A Planar Violet Electroabsorption Modulator and Modeling of Electric Field Effects on Zinc Oxide Excitons

modulator;zinc oxide;exciton;ZnO;semiconductor;temperature dependent;absorption;electroabsorption;wide bandgap semiconductors;electrical field

Zhang, Xiyao ; John F. Muth, Committee Chair,David E. Aspnes, Committee Co-Chair,Robert J. Nemanich, Committee Member,Leda Lunardi, Committee Member,Zhang, Xiyao ; John F. Muth ; Committee Chair ; David E. Aspnes ; Committee Co-Chair ; Robert J. Nemanich ; Committee Member ; Leda Lunardi ; Committee Member

University:North Carolina State University

关键词: modulator;    zinc oxide;    exciton;    ZnO;    semiconductor;    temperature dependent;    absorption;    electroabsorption;    wide bandgap semiconductors;    electrical field;   

Others  :  https://repository.lib.ncsu.edu/bitstream/handle/1840.16/4660/etd.pdf?sequence=1&isAllowed=y

美国|英语

来源: null

【 摘 要 】

First principle electroabsorption calculations based on WTK spectral density theorem and Dow and Redfield theory were performed and used as a basis of a model to fit experimental electroabsorption data. Absorption measurements were taken from 4.5 K to 300 K on a c-plane ZnO thin film sample with a high-resolution spectrometer to obtain the temperature broadening linewidth. The free A and B exciton peaks and two major neutral donor bound excitons were observed. The calculated zero temperature electroabsorption spectrum from Dow and Redfield theory was convolved with Lorentzian and Gaussian temperature dependent linewidth to model these absorption spectra at various temperatures. It is found that Gaussian lineshape works better due to the strong electron-LO-phonon interaction of ZnO, especially at higher temperatures. Gaussian broadening parameters are then extracted and expressed as a function of temperature. Two material related coefficients in the broadening linewidth expression are the exciton-acoustic-phonon interaction strength = 79.6 ± 3 μeV⁄K and exciton-LO-phonon interaction strength= 242 ± 10 meV. This expression is independent of sample qualities, or is 'generalized' to accommodate the exciton band edge. A concept called "effective microfield intensity" was introduced to represent the crystalline quality. The microfield intensities of two ZnO samples with rocking curve FWHMs ˜ 0.25° and ˜ 0.42° are ˜ 1.3 105 and ˜ 2.26 105 kV⁄cm respectively. Buffer-assisted growth technique was applied to improve the quality of PLD grown ZnO on sapphire, with the optimal buffer condition of 8 nm thickness, 500 °C temperature and 35 mTorr pressure. The same technique can also be applied to MgxZn1-xO. ZnO thin films were deposited on ATO⁄ITO⁄glass substrate at 500 to 700 K, with higher temperature producing better films. XRD of these samples show highly c-oriented ZnO growth on the ATO substrate. TLM measurements shows that the contact resistivity of Ti⁄Au (50⁄150 nm) on Al-doped MgxZn1-xO is reduced by rapid thermal annealing (RTA) from ˜ 1.2 10-1 ΩΩcm2 to 4 10-2 ΩΩcm2. ZnO electroabsorption (EA) optical modulators were fabricated based on ZnO⁄ATO⁄ITO⁄glass structure. Two types of top electrodes, Ni semi-transparent electrode (TE) and conducting indium gallium zinc oxide (IGZO) were experimented with the IGZO-coated device reducing the insertion loss by ˜50%. In both device, I-V characteristics shows leakage less than 20 nA within the device operation voltage range. The DC percentage modulation of these devices have two peaks, over 40% near ˜ 370 nm and around 20% near ˜ 385 nm at 140 V bias. AC testing confirms purely field modulation and shows no evident of frequency cut-off up to 100 kHz. A simple device model attributes the threshold voltage of ZnO EA modulators to the charge screening effects caused by the free electrons in the ZnO active layer. Applied voltage was obtained both by converting the electric field based on device model and fitting the calculated EA spectra with the first principle calculation and Gaussian broadening. The field strengths obtained from the two approaches is consistent, with only a small proportional discrepancy of ˜ 30%.