Student Research

Ian Storer, David Starling

Spectroscopy is a useful technique in the sciences because it allows one to reliably identify unknown substances. Due to the precision required, modern spectrophotometers can exceed the $5,000 mark for standard commercial products. Additionally, typical spectrophotometers have large footprints and slow acquisition rates. To that end, we have implemented a modern and unique imaging technique known as compressive sensing that can both decrease acquisition time and cut cost.

The standard method for data acquisition is a process called "raster scanning" that results in a low signal to noise ratio. In contrast, a straightforward optical setup along with a modern technique called compressive sensing allows us to get viable data at low comparative cost. The setup consists of an LED light source, a diffraction grating, a lens, some mirrors, a digital micro-mirror device (DMD), and a custom photodiode sensor. The DMD is the key to compressive sensing and takes the place of a moving slit in the case of standard raster scanning. Random portions of the light are reflected by the DMD and measured by the sensor. This data is recorded and then used to reconstruct the image, resulting in the spectrum of the light that passes through the setup. To use this system for spectroscopy, a sample would be placed in the path of the beam and the new data set would be compared to the original white data set to determine what light is absorbed. This data can then be used to determine the properties of a substance. The results from this configuration are preliminary; however, we can create a 230 nm spectrum with 0.38 nm resolution in 10 s.

Mari Magabo, Dorothy Carter, Joseph Ranalli, David Starling

With rapid advances in technology, smart phones have become more accessible to average individuals, replacing and sometimes even exceeding the performance of other single-purpose tech gadgets. One such example is the camera. In today’s day and age, people are more inclined to take pictures with their smart phone due to their portability and convenience. As such, amateur stargazers may benefit from an apparatus that allows them to take photographs of the night sky with their smart phones. While products exist on the market today designed to mount cellphones to telescopes, their high prices make them undesirable to most amateur astro-photographers.

To address this problem, our group has designed a simple, adjustable phone mount intended to work with most smart phones. The mount can be easily printed by a 3-D printer for a fraction of the cost compared to the commercial mounts sold today. After several iterations, the final design comprises two parts made from thermoplastic that not only ensure safe attachment of the user’s phone but sturdy mounting to a 1.25 inch telescope eyepiece. Additional supplies needed to secure the two parts together are available in any retailer that offers home improvement or hardware supplies. In total, the phone mount costs approximately $2 to print and assemble. Tests are currently underway to evaluate stability and durability.

Tyler Fuehrer, David Starling

Computer programs, much like the one developed here, are constantly being created by astrophysicist in order to process data. The specialized nature of their research requires such custom programs. In particular, astrophysical data requires many hidden layers of processing before it is possible to analyze. Here, we developed code in the Python programing language [1] that will process images from a custom Android smartphone app (developed by Dr. Joseph Ranalli) that was used to take photographs of celestial bodies. The android app takes a quick burst of photographs, waits a set amount of time, and then repeats this process. The Python program takes this raw data and compiles it into a single file as discussed below.

Our program first enumerates the number of files it will need to process using the file type of our photos. Using a specific naming standard for the images the program is able to search through the file names to find the amount of images in a single burst. We use this information to stack each burst of images into a single composite image. This process repeats on each set of bursts. Due to the time between each burst of images, the rotation of the Earth has caused a rotation between the images. Therefore, we rotate the composite images into the same orientation [2]. We take these rotated images and store them in a data cube or stack them together, as needed. The program has been tested successfully on simulated images and is ready for testing with field data. We will present portions of the code along with the resulting data.

Richard Michael, Thomas Klein, Edward Miller, Blake Burger, Janak Jethva, Joe Zolnowski, Joseph Ranalli, David Starling

An interesting property of a light emitting diode (LED) is that it can actually act as a light absorber for use as a detector of single photons. When a photon hits a negatively biased LED, the one-way current flow characteristic of the LED is disrupted and a short current pulse will flow through the LED. The team set out with the initial goal of counting single photons at a high rate, requiring the construction of an active quenching circuit (i.e., a circuit that rapidly resets the LED after each photon detection event).

Throughout the course of the research, the team experimented with various circuits to compare the effectiveness of each. The first goal was to research active quenching circuits in the literature in order to find a suitable choice for the detector. After narrowing down potential circuits, each one was tested with a variety of common LEDs in order to determine which circuit was the most effective.

Once the most effective circuit had been determined, the team optimized the capacitance to decrease the reset time of the diode; this allows for an increased rate of single photon detection events. After attempting various capacitor combinations, it was determined that a 150 microfarad capacitor was able to reduce the reset time of the LED by 22.5%. Therefore, the goal of decreasing the reset time of the LED with a quenching circuit was fulfilled; however, there are still several avenues for improvement worth exploring. Several components of our existing circuit could be limiting our detection event frequency. For instance, a different range of capacitors and possibly higher quality LED’s can be investigated in order to further optimize our quenching circuit.

Scott Gauer, Joseph Ranalli, David Starling

Flame imaging is a method of recording spatial information on the combustion of a given material. This kind of imaging is very useful for gathering information on a flame's intensity, combustion efficiency, emission spectra, and more. Data compiled from flame research has applications in optics, petroleum research and alternative energy. Most of the visible light from a flame is emitted from hot soot, whereas the light from the chemical process of combustion is much dimmer. As a result, many researchers encounter an obstacle when working in this field. Currently, intensified charge-coupled devices (ICCDs) are the primary instrument used to capture and study dim flame images. However, these ICCDs are prohibitively expensive, and many smaller research institutions simply can not afford one. In our experiment, we test the viability of a less expensive alternative to ICCDs, known as compressive sensing. Our experiment uses a pair of lenses to focus the image of a flame onto an array of switchable micromirrors, called a digital micromirror device. These mirrors are laid out in much the same way that pixels are laid out in a camera sensor or ICCD. The photons from the image reflects off these mirrors into another pair of lenses that focuses the light down into an optical fiber connected to a single photon detector. Using a customized Python program, we command a random pattern of mirrors to switch “on” and reflect the flame light into the fiber. On average, half of the mirrors on the array are “on”, while the other half are “off” and dump the photons that hit them. We record the number of collected photons with a single photon detector from a predetermined number of patterns. We then reconstruct the original image using an algorithm developed at Rice University known as TVAL3. This algorithm converts the photon counts and patterns into what is essentially a large under-determined matrix equation, which it then solves subject to a specific constraint. Since most objects in nature do not have sharp contrasts, the matrix equation is solved for a minimized contrast gradient. The main benefits of compressive sensing over ICCD imaging are its lower cost, for which resolution is sacrificed. Our experiment was successful and performed much better than another common type of imaging, a raster scan. While the raster scan was not able to discern the image signal from shot noise, our compressive sensing scan had a high enough signal to noise ratio to provide a discernible image of the light produced by the excitation of the C-H bond in propane gas.

Mark Wychock, Tyler Gregory, Meagan Pandolfelli, Chris Moran

Nuclear fusion is a thermonuclear reaction in which two or more light nuclei collide together to form a larger nucleus, releasing a great amount of binding energy the in the process. Fusion and fission are natural processes that occur in stars. Fission is the process in which an unstable nucleus splits into two nuclei over a period of time or by induced fission of a neutron bombarding a radioactive atomic nucleus. In stars, it is understood that the fusion-fission process provides a near constant source of energy from proton-proton chain reactions. Although a fusion reaction generates more energy than a fission reaction, modern nuclear power plants utilize fission processes due to the stability of the fission reaction, convenience, and cost of production. If nuclear fusion could be produced in a commercial setting, it could provide 3-4 times the energy a fission reaction generates. Fusion material such as deuterium, a key component in thermonuclear reactions, can be distilled from seawater providing a virtually infinite and promising source of energy in the future. The understanding and development of thermonuclear reactions and reactors is accomplished by the aid of differential equations. Engineers and scientists are able to observe behaviors, such as mass conservation, hydrostatic equilibrium states, and energy generation, of induced nuclear reactions from differential equations when developing fusion reactors. Nuclear fusion reactors are undergoing development to replace obsolete nuclear fission reactors. Two potential types of fusion reactors in development are laser ignition and magnetic confinement. In the research project, the scientific and mathematical applications behind magnetic confinement will be explored since it is more economical and efficient than laser confinement. Specifically, we will examine how the Grad-Shafranov equation plays a pivotal role in the operation of tokamak and stellarator reactors.