| This thesis reports on two major activities and the results obtained with them. The first is the conception, design, fabrication and testing of a facile double-channel digital system for obtaining quantitative noise spectra. It consists of four main sub- systems. One is a shielded enclosure containing the device being tested, a probe station for contacting the device and a blocking capacitor to pass frequencies in the 1 Hz to 100 kHz region. The second major component is a pair of Stanford Research SR560 Low Noise Amplifiers. Then, another shielded box contains the 200 kSample/second, 16 bit analog-to-digital converter 9223 from National Instruments in a Compact DAQ chassis. Finally, a laptop computer is employed to store the twenty million data points obtained in each 100 second run, and to convert such time series into spectra.;MATLAB was used for the bulk of the computations, and LabVIEW was employed for some operations. Fast Fourier Transform algorithms in MATLAB were used to obtain single channel spectra or cross correlated spectra from the dual channels, each spectrum containing ten million points in the 1 to 105 Hz frequency range. Cross correlation removed some of the LNA and ADC noise and improved the noise floor of the system by somewhat less than a factor of 10. The single channel noise floor of the system is 10-17 V2/Hz, which is competitive with most of the systems described in the literature. However, our system is much less expensive and faster than earlier approaches. It has the potential to be a commercial product. The system development and testing part of the work was highly successful.;The second part of the work was to perform exploratory measurements with the new system on several types of solar cells. The fundamental idea is that both carrier transport through solar cells and noise spectra depend on geometry, composition, and defects in solar cells. Hence, noise spectra offer the possibility of providing information on solar cell performance and defects. Silicon, Gallium Arsenide and more complex compound semiconductor solar cells, mostly with single junctions but some with multiple junctions, were measured under various conditions. They included varying illumination with the cell either not loaded or connected to a resistor load. The triple junction complex compound semiconductor cells had been irradiated with heavy doses of protons, so noise spectra were obtained as a function of the proton fluences. Many differences were observed in the dozens of measured spectra. They included the appearance of some lines of unknown origin, but those lines do not degrade the utility of the noise spectra. Exhaustive parametric studies in future work should show that noise spectra are a useful tool for characterization of solar cells. The possibility of using noise measurements for on-orbit assessment of solar cell radiation degradation was considered.;A creative combination of hardware and software has produced the GWU Noise Spectroscopy System. It is a low cost, user-friendly, rapid, non-destructive digital capability to acquire and exploit noise spectra. The very low noise floor and broad frequency range of the system are both important. The new capability has a remarkably wide range of applications in physics, chemistry, biology, brain research, medicine, materials, chemical processing, mechanical and structural engineering, and acoustics.;To demonstrate additional applications of the entire new system, noise spectra were measured from a new MEMS microphone and from batteries with different chemical systems. Spectra quite different from those of solar cell were obtained. The analysis part of the system was employed to produce spectra from time series gotten from electrical probes inserted into a live rat brain. Very unusual spectra resulted. The results indicate that such spectra can be obtained in real-time during neural probe experiments. |
PDF Full Text Request