First Week of the Program From Left to Right: Grace Jeanpierre, Indra Gonzalez Ojeda, Amber Perry, Robert Weinbaum, Jorin Graham, Tyler Howard, Claire Onsager, Gabe Seymour
Last night of the program
Amber Perry: Mask Lithography and Direct Laser Writing of 2D and 3D Magneto-Photonic Microstructures for Biomedical and Quantum Applications
Whiteboard Video: This summer I worked in the Quantum and Molecular Photonics Laboratory (Laboratoire de Photonique Quantique et Moléculaire – LPQM), École Normale Supérieure Paris-Saclay (ENS Paris-Saclay) under the supervision of Dr. Ngoc Diep Lai and PhD student Thi Huong Au. My work explores the coupling of a single photon source based on a colloidal semiconductor nanocrystal (NC), also known as a colloidal quantum dot (QD), into a polymer-based photonic structure doped with magnetic nanoparticles. In the field of nanophotonics, achieving the control of single emitter properties embedded in a hybrid nanocomposite is highly desired for a wide range of applications in opto-electronic devices, microrobotics, biological-labeling, drug delivery, quantum technology, etc. Before structure fabrication, samples must be prepared. This involves spin-coating various layers onto a clean glass substrate. The first layer is a PMMA buffer layer, which prevents the structures from sticking to the glass surface and hence allows them to move in response to the applied magnetic field. The second layer, on which the structures will be fabricated, contains CdSe/CdS colloidal QDs and SU-8 2005 or 2000.5 doped with magnetite (Fe3O4) magnetic nanoparticles (MNPs). SU-8 is a type of photoresist, meaning that it's a light-sensitive material used to form a patterned coating on a surface in photolithography or photoengraving. Note that sample preparation including the PMMA layer is used only if structure movement is desired. I began the summer by fabricating microstructures using a mask optical lithography technique. This involves applying a patterned mask with micro-sized cutouts of various shapes directly against my sample and exposing it with a 355nm continuous-wave (cw) laser. Since SU-8 is a negative photoresist, the developing process dissolves away areas of the photoresist not exposed to the laser (the opposite occurs for positive photoresist) and leaves the fabricated microstructures on the substrate. If a PMMA layer was applied, the structures are free-floating in the developer and their movement can by manipulated with an applied magnetic field. Once successfully fabricating and manipulating structures made with mask lithography, I began to work with a second method of fabrication called low one-photon absorption (LOPA)-based direct laser writing (DLW). DLW employs various absorption mechanisms, such as one-photon absorption or two-photon absorption. Our research presents LOPA-based DLW, which is simpler, more compact, and less expensive than other methods of fabrication. LOPA is a process by which the photoresist absorbs one photon at a time at a low rate, meaning a higher intensity of photons is necessary for the photoresist to be fully polymerized. Since SU-8 photoresist and QDs have low absorption at an excitation wavelength of 532 nm, DLW of such composite using a 532 nm laser operates in the LOPA regime. In this DLW setup, a piezoelectric translator and an objective lens with high numerical aperture are used to control where the laser is focused onto the photoresist. Using LabVIEW, we can program a path for the focused laser to take, allowing fabrication of arbitrarily-shaped 2D and 3D structures. As with mask lithography, development of the sample dissolves away areas not exposed to the laser, forming microstructures from the composite. Both the photoresist and QDs have fluorescent properties, meaning that when excited by the laser they give off light. Using a LabVIEW program, we can track this emitted light and thus determine what layer of our sample contains the photoresist and what position within the composite the QDs are located at. By mapping the location of the QDs we are able to fabricate individual structures containing a single QD. Once the structures are fabricated and developed, we can manipulate their movement with an applied magnetic field. This achieves our goal of fabricating movable microstructures with coupled magnetic and fluorescent properties and containing a single photon source. As a result of this work, we are submitting two abstracts, one for a poster and one for an oral communication, to the Photonics West conference (San Francisco, USA) in February 2019. During this experience at ENS Paris-Saclay, I had the opportunity to learn not only inside the lab, but outside it as well. Working closely with a PhD student gave me insight into what it’s like to pursue research at a PhD level and being involved with a French lab allowed me to learn more about the European educational system in the domain of science. This experience was extremely rewarding and has motivated me to return to France or Europe to pursue higher education, research, and a professional career in the sciences. Top
Grace Jeanpierre: The effect of uniaxial stress in magnetic materials using using eddy current, a nondestructive testing
Whiteboard Video: This summer, I worked under DrPr. Laurent Daniel and DrPr. Yann Le- Bihan with the CentraleSupelec in Group of Electrical EngineersEngineering- Paris (GeePs). Previously, DrPr. Daniel and Dr. Pr. Le- Bihan studied and modeled the effect of uniaxial stress in magnetic materials using using eddy current, a nondestructive testing (testing (NDT) technique (NDT). This summer, I was responsible for creating a 2D and 3D model in COMSOL Multiphysics that could detect stress in a iron steel pipe by looking at changes in Earth’s terrestrial magnetic field. Pipelines are the most favorable way to transport large quantities of liquids and gases. Improper maintenance and threats such as defects, corrosion, damage from construction, cracking, earthquakes, landslides, and extreme temperatures can cause environmental hazards, property damage, structural damage, and threaten life. Stress not only affects the performance of pipes but determines lifetime. It has a great effect on the magnetic permeability of ferromagnetic materials. This is known as magneto-mechanical behavior. A magnetic NDT is used a technique used toto determine determine a pipeline’s integrity by measuring changes in magnetic flux.Stress not only affects the performance of pipes but determines lifetime and has a great effect on the magnetic permeability of ferromagnetic materials. This is known as magneto-mechanical behavior. In the COMSOL model, the steel pipe was located 1 meter below the soil. The pipe had a radius of 8.4 cm and thickness of 7 mm. The relative permeability for air and soil were chosen to reflect real world parameters. The induced terrestrial magnetic field was chosen to be 50 uT. The goal of the project was to determine whether a small defect (in this case a 10-degree angular sector) and the defect’s position had a measurable effect on earth’s magnetic field. Since stress is known to change the magnetic permeability of ferromagnetic materials, the pipe was given a relative permeability of 1000 while the defect’s relative permeability ranged from 100-1000. A 1D sensor was placed 1m away from the pipe to collect data on magnetic flux density using a finite element method. Once it was determined that a change in magnetic permeability and a rotation in angle influenced earth’s magnetic field, a new defect was introduced. This new defect contained a missing angular sector that penetrated 50% of the pipe’s radius (3.5 mm deep). A new study was conducted to determine how the size and location of the defect affected the terrestrial magnetic field. Future work includes refining the mesh to reduce meshing noise, optimizing the study domain so that less computation time is needed, and implementing nonuniform stress in the pipe to create a more accurate model. I would like to thank the National Science Foundation for the opportunity to do research abroad in Paris, France. Thank you to DrPr. Laurent Daniel, DrPr. Yann Le- Bihan, and DrMr. Laurent Santandrea for hosting me in the GeePs lab, guiding me through this project, and checking in on my progress regularly. Thank you to my colleagues Ghida Al Achkar, Anderson Santos Nunes, and Diane Phan for all the laughs and lunches. Lastly, thank you to Dr. Steve Yalisove and John Nees for leading and maintaining this REU program. Top
Jorin Graham: Stationary and ultrafast transient optical properties of nanoparticles
Incident light on a noble metal nanoparticle causes the conduction electrons to oscillate. Resonance of this oscillation due to excitation at a characteristic frequency is referred to as localized surface plasmon resonance (LSPR). Excitation of noble metal nanoparticles at LSPR frequencies leads to efficient conversion of light into heat, strongly enhanced electromagnetic waves in the near field regime, and, when excited by an ultrafast pulse, ejection of electrons from the nanoparticle. In addition, the characteristics of the LSPR strongly depends on the shape, size, and composition of the nanoparticle and on the dielectric properties of the surrounding medium. This makes noble metal nanoparticles promising candidates for a variety of applications. My group is currently re-searching the effects of substrates on the optical proper-ties of gold-core, silver-shell nanorods (Au@Ag NRs) and the ultrafast properties of Au@Ag NRs. I was involved in both of these areas of research. As part of the research into substrate effects on Au@Ag NR properties, I worked to develop a microscopy system for studying the stationary and ultrafast transient optical properties of nanoparticles. In order to minimize noise in the spectral data, it is desirable to study homogeneous samples. Unfortunately, it is difficult to homogeneously deposit nanoparticles on a substrate. One solution is to study only a small area of the sample, so that the area under study is approximately homogeneous. However, with the equipment currently available to my group, in order to align the desired region with the probe beam and to limit the light to a small region requires use of a pinhole which may cause undesirable diffraction effects. The solution to this problem on which I worked was to couple a spectrometer to a microscope. I participated in the system design and performed the alignments for the prototype system. I successfully passed a signal through the system. However, the transmission was less than 1%, far to low for practical use. It was determined that this loss was due to the system and not misalignment of the optical components. New optical components will be ordered which should allow for a reduction in the losses. However, they will arrive after the end of my internship. As part of the research into the ultrafast properties of Au@Ag NRs, I performed data analysis on stationary absorbance spectra of colloidal Au@Ag NR samples. In order to understand the ultrafast properties of nanoparticles, one must first understand the stationary properties of these particles. The stationary properties can be studied with regular absorbance spectroscopy. Several studies have already performed detailed analyses on the stationary optical properties of Au@Ag NRs, connecting these properties to particle morphology and plasmonic proper- ties. As such, my goal was not to analyze new properties of Au@Ag NRs. Instead, my objective for this portion of my research was two-fold. My first goal was to provide a summary of the stationary optical properties of Au@Ag NRs. The second was to characterize the particular samples that my group will study. In order to accomplish this portion of my research, I drew data from three main sources: literature, experiment, and simulation. From the literature, I crafted a a review which de-tailed the expectations of the behavior of spectra of the Au@Ag NR samples for different thicknesses of the Ag-shell. Specifically, I discussed the number of peaks expected to be present in each spectrum, the relationship of these peaks to particle morphology and plasmonic properties, the shift of each of the peaks with changing Ag-shell thickness, the change in the intensity of the peaks with changing shell thickness, and the effect of the presence of a distribution of particle size in the experimental samples on the broadness of the peaks. For the experiment, I collected spectra for Ag-shell thicknesses of 0 nm, 2.5 nm, and 10.2 nm. I based my simulations on a code written by Dr. Cyrille HAMON, who collaborates with my group (Dr. HAMON also performed the synthesis of the particles). This code models the Au-cores as spherocylinders and the Ag-shells as rounded nanoprosims. It then solves for the extinction spectrum using the boundary element method. I began my simulation work by determining the aspect ratio (ratio of length to diameter; abbreviated as AR) of the Au-cores by finding the AR which produced a simulated spectrum with peaks in the same location as in the experimental spectrum. I then simulated spectra corresponding to the three Au@Ag NR samples, allowing me to directly compare the simulation and experimental results. After this, I performed simulations for various Ag-shell thicknesses and analyzed how the simulated peaks shift and how the intensities of the peaks change with changing shell thicknesses. From this, I determined that the Au@Ag NR model does not produce accurate results for shell thicknesses below about 2 nm. I then simulated spectra for a variety of particle sizes. Using these spectra, I created a code which finds a best fit to the experimental data from a linear superposition of simulated spectra for various particle sizes. This demonstrated that broadness of the primary peak in the experimental spectra could be largely attributed to the presence of a size distribution of particles in the experimental sample. Generally, I found that the literature, experiment, and simulation agreed well, albeit with some differences which may require further investigation.
This summer I worked with Dr. Mélanie Lebental at ENS Paris-Saclay in the Laboratoire de Photonique Quantique et Moléculaire doing research with 3D micro-lasers. Micro-lasers have multiple applications in the fields of telecommunications, sensing, and quantum chaos. Micro-lasers in 2D (thickness is on order of wavelength) have been well studied and characterized, but when you get to 3D there is little known work and the complexity greatly increases. This summer my primary focus was on the analysis of the orbit/s generated by barrels, the 3D equivalent of a 2D plano-concave Fabry-Perot resonator that was previously tested. To excite the microcavities, a 532 nm frequency-doubled Nd:YAG laser operating at 10 Hz pulses pumped the cavities overhead, and the resulting spectrum captured; the resulting orbit may be determined by the Fourier transform of the spectrum. My first task was to determine the group index of refraction of light by analyzing the spectrum of 50x50x40 μm cubes. Once, I had the group index I began to examine the spectrum produced by the 3D barrels. Initial results of the samples showed a lot of noise in the Fourier transform of the spectrum, with no visibly, distinct pattern. Thinking this was caused by interactions with the substrate, I moved on to test barrels that were supported on posts, however, these barrels showed similar results. After some calculations, we determined these barrels are emitting the whispering gallery mode (WGM), mode that follows the circular perimeter of the barrel. To reduce the WGM, I redesigned barrels to have a smaller diameter, with varying lengths, and curvature of the concave surface. To design the new barrels, I alternated between the CAD software OpenSCAD and COMSOL depending on resolution requirements. While waiting on the samples to arrive, my professor noticed how “dirty” the beam was and assigned me the task of realigning the system with a microscope and pinhole apparatus to clean the beam. Experimentation of these new samples still showed a dominant WGM in the samples, which puzzled us. So, in hopes to eliminate the WGM completely, I designed two barrels that included a series of light traps along the exterior. Analysis of these barrels showed the opposite result of what we had expected to see, the WGM was enhanced and the only visible mode in the spectrum. Reasons for this are still unknown. Along with analyzing the spectrum produced by the barrels, I ran some calculations on the ray an expected orbit should take and the stability it should have based on the geometric properties of the barrel. For this I did multiple configurations: i) plano-concave, ii) plano-elliptical, iii) concave-concave, iv) concave-ellitpical, v) elliptical-elliptical. Not only did I design barrels, I also designed additional 3D microcavities that will be tested later such as a circular strip, mobius strip, and geometry with a unique five bounce orbit. These results and designs will be used for future research. I would like to thank Dr. Lebental and Melissa Guidry for their help and supervision over the summer, Dr. Dominique Decanini at Centre de Nanoscience et de Nanotechnologies for assistance in the fabrication of these samples, University of Michigan who hosts the program, and NSF for their support and sponsorship. Top
Indra González Ojeda: Coupling fluorescence microscopy with scanning electrochemical microscopy
Whiteboard Video: This summer I worked in the ENS-Cachan campus, under the supervision of Prof. Fabien Miomandre in the supramolecular and macromolecular photophysics and photochemistry laboratory (PPSM). My lab’s specialty lies in electrochemistry and its applications. This is a very interdisciplinary field with applications in chemistry, physics and even biology. The specific sub-projects that I worked on this summer were directly related to the lab’s innovative coupling of fluorescence microscopy and scanning electrochemical microscopy. Here, a 3-electrode electrochemical cell is combined with a fluorescence microscope to monitor the changes in fluorescence brought upon by the application of a specific current or potential on a sample; My work can be divided in two main parts. Firstly, there has been no water-soluble dye characterized in this set-up. Doing this was a priority because through a water-soluble dye, biological systems can be studied, thus greatly expanding the applications for this new microscopy technique. I assisted with the characterization of resorufin (RF), a water-soluble molecule with a strong fluorescent signal. We first studied the “switching off” phenomenon in which a potential was applied to the RF solution using an electrochemical probe (platinum tip, 20 µm diameter). The potential “switches off” the fluorescence because it converts the RF into its non-emitting reduced form (dihydroresorufin). We specifically analyzed the intensity modulation between the fluorescent state and the “switched-off” state to evaluate the sensitivity of our instrument. We determined that this modulation amplitude is sensitive to the kinetics of the electrochemical reaction at the conducting substrate. The fluorescence intensity allows detecting the species produced at the substrate with more accuracy than the electrochemical current measured at the tip does. All good news as it proves that resorufin can serve as a good redox mediator in the future study of reactions. The second part of my project was modifying the excitation system to perform lifetime measurements. You see, our current set-up uses a wide-field white light continuous excitation system. This allows for the measurement of fluorescent intensity, but it does not offer a direct way to measure the decay of this fluorescence. We tried implementing a pulsed LED as the light source, but it sadly died shortly after we got it set-up. This was no problem however because Prof. Miomandre rented a pulsed laser that we successfully coupled into the system. After this, I was able to make lifetime measurements on a molecule called tetrazine. We used a time-correlated single photon counting card to detect the emitted molecule’s fluorescence. This gave us the delay time and absolute arrival time of each photon. We then performed an histogram analysis of these time values to observe the decay in fluorescence between pulses. We also measured the instrument response function (IRF) to obtain the time resolution at which we can have confidence in our measurements. We found that our time resolution is around 500 ps which is consistent with what we expected based on our equipment, and more than enough to study lifetime measurements in tetrazine. In the future my lab will try to achieve a total internal reflection fluorescence (TIRF) mode with the laser. We have already started thinking about ways to achieve this but, sadly, my two months went by very quickly. I thank the Optics in the City of Light program for this amazing experience (and to the NSF for funding us!). I also want to extend my deepest gratitude to Jeff Audibert and Laetitia Guerret for constantly guiding me in the lab, to Prof. Miomandre who welcomed me in his lab and took the time to explain the concepts to me and to the staff at ENS-Paris Saclay who made us feel at home. Top
Robert Weinbaum: High-resolution frequency spectrum for an attosecond pulse
Whiteboard Video: This summer, I worked in the Charles Fabry Laboratory at the Institut d’Optique under the supervision of Charles Bourassin-Bouchet. Physicists have become more and more interested in recent years with producing ultrashort laser pulses for studying many ultrafast phenomena in atomic or molecular physics. Scientists have been able to produce laser pulses with durations on the order of attoseconds, which is the primary time scale of electronic dynamics. One major challenge in this field is obtaining accurate measurements of the temporal profile of the laser pulse, which is critical if scientists want to use these pulses for any practical imaging. Many pulse reconstruction algorithms exist to accomplish this given a set of measurements, including a high-resolution frequency spectrum of the laser pulse. The challenge Charles and I faced is improving one such nondestructive technique of obtaining a high-resolution frequency spectrum for an attosecond pulse, called Time of Flight (ToF) Spectroscopy. In ToF Spectroscopy, some of the photons in the laser pulse interact with a gaseous sample and ionize the atoms. The kinetic energy of the electrons is the photon energy minus the ionization potential of the atom (which is well-known). The electrons travel down a flight tube and reach a microchannel plate (MCP) after some measured time, which we call their “time of flight.” By knowing the length of the flight tube and the electrons’ times of flight, we can deduce their speed, and work backwards to determine the electrons’ energy spectrum, which tells us directly the photon energy spectrum. By Planck’s Law, this directly tells us the frequency spectrum of the attosecond pulse. There are many factors that influence the resolving power of a ToF Spectrometer. One is that electrons with higher energies arrive with smaller time separations between one another, while the MCP maintains a constant sampling rate, making them more challenging to resolve. Two existing solutions to this are (1) to lengthen the flight tube – some as long as 8 meters have been constructed – or (2) to apply a retarding potential within the flight tube. The first solution is expensive and often impractical in a laboratory setting, while the second technique loses any electrons with kinetic energies insufficient to climb the potential well. Another challenge is that the exact flight path is not known, as the electrons can have trajectories slightly off-axis or could originate from anywhere within the volume of the gas source. Charles and I have spent the summer developing mathematical and numerical models to better understand these limits to the resolution, and have worked towards modifying existing ToF Spectrometers to dramatically improve the resolution – in particular at high energies, but also over broader energy ranges. This is useful because the shorter the laser pulse becomes in the time-domain, the longer it will become in the frequency-domain. Personally, I developed analytical approximations that characterize different sources of error. I then wrote a Python code that implemented a fourth-order Runge-Kutta technique to solve for the electron trajectories in 3D space to high precision, and to measure the resulting ToF profiles. We were able to compare my model with a commercial software called SIMION that simulates electron and ion trajectories. This code has proven invaluable for checking our analytical models against simulations and for running optimization algorithms on different components of the spectrometer geometry, and can calculate the trajectories to higher numerical precision and with less computational difficulty than the commercial software, as it was built specifically to test this particular case. I would like to thank the Insitut d’Optique for being such welcoming and kind hosts to me this summer, the University of Michigan – in particular Steve Yalisove, John Nees, and Bett Weston – for organizing the program, the National Science Foundation for their generous funding that allowed me to participate in this amazing experience, and Charles for being an amazing mentor and for allowing me the opportunity to be very independent and creative in conducting this research project.
Gabrial Seymour: Direct Laser Writing of metallic nanostructures in the applications of data storage and color nanoprinting
This summer, I worked in the Quantum and Molecular Photonics Laboratory (LPQM), École Normale Supérieure Paris-Saclay (ENS Paris-Saclay) under the supervision of Dr. Ngoc Diep Lai and Ph.D. student Mao Fei. My work focused on Direct Laser Writing (DLW) of metallic nanostructures in the applications of data storage and color nanoprinting. DLW is the process of focusing a laser beam on a material and realizing desired microstructures. Usually, this technique is used to fabricate submicrometer 2D and 3D polymeric structures. Very recently, the LPQM lab demonstrated that the DLW technique could also be used for the production of metallic nanostructures. In this case, the light beam is focused onto a thin film of metal to induce a local hot spot which causes the dewetting effect to occur. When the metallic thin film is dewetted, metallic nanoparticles that are of size on the order of tens of nanometers are formed. Under the illumination of a light beam, the metallic nanoparticles present a plasmonic effect, which means that light interacts with free electrons of these nanoparticles to create induced resonant dipoles. These induced resonant dipoles cause the absorption of light at a specific wavelength called the plasmonic resonance. Viewing the illuminated nanoparticles under either a transmission or reflection microscope shows the effect of the absorption and/or scattering of light by the induced resonant dipoles, resulting in the sample appearing as different colors. The plasmonic resonance spectra of the nanoparticles are dependent on the average size of the nanoparticles, so by changing the average size of the nanoparticles, we are effectively able to shift the absorption spectra, and tune the corresponding color of the sample. In practice, by changing the laser power or the scanning speed, we are able to change the average size of the nanoparticles, and thus, we are able to change the color of the fabricated structures. Using this idea, I explored a wide range of laser parameters on thin films of gold, silver, and a mixture of both gold and silver to realize structures such as letters and images with different colors. In past research, DLW was used on a thin layer of gold to successfully realize structures, so the first step of my research was to confirm these results, and to see if we could get similar, if not identical, results with a layer of gold of a different thickness than was originally used. After getting successful results with gold, I moved on to silver to see if the realization of these structures was possible and if a range of colors could be achieved. It was found that the realization of nanostructures with silver is possible, and while the range of colors achieved is not broad, the colors are finely tunable. To wrap up my research, I looked at the realization of structures on different mixtures of gold and silver. The results from these inhomogeneous mixtures were not what we expected, but the possibility of using these mixtures in real-world applications is a possibility after more testing and further understanding of the optical properties of these materials.
Top Claire Onsager: Fabrication of a 3D printed micromirror for microfluidic ray optics experiments
This summer I had the pleasure of working at the PPSM lab at ENS-Cachan with Dr. Robert Pansu and Jean Frederic Audibert. I was tasked with the fabrication of a 3D printed micromirror to work with the lab’s microfluidic ray optics experiments. In their current setup, they are observing crystallization in flow within a small capillary tube of appx. 300µm in diameter. During this study, they observed the capillary through a microscope with a bottom view. They hoped that with the use of micromirrors they could obtain a view of other areas of the tube. This created the need for a very specific part. The mirror needed be large enough to fit around the capillary without concealing the view. Another requirement was that the mirror cannot block the microscope light coming from above, requiring a small observation slit in the part. In addition, due to the images they wished to produce, it was necessary to fabricate two mirrors, a parabolic mirror and a right-angle mirror. Before we could do any fabrication, we needed to determine the restrictions of our 3D printer. I spent a week creating test prints to check how small we could print lines and holes. These tests helped me determine the printer’s accuracy and consistency. Once that process concluded, I began work on the curved mirror with a 600µm diameter. I created the design in Google SketchUP, making sure to include enough material for support and easy handling. The program Cura was then utilized to modify the printer settings to produce small details. In order to be used as a mirror, the curved surface needed to be optically smooth. This smoothness is below the resolution of the printer (250µm). To overcome this issue, we tried printing on a mold, a difficult process to do with our 3D printer. After many tests of the printer settings, we were able to print a design with the insertion of a wire during the print. The printer would then fill in the area around the wire with material, creating a circular mirror. However, upon examination of the mirror surface, we could see the texture of our wire. To remedy this, I found a way to use Acetone vapor to smooth the surface. This process allowed us to eliminate the ridges created by the wire, as well as any visible layers from the printer. Once the mirror was printed, we needed to make it reflective. This was done by use of a silver coating using Tollens reaction. I tested a variety of Tollens recipes before finally adjusting one from North Carolina State University to produce the desired result. The Tollens process is a back-coating method. This means that when coating a sample such as the inside of a glass jar, the best mirror shine will appear on the outside of the glass. However, we were using the process to front-coat our samples. The result left our mirrors dull and yellow tinted. With the use of chemical and mechanical polishing methods we found that we can increase this shine. After the circular mirror, I created the right-angle mirror. In this design, we added a support to keep the capillary a certain height from the mirror surface. Instead of the mold method, I created the shape with the printer directly. I then smoothed and silvered it in the same manner I had for the circular mirror. To prove that the mirrors worked, I needed to test them optically. Due to their small size, many of these tests involved a microscope that could take measurements and detailed images. With the right-angle mirror, it was easy to see the images created when a colored capillary is centered. By shining light onto the sample, one can easily see two reflections of our capillary on the mirror faces. Their location revealed that the printer was imprecise and resulted in the mirror angle to be closer to 120 degrees than 90. We believe that with better equipment and more adjustments these mirrors will work. The circular mirror was much trickier. Our original plan to find a point where reflected light converged soon appeared impossible given the small size. We then investigated new testing methods such as the use of a laser beam to provide a visible path of light. In the end, while we did not have a fully functional mirror, we were aware of the steps needed to make one and had many ideas for improvement.