We study the fundamental properties of femtosecond-laser (fs-laser) hyperdoping by developing techniques to control the surface morphology following laser irradiation. By decoupling the formation of surface roughness from the doping process, we study the structural and electronic properties of fs-laser doped silicon. These experiments are a necessary step toward developing predictive models of the doping process. We use a single fs-laser pulse to dope silicon with sulfur, enabling quantitative secondary ion mass spectrometry, transmission electron microscopy, and Hall effect measurements. These measurements indicate that at laser fluences at or above 4 kJ m-2, a single laser pulse yields a sulfur dose > (3 ± 1) x 1013 cm–2 and results in a 45-nm thick amorphous surface layer. Based on these results, we demonstrate a method for hyperdoping large areas of silicon without producing the surface roughness.
We demonstrate amorphous and polycrystalline anatase TiO2 thin films and submicrometer-wide waveguides with promising optical properties for microphotonic devices. We deposit both amorphous and polycrystalline anatase TiO2 using reactive sputtering and define waveguides using electron-beam lithography and reactive ion etching. For the amorphous TiO2, we obtain propagation losses of 0.12 dB/mm at 633 nm and 0.04 dB/mm at 1550 nm in thin films and 3 dB/mm at 633 nm and 0.4 ± 0.2 dB/mm at 1550 nm in waveguides. Using single-mode amorphous TiO2 waveguides, we characterize microphotonic features including microbends and optical couplers. We show transmission of 780-nm light through microbends having radii down to 2 μm and variable signal splitting in microphotonic couplers with coupling lengths of 10 μm.
We fabricate submicrometer-width TiO2 strip waveguides and measure optical losses at 633, 780, and 1550 nm. Losses of 30, 13, and 4 dB/cm (respectively) demonstrate that TiO2 is suitable for visible-to-infrared on-chip microphotonic devices.
Chemotaxis plays a critical role in tissue development and wound repair, and is widely studied using ex vivo model systems in applications such as immunotherapy. However, typical chemotactic models employ 2D systems that are less physiologically relevant or use end-point assays, that reveal little about the stepwise dynamics of the migration process. To overcome these limitations, we developed a new model system using microfabrication techniques, sustained drug delivery approaches, and theoretical modeling of chemotactic agent diffusion. This model system allows us to study the effects of 3D architecture and chemotactic agent gradient on immune cell migration in real time. We find that dendritic cell migration is characterized by a strong interplay between matrix architecture and chemotactic gradients, and migration is also influenced dramatically by the cell activation state. Our results indicate that Lipopolysaccharide-activated dendritic cells studied in a traditional transwell system actually exhibit anomalous migration behavior. Such a 3D ex vivo system lends itself for analyzing cell migratory behavior in response to single or multiple competitive cues and could prove useful in vaccine development.
Femtosecond (fs) laser irradiation of a silicon substrate coated with a thin film is a flexible approach to producing metastable alloys with unique properties, including near-unity sub-band gap absorptance extending into the infrared. However, dopant incorporation from a thin film during fs-laser irradiation is not well understood. We study the thin film femtosecond-laser doping process through optical and structural characterization of silicon fs-laser doped using a selenium thin film, and compare the resulting microstructure and dopant distribution to fs-laser doping with sulfur from a gaseous precursor. We show that a thin film dopant precursor significantly changes the laser-material interactions, modifying both the surface structuring and dopant incorporation processes and in turn affecting p-n diode behavior.
We observe an insulator-to-metal (I–M) transition in crystalline silicon doped with sulfur to non-equilibrium concentrations using ion implantation followed by pulsed laser melting and rapid resolidification. This I–M transition is due to a dopant known to produce only deep levels at equilibrium concentrations. Temperature-dependent conductivity and Hall effect data measured for temperatures T > 1.7 K both indicate that a transition from insulating to metallic conduction occurs at a peak sulfur concentration between 1.8 and 4.3 × 1020 cm–3. Conduction in insulating samples is consistent with variable range hopping with a Coulomb gap. The capacity for deep states to effect metallic conduction by delocalization is the only known route to bulk intermediate band photovoltaics in silicon.
Silicon hyperdoped with chalcogens via femtosecond-laser irradiation exhibits unique near-unity sub-bandgap absorptance extending into the infrared region. The intense light-matter interactions that occur during femtosecond-laser doping produce pressure waves sufficient to induce phase transformations in silicon, resulting in the formation of metastable polymorphic phases, but their exact formation mechanism and influence on the doping process are still unknown. We report direct observations of these phases, describe their formation and distribution, and consider their potential impact on sub-bandgap absorptance. Specifically, the transformation from diamond cubic Si-I to pressure-induced polymorphic crystal structures (amorphous Si, Si-XII, and Si-III) during femtosecond-laser irradiation was investigated using scanning electron microscopy, Raman spectroscopy, and transmission electron microscopy. Amorphous Si, Si-XII, and Si-III were found to form in femtosecond-laser doped silicon regardless of the presence of a gaseous or thin-film dopant precursor. The rate of pressure loading and unloading induced by femtosecond-laser irradiation kinetically limits the formation of pressure-induced phases, producing regions of amorphous Si 20 to 200 nm in size and nanocrystals of Si- XII and Si-III. The surface texturing that occurs during femtosecond-laser irradiation produces inhomogeneous pressure distributions across the surface and causes delayed development of high-pressure silicon polymorphs over many laser pulses. Finally, we find that the polymorph phases disappear during annealing more rapidly than the sub-bandgap absorptance decreases, enabling us to decouple these two processes through post-treatment annealing.
We describe two ways in which pulsed lasers can be used to increase efficiency in photovoltaic devices. First, pulsed-laser hyperdoping can introduce dopants into a semiconductor at non-equilibrium concentrations, which creates an intermediate band within the bandgap of the material and modifies the absorption coefficient. Second, pulsed-laser irradiation can enhance geometric light trapping by increasing surface roughness. Hyperdoping in silicon enables absorption of photons to wavelengths of at least 2.5 μm, while texturing enhances the absorptance to near unity at all absorbing wavelengths. In this paper, we review both effects and comment on outstanding questions and challenges in applying each to increasing the efficiency of photovoltaic devices.
The Force Concept Inventory (FCI) has influenced the development of many research-based pedagogies. However, no data exists on the FCI’s internal consistency or test-retest reliability. The FCI was administered twice to one hundred students during the first week of classes in an electricity and magnetism course with no review of mechanics between test administrations. High Kuder–Richardson reliability coefficient values, which estimate the average correlation of scores obtained on all possible halves of the test, suggest strong internal consistency. However, 31% of the responses changed from test to retest, suggesting weak reliability for individual questions. A chi-square analysis shows that change in responses was neither consistent nor completely random. The puzzling conclusion is that although individual FCI responses are not reliable, the FCI total score is highly reliable.
Silicon doped with non-equilibrium concentrations of chalcogens using a femtosecond laser exhibits near-unity absorption of sub-bandgap photons to wavelengths of at least 2500 nm. Previous studies have shown that sub-bandgap absorptance decreases with thermal annealing up to 1175 K, and that the absorption deactivation correlates with chalcogen diffusivity. In this work, we show that sub-bandgap absorptance can be reactivated by annealing at temperatures between 1350 K and 1550 K followed by fast cooling (> 50 K/s). Our results suggest that the defects responsible for sub-bandgap absorptance are in equilibrium at high-temperatures in hyperdoped Si:chalcogen systems.
In this dissertation we examine several issues related to the retention of underrepresented minority students in physics and science. In the first section, we show that in calculus-based introductory physics courses, the gender gap on the FCI is diminished through the use of interactive techniques, but in lower-level introductory courses, the gap persists, similar to reports published at other institutions. We find that under-represented racial minorities perform similar to their peers with comparable academic preparation on conceptual surveys, but their average exam grades and course grades are lower. We also examine student persistence in science majors; finding a significant relationship between pedagogy in an introductory physics course and persistence in science. In the second section, we look at student end-of-semester evaluations and find that female students rate interactive teaching methods a full point lower than their male peers. Looking more deeply at student interview data, we find that female students report more social issues related to the discussions in class and both male and female students cite feeling pressure to obtain the correct answer to clicker questions. Finally, we take a look an often-cited claim for gender differences in STEM participation: cognitive differences explain achievement differences in physics. We examine specifically the role of mental rotations in physics achievement and problem-solving, viewing mental rotations as a tool that students can use on physics problems. We first look at student survey results for lower-level introductory students, finding a low, but significant correlation between performance on a mental rotations test and performance in introductory physics courses. In contrast, we did not find a significant relationship for students in the upper-level introductory course. We also examine student problem-solving interviews to investigate the role of mental rotations on introductory problems.
Researchers and practitioners routinely use the normalized gain (Hake, 1998) to evaluate the effectiveness of instruction. Normalized gain (g) has been useful in distinguishing active engagement from traditional instruction. Recently, concerns were raised about normalized gain because it implicitly neglects retention (or, equivalently, "losses"). That is to say, g assumes no right answers become wrong after instruction. We analyze individual standardized gain (G) and loss (L) in data collected at Harvard University during the first five years that Peer Instruction was developed. We find that losses are non-zero, and that losses are larger among students with lower pre-test performances. These preliminary results warrant further research, particularly with different student populations, to establish whether the failure to address loss changes the conclusions drawn from g.
In this thesis, we present the results of three experiments that combine techniques from the elds of ultrafast nonlinear optics and plasmonics, with the aim of developing tools for improved surface-enhanced Raman spectroscopy and biological cell transfection. We fi rst describe the use of femtosecond laser pulses to generate large areas of a nanostructured silicon surface which is used as a new type of substrate for surface-enhanced Raman scattering (SERS). We perform spectroscopic characterization of this substrate and nd its Raman cross-section enhancement factor to be on the order of 10^7. This large, spatially-uniform, and reproducible enhancement factor is nearly constant across the near-infrared spectral region. In a second experiment, we develop a technique to spatially isolate the \hot spots" on SERS substrates. This technique leverages the plasmonic near eld enhancement of metallic nanostructures to preferentially expose a commercial photoresist using femtosecond laser pulses. By isolating the hot spots, analyte molecules adsorb only to the regions of largest electromagnetic enhancement. Compared to an unprocessed substrate covered with a sub-monolayer of benzenethiol molecules, a processed substrate shows a 27-fold im- provement in its average Raman cross-section enhancement factor. Finally, we present a proof-of-principle experiment which demonstrates high-throughput ultrafast laser transfection of biological cells using large-area plasmonic substrates. Utilizing the fi eld localization properties of a substrate fabricated using photolithography, wet etching, and template stripping, we demonstrate the introduction of silence RNA (siRNA) molecules into cells with an efficiency of approximately 50% after exposure to femtosecond laser pulses.
In this Manuscript, we present the fabrication and spectroscopic characterization of a large-area surfaceenhanced Raman scattering (SERS) substrate, as well as a method for improving femtomole-level trace detection (109 molecules) using this substrate. Using multiphoton-induced exposure of a commercial photoresist, we physically limit the available molecular adsorption sites to only the electromagnetic "hot spots" on the substrate. This process prevents molecules from adsorbing to sites of weak SERS enhancement, while permitting adsorption to sites of extraordinary SERS enhancement. For a randomly adsorbed submonolayer of benzenethiol molecules the average Raman scattering cross-section of the processed sample is 27 times larger than that of an unprocessed SERS substrate.
This study explored 20 tenured professors' teaching improvement efforts in introductory undergraduate science, technology, engineering and mathematics (STEM) classrooms at two major American research universities (MRUS). It identified the mechanisms central to these professors' efforts to improve their undergraduate teaching and the influences and resources shaping those efforts. Despite the billions of federal dollars invested in STEM educational enhancement, STEM professors' teaching improvement efforts are little understood. This research examined this problem through analysis of 40 in-depth interviews (two per participating professor), 36 observations of professors' classroom teaching, and hundreds of documents representing professors' instructional efforts and career progression. The following propositions summarize key study findings: First, contrary to prevailing views, some STEM professors in MRUs do engage in efforts to improve their introductory teaching. Second, some of these professors employ creative, strategic, and systematic designs in so doing. Third, STEM professors' teaching improvement efforts are contextualized by internal and external forces that may facilitate or stymie their teaching improvement endeavors. Finally, STEM professors may be aware of institutional and external resources available to them as supports toward introductory STEM teaching improvement. Taken together, the data suggest that some university STEM professors do engage in efforts to improve their teaching and that such effort may be more common than popular opinion holds. The study revealed the inaccuracy of common beliefs and policy assumptions that the large majority of MRU-based STEM professors neglect their introductory teaching, do not care about it, lack knowledge about students and pedagogy, and prefer use of (and consistently rely on) conventional teaching approaches. To the contrary, all 20 participating professors were found to devote extensive energy toward improving their introductory teaching. Further, all 20 participants indicated extensive knowledge of introductory STEM subject matter, students, and pedagogies. The study also identified over 30 innovative pedagogies that participating STEM professors employed in their classrooms. Drawing on these findings, the study calls for public and policy-level reconsiderations of what it means to be a research-active STEM professor, including revision of extant ideas about how these professors invest their energies and time, and what they care about.