Physics Program Presents
Summer Research Presentations
Friday, September 23, 2016
Hegeman 107
12:00 pm EDT/GMT-4
Testing the quantum hall effect in hybrid graphene12:00 pm EDT/GMT-4
Nowell Stoddard and Kai Naor
Bard Nanolab
Mentor: Paul Cadden-Zimansky
Over the summer we worked in the Bard Nanolab to fabricate graphene samples and create devices capable of measuring hall resistance at low temperature and high magnetic field. The goal of these devices was to study the quantum hall effect, a unique property of two dimensional semiconductors at high magnetic field. This effect has been studied extensively in monolayer graphene but our aim was to study it in Hybrid graphene, that is, graphene that contains both a single and a double atom layer on the same piece. On the last week were able to go to down to the national magnet lab to test our devices.
Modelling Exoplanetary Microlensing
Eleanor Turrell
Quantum Gravity Group, Marseille France
Bard College & ISSI Bern Switzerland
Mentors: Eleni Kontou & Hal Haggard
Exoplanets act as gravitational lenses by bending light as it moves from a distant source star toward Earth. We created hypothetical light curves to characterize exoplanets if microlensing is observed. We looked at configurations including one lens, two lenses, and plan to examine three lens configurations next.
Pre-Processing and Classification of Hyperspectral Imagery Via Selective Inpainting
Victoria Chayes
University of California, Los Angeles
Mentors: Wei Zhu, Andrea Bertozzi, & Stan Osher
We devise a semi-supervised means to classify and sharpen a hyperspectral image. Hyperspectral imagery (HSI), wherein sensors capture data at hundreds of different wavelengths, has numerous applications in agriculture, environmental science, mineralogy, medical imaging, and surveillance, because of its fundamental ability to allow the identification of separate objects or materials that cannot be differentiated on sight. By discarding identified "noisy" pixels and applying an accelerated proximal gradient inpainting scheme, we are able to de-noise pixel signatures and sharpen bands, leaving a clear enough hyperspectral image that clusters can be identified purely by Euclidean distance from pre-known endmembers.
Pizza and refreshments will be provided!
For more information, call 845-752-7302, or e-mail [email protected].
Time: 12:00 pm EDT/GMT-4
Location: Hegeman 107