Date of Graduation

5-2013

Document Type

Thesis

Degree Name

Master of Science in Microelectronics-Photonics (MS)

Degree Level

Graduate

Department

Microelectronics-Photonics

Advisor/Mentor

El-Shenawee, Magda O.

Committee Member

Naseem, Hameed A.

Second Committee Member

Ware, Morgan E.

Third Committee Member

Vickers, Kenneth G.

Keywords

Pure sciences; Applied sciences; Nanoparticles; Plasmonics; Solar cells

Abstract

In this work, computational investigation of plasmonic nanostructures was conducted using the commercial finite element electromagnetics solver Ansys® HFSS. Arrays of silver toroid nanoparticles located on the surface of an amorphous silicon thin-film absorbing layer were studied for particle sizes ranging from 20 nm to 200 nm in outer diameter. Parametric optimization by calculating an approximation of the photocurrent enhancement due to the nanoparticles was performed to determine optimal surface coverage of the nanoparticles. A comparison was made between these optimized nanotoroid arrays and optimized nanosphere arrays based on spectral absorption enhancement and potential photocurrent enhancement in an amorphous silicon absorbing layer.

In addition to these nanotoroids, highly irregular nanostructures were investigated. These structures were inspired by surface structures that were observed by others in the literature to be forming during the top-down aluminum induced crystallization of amorphous silicon. A 3D model of these irregular nanostructures was studied considering the structure's material to range from pure aluminum to a weighted mix of aluminum and silicon. Absorption enhancement in the underlying silicon layer was calculated and multiple, broadband spectral resonant peaks were observed. Parallel computation based on the Message Passing Interface (MPI) of two HFSS parallelization methods was employed on the Arkansas High Performance Computing Center. A 2.6 times speedup and 34% reduction in memory requirements was achieved when using the domain decomposition scheme of the package as compared to the basic multiprocessing parallelization.

Download

Share

COinS