeness centrality of the stops around worksites weighted for travel time and headways, and built environment variables measured around each worksite. The built environment variables measured include, among other things, density, free parking price, demographics, and the design of the street network. I expect that higher centrality of worksites and higher pass subsidies will both increase transit utilization.

The University of Washington CubeSat program needs a space environment simulator to test their satellite. This was done by simulating the magnetic field and light environments of low-earth orbit. We have accomplished this through the construction of a 3-axis Helmholtz cage and sun simulator. The cage is used as a testing rig for a small satellite's, DubSat-1, attitude determination and control systems. The cage is large enough to fit a 3U cubesat (30cmx10cmx10cm) inside and fully cancel Earth's magnetic field, in addition to generating an arbitrary magnetic field vector of the magnitude expected to be found in low-earth orbit. The sun simulator generates a sun vector over the entire sphere within 10 degrees of accuracy using a strip of LEDs. The LED strip is mounted on the cage and a motor is used to rotate the strip around the cage to change the angle of the light. The Helmholtz cage is controlled using feedback from a magnetometer placed inside the magnetic field of the cage. Based on the data from the magnetometer a control loop is used to control the current through the coils to adjust the magnetic field strength to its desired value.

For our research, we are focusing on dice and how we can relate them to voting systems. We want to see if we can create dice systems that mimic voting systems (with >2 candidates) and the traits that the systems possess. Two ideal traits of these voting systems are Pareto, which means that if everyone likes candidate A more than B, candidate B should lose; and IIA, which is if candidate A wins, then everyone mixes their ballot but keeps A's rank relative to B the same, then A should still win. However there is Arrow's Theorem, a central focus of our research, which states that if your system is Pareto and IIA, you must live in a dictatorship. One major topic of our research is trying to show whether or not this holds for our die systems. In order to accomplish this, we set out to define what these traits would be in terms of dice and how to translate dice outcomes to voter results, which we have successfully done for Pareto and IIA thus far. Our methods for this included testing certain systems that would reflect moving candidates/voters around, and testing these systems using some code we wrote for this purpose. Some other topics for our research are translating specific voting systems, such as popular vote, pairwise competitions, and point systems to dice, and showing whether or not these ideal traits hold for these translations, which we have successfully done for some. With this, we hope to gain a better understanding of these systems and how using something truly random, such as dice, has any reflection on the systems we have in place.

The purpose of this research is to better understand the connections between dice and elections in voting theory. We explored the phenomenon of nontransitivity in dice and elections. In a set of 3 dice, labeled A, B, C, it is straightforward to designate sides such that, on average, A beats B, B beats C, and C beats A. Similarly, in an election of more than two candidates, where voters create a preference list for the set of candidates, it is simple to make nontransitive relationships in voter preference. Motivated by these situations, we aimed to find a correspondence between these two mathematical objects. Additionally, we defined the concept of an overall winner for a set of dice and an election, and explored the effects of nontransitivity in determining it. We used tournaments (from graph theory) as a tool to visualize the relationship structure among dice and elections. As a result of this work, we have shown that, from any dice set, an election can be constructed such that the dice possess the same winning-losing relationships. The converse is also true: given an election, a set of dice can be constructed with identical winning structure to the election. We defined the notion of a contest, which serves as a classification for both objects. Finally, we developed a triangular inequality for contests, and proved that dice and elections both satisfy this inequality.

 A common class of problems is that of understanding the large-scale properties of a system defined by small-scale relationships. These problems arise in fields such as statistical mechanics. We have been considering such problems as they apply to graphs, mathematical structures that can be used to represent networks, by investigating the limiting behavior of finite approximations to infinite graphs. In particular, we have been studying the growth rate of the number of spanning trees on these finite graphs. This allows us to quantify the complexity of various infinite graphs. The Matrix-Tree Theorem gives a formula for the number of spanning trees on a finite graph using techniques from linear algebra. This allows us to numerically approximate growth rates. These calculations can help us see patterns and develop theoretical approaches for obtaining exact values, as well as test theoretical results. In the case of graphs with translational symmetry, we have used techniques from group theory, a powerful area of mathematics, to derive an elegant expression for the exact growth rate in terms of Mahler measure. This is an integral formula connected to dynamical systems and number theory, among other research areas.While these considerations on spanning trees are facilitated by the Matrix-Tree Theorem, there is no known formula for counting spanning forests, a structure consisting of one or multiple trees which is more natural in some ways. Our current research generalizes the questions on trees to various types of spanning forests. We have also been exploring different ways to consider how a graph grows. This work can be used to build on this active research area of determining properties of large-scale structures.

Prime numbers are fascinating objects in mathematics, fundamental to number theory and cryptography in particular. A primality test takes an integer as input and outputs whether that number is prime or composite. Most practical applications require primality tests to be efficient. Today, the largest known prime number has over twenty million digits. Proving that a number of this size is prime can be computationally expensive. Most tests are probabilistic and do not guarantee that any given prime number is in fact prime. Other tests fail to filter pseudo-primes, like Carmichael Numbers, that are in fact composite integers. In 2002, Agrawal, Kayal, and Saxena published the first deterministic primality test that also runs in polynomial time relative to the binary representation of the input. Although their algorithm represents an important breakthrough in the field of computational number theory, it is seldom used in practice. Our objective was to determine why this idealized algorithm is not practical enough compared to other primality tests. We have re-created the algorithm and optimized our initial naïve implementation by studying each step to achieve optimal complexity using Fermat’s Little Theorem, modular arithmetic, and the Fast Fourier Transform for multiplicative efficiency. To confirm that probabilistic methods are still preferred, we compared execution times and accuracy of other tests to our own results.

Increasing atmospheric CO2 has lead to increased uptake of CO2 by oceans causing ocean acidification. Ocean acidification is the long-term reduction of pH that can have negative effects on marine organisms, especially young calcifying organisms. In Puget Sound, ocean acidification has already caused large-scale oyster larval mortality, negatively affecting a significant contributor to the Pacific Northwest economy. The native eelgrass in Puget Sound, Zostera marina, may be a local mitigation strategy for ocean acidification. Marine plants take up CO2 through photosynthesis to create biomass increasing the pH of the water. In this study, water samples were collected from several depths within the water column above eelgrass beds and surrounding mudflats in Padilla Bay. Water samples were analyzed for pH, total alkalinity, dissolved oxygen, dissolved inorganic carbon, and chlorophyll A. Eelgrass shoots and density were measured for leaf area index and the Padilla Bay National Estuary Research Reserve weather stations were used to calculate solar irradiance to estimate the photosynthetic activity of the eelgrass beds. Results suggest that eelgrass beds, in addition to providing critical habitat, may also help locally mitigate ocean acidification beyond the area of the eelgrass beds. Future research analyzing the input and output of carbon in eelgrass beds will be critical to developing large-scale carbon and pH models in marine environments.

People have an instinctive understanding of what it means for an object to be three-dimensional: it has height, width, and length. There is similar familiarity with two-dimensional shapes like squares and triangles, as well as one-dimensional shapes such as lines. The dimensions of some objects, on the other hand, cannot be understood in this familiar way. Take, for example, a set of points in the xy-plane that do not form a straight line or any familiar shape from geometry. One way to measure the dimension of these objects is to calculate the Hausdorff dimension of the set of underlying points. Intuitively, the Hausdorff dimension is a measure of how many small boxes it would take to cover an object. Our research aimed to calculate the Hausdorff dimensions of more complicated curves, such as those arising from Brownian motion, which is a probabilistic model for a randomly moving particle. In particular, we set out to corroborate known results about the Hausdorff dimensions of Brownian motion and the Brownian frontier. Further, we wanted to strengthen unsolved conjectures regarding the Hausdorff dimension of the Brownian earthworm model, which is a model of mass redistribution on the d-dimensional integer lattice introduced by Professor Krzysztof Burdzy. We attempted to develop answers to these problems by simulating Brownian motion in two dimensions and performing statistical analysis on our results. Our model correctly calculated the known dimensions and was in line with Professor Burdzy’s previous conjecture regarding the Brownian earthworm. It is our goal that this research will ultimately contribute to a greater understanding of the earthworm model.

In the age of self-branding, viral memes, and the Kardashians, Instagram has played a critical role in developing the attention economy. Attention has become a form of capital itself, where value is assigned to the most popular, aesthetically-pleasing content on channels like Instagram. When users construct their online identities within the context of late consumer capitalism, they participate in a cycle of desiring and being desired, in order to generate attention. This process enforces the notion that you are what you post, where one’s identity is represented by the signifiers one displays online. As a consequence, Instagram has radically reshaped the ways in which individuals conceive of themselves and interact with others in contemporary society.

Many scholars have constantly claimed that a shift from defined benefit plans to defined contribution plans has been the wrong course of action in policy and it has brought direr effect to the economy. Economists have argued that defined contribution plans are generating inefficiency pension income for employees in the private sectors, creating further cost burden for taxpayers and discouraging employees to stay at a job. While most observers are accepting the current situation by sticking to the traditional defined benefit plans, there are significant indicators in which the shift to defined contribution plans is inevitable. Economic shocks such as financial crisis, new health policy like Affordable Care Act and Congress abolishment of the Aid to Families with Dependent Children federal assistance program have changed the tides. They have brought a great deal of attention on how defined contribution pension plans can shape agents’ behavior in the labor market. Using datasets from the Survey of Income and Program Participation (SIPP) Synthetic Beta file and the National Quarterly Workforce Indicators, this paper examines the proposition that defined contribution plans motivate newly hired workers to stay at a job for a much longer period and gain higher weeks at pay without federal assistance programs taken place. This occurs particularly for male employees in public sectors. Regression analysis (OLS and logistic) demonstrates that workers enrolled in defined contribution pension plans have increasing personal level annual earnings taxed under FICA, higher probability in having employer- based health insurance and a larger family size in their individual households. Furthermore, this paper attempts to explain whether defined contribution plans have transitory or permeant effect on workers’ wellbeing during economic shocks using unit root test to determine whether there exists a stochastic trend.

Keywords: defined contribution pension plans, labor market, individual households, workers

The U.S is projected to face a shortage of a million graduates in STEM fields by 2022 according to the President’s Council of Advisers on Science and Technology. Much of this is the result of the current attrition rate from university STEM programs. Only half of all students entering the university intending to major in STEM fields graduate with a STEM degree. Decreasing this attrition rate by just 20% (or by 10 percentage points from 50% to 40%) will address over three quarters of this million-individual STEM gap. Therefore, it is of great interest to study the systemic factors that influence STEM attrition. This study will contribute to the existing body of knowledge on STEM attrition by studying the trajectory of students into and out of STEM fields by examining transcript records for every undergraduate student at the University of Washington from 1998 to 2007. As a first step, we will build a model that intuits students' intended majors by calculating affinity scores based on prerequisite classes taken by students. Examining changes in these affinity scores over time will allow for the identification of inflection points where a student's interests are likely changing. With this information, we can identify when students change fields of study and, on a broader level, also identify when they transition away from STEM fields. We hope to then use a supervised machine learning model to identify the features of students' transcript records that predict if and when this attrition will occur. Ultimately, we hope this information will help inform institutional strategy to improve STEM curriculum design, resource allocation, and retention strategies.

Many mainstream video games involve difficult game levels and challenges such as boss-battles and puzzles, leading some gamers to become frustrated as they struggle to advance to the next game level. In some cases, this frustration may lead gamers to give up on playing the game entirely. Although walkthroughs for video games are available online to help gamers when they get stuck, the gamers need to manually locate where they are in the game to obtain appropriate hints. Not only can manually looking for hints be inconvenient and time consuming, but it can also expose gamers to spoilers. We hypothesize that we can improve the gaming experience by providing gamers with a system that can automatically determine their current location in a game and serve up a link to a walkthrough video that demonstrates how to proceed. The goal is to queue up a video from the spot where gamers are stuck without the gamer having to know the level, scene, and time in the walkthrough for the hint. Our proposal is to use Deep Neural Networks (DNN) to learn the association of location and appearance pattern in the game(s) in order to locate the current game level. We will use recorded game play to train a DNN classifier for the games INSIDE and Portal, and evaluate the performance of the classifiers with gamers. Our goal is to design a system for NVIDIA Shield™ that provides gamers with help when needed. We hope by providing gamers with such a system, we improve the overall gaming experience and fewer players will give up on a game.

High-powered electromagnets are used in many different fields of science such as quantum physics and synthesis chemistry. Some projects require a very strong magnetic field to work. This magnet will allow students the freedom to design complex projects that utilize a strong magnetic field However, they are generally quite expensive and only available for use at well-funded research facilities and universities. Thus, we present our research in the design and construction of a cost-effective electromagnet capable of producing a highly-uniform magnetic field that can be used to further undergraduate research at Edmonds Community College. Our principal focus is on the reduction of engineering, fabrication and maintenance costs through a novel radial-flow cooling system which can be easily 3D-printed using high-temperature plastics in conjunction with a solid Bitter-plate solenoid design, requiring far less intricate and less expensive machining processes than conventional staggered-channel Bitter-plate designs. In order to make meaningful comparisons of generated field and thermal profiles with other published designs, our design maintains general shape and structural congruence to a design by Sabulsky, et. al., yet differing significantly in the design of the cooling system. 

The purpose of this research project is to demonstrate the feasibility of incorporating a new photocatalytic Diels-Alder reaction into the curriculum of undergraduate organic chemistry laboratories at University of Washington. Although the traditional Diels-Alder reaction is thermally allowed and does not require UV light catalysis, it fails to succeed when both the dienophile and diene are electron rich due to electrostatic repulsion between the reagents. Recently, an alternative photocatalytic Diels-Alder reaction was reported in the chemical literature involving a ruthenium bipyrazine catalyst, Ru(bpz)₃X, where X denotes the counter anion. This new photocatalytic method is advantageous because it allows the Diel-Alder reaction to work when both the diene and the dienophile are electron rich. Consequently, we believe that the new photocatalytic Diels-Alder reaction would be an attractive addition to our undergraduate laboratory curriculum by increasing the versatility of Diels Alder reagents to accommodate both electron rich substrates. Because the ruthenium catalyst is not commercially available, it must be synthesized from chloropyrazine. Two equivalents of chloropyrazine were coupled together under heated conditions and a Pd(Ph₃)₄ catalyst to yield a solution mixture containing bipyrazine, which was purified via silica flash chromatography. Subsequently, bipyrazine was reacted with RuCl₃xH₂O in ethylene glycol at 170°C to afford a solution mixture containing Ru(bpz)₃Cl₂. This chloride salt was reacted with NaBArF to precipitate the desired catalyst, Ru(bpz)₃X, where X represents the anion BArF-. Finally, the catalyst was purified through alumina flash chromatography. We intend to test the efficacy of this new photocatalytic method using the electron rich substrates trans anethole (dienophile) and isoprene (diene) in the upcoming months. We expect that the two photocatalysts we have synthesized will catalyze the cycloaddition reaction between isoprene and trans anethole with endo stereoselectivity.