No description available.

In this study, the Jones matrix eigenanalysis (JME), optical spectrum analyzer (OSA) and polarization scrambling methods were used to investigate polarization dependent loss (PDL) in the presence of polarization mode dispersion (PMD) in optical components and fibres. The PDL measurements were conducted both in the laboratory and in the field. For field measurements, a buried link (28.8 km) and an aerial fibre (7.1 km) were extensively studied. The findings obtained from these studies are very important for network operators who must assess the impact of PDL on the network reliability. The three different PDL measurement methods (JME, OSA and polarization scrambling) were compared and their PDL values were found to agree very well at the selected wavelength of 1550 nm. Concatenation of PDL components showed that as expected, PDL increase as the number of PDL components were added. The interactions between PMD and PDL measurements were analyzed. A PMD/PDL emulator was constructed. We observed that PMD decreased while PDL increased. The PMD decrease was a result of the PMD vector cancellation enhanced by the randomly distributed mode coupling angles while PDL increase was a result of each PM fibre segments contributing to the overall global PDL. It was observed that the presence of PMD in a link containing PDL, results in PDL being wavelength dependent and this resulted in the extraction of the PMD information from the PDL data. PDL was found to be Maxwellian distributed when considering low values of PMD. High PMD values resulted in the PDL distribution deviating from Maxwellian. Long-term PDL and PMD (average DGD) measurements indicated that the PDL and PMD varied slowly with time and wavelength for both the laboratory and field measurements. It was observed that the BER increase as both PDL and PMD increased for simulated optical link.

Polarization mode dispersion (PMD) causes significant impairment in high bit-rate optical telecommunications systems. Knowledge of a fibre’s PMD mean value, and the relevant confidence interval, is essential for determining a fibre’s maximum allowable bit-rate. Various methods of confidence interval construction for time series data were tested in this dissertation using simulation. These included the autocovariance-matrix methods as suggested by Box and Jenkins, as well as the more practical and simpler batch means methods. Some of these methods were shown to be significantly better than the standard method of calculating confidence intervals for non time series data. The best of the tested methods were used on actual PMD data. The effect of using polarization scramblers was also tested.

In this study, two polarization effects, namely the state of polarization (SOP) and polarization mode dispersion (PMD) in optical fibre cable are investigated. The change in polarization effects introduces errors in optical fibre communication system. We find that the SOP drifts slowly in buried cables, and rapidly in aerial cables. This is because buried cables are located in a static environment, whereas aerial cables are exposed directly to a dynamic environment. The SOP change in aerial cable shows significant correlation with its environment (the global radiation, temperature and wind). The autocorrelation function (ACF) was not performed in buried cable, since they do not satisfy the ACF assumption, whereas in aerial cable it is found that the ACF of the SOP decorrelates quite quickly during the day. The 50 percent decorrelation time during the day and night are 9.6 and 30.4 minutes, respectively. During the day the properties of the optical fibre change rapidly as a result of the rapidly changing environmental conditions, whereas at night the environmental conditions change relatively slowly. Fast Fourier Transform (FFT) of the SOP fluctuations show discrete peaks, which corresponds to the wind induced vibrational frequency of the cable. The PMD fluctuations for undeployed and deployed aerial optical fibre cable are monitored using the generalized interferometric technique (GINTY). It is found that the PMD measured with polarization scrambling is more scattered but more reliable than the PMD measured without polarization scrambling. This is because the PMD obtained with polarization scrambling is averaged over different input and output (I/O) SOP pairs. For deployed aerial cable, it is found that the PMD measured without polarization scrambling fluctuates rapidly during high wind speed conditions. Furthermore, there is a correlation between the measured PMD and the change in temperature. It is found that the change in temperature has a stronger influence than the wind on the PMD of the optical fibre link.

The work presented in this dissertation is a comparative study of polarization mode dispersion (PMD) measurement performance where the fixed analyzer (FA) technique was built and tested for the first time in South Africa. Techniques involved in the study are: the Jones matrix eigenanalysis (JME), generalised interferometric technique (GINTY) and the FA technique, with a particular focus on the FA technique. The FA PMD measurement technique determines the average differential group delay (DGD) from the transmitted intensity spectrum through a polarizer and has three analysis methods (extrema counting, mean level crossing and Fourier analysis) which were all evaluated. PMD measurements were performed in the laboratory on several different fibre types and in the field on buried deployed Telkom fibre links (28.8 km). The techniques showed good agreement in the measured PMD value, both in the laboratory and field measurements. In particular very good agreement was found between the JME average DGD and the extrema counting analysis PMD value. The GINTY and FA Fourier analysis method also gave very similar PMD values. It was found that the fibre birefringence and the mode coupling manifest in different ways on the intensity spectrum. By using the FA ratio method, the length regimes of the different fibre types were determined. Three characteristics of the FA technique were investigated, namely: wavelength window variation, sampling and input SOP scrambling. It was found that the wavelength window and the PMD are inversely proportional. Correct sampling plays a significant role in determining the correct measured PMD value. Lastly an average PMD value over the PMD values for different input SOPs serves as a better representation of the true PMD value. An additional study showed that the FA technique and a developed Poincaré sphere analysis method agree very well regarding the PMD value.

In this study, the generalized and traditional interferometric techniques (GINTY and TINTY) are used to investigate some of the important aspects of Polarization Mode Dispersion (PMD) phenomenon in optical fibres. Measurements of PMD and chromatic dispersion (CD) are performed on buried optical fibre cable in the Pretoria Telkom’s metropolitan network. The upgradeability of this network was investigated, and it was observed that just over a half of the fibres are upgradeable to 10 Gb.s-1 transmissions. Long-haul aerial network in the Northern Cape was also tested for upgradeability to 10 Gb.s-1. It was found that 41 percent of the fibres tested are upgradeable to 10 Gb.s-1. Long-term monitoring showed that PMD varies rapidly in aerial fibres and, on the other hand, it is relatively stable in buried cables. Investigations on the accuracy of the techniques showed that polarization scrambling is essential for the reduction of the measurements uncertainties. Furthermore, it was observed that TINTY underestimates the single scan PMD distributions. The study on the effect of the change in mode coupling on various fibre configurations was performed, and fibres showed a reduction in PMD after the introduction of mode coupling. Measurements of PMD conducted in the laboratory on cabled fibre with low PMD showed the floor sensitivity of TINTY. Comparison between GINTY and Jones Matrix Eigeanalysis (JME) PMD measurements methods were performed on an emulator, and the results showed a good agreement in the measured PMD.

The scattering of laser light by laboratory plasmas has been observed. When the scattering was from a plasma formed in a θ-pinch, with scattering angle of 90° , a nearly Gaussian profile of the scattered intensity as a function of wavelengths was observed, corresponding to scattering by non-interacting electrons. When the scattering was from a plasma jet, with scattering angle of 45° from the forward direction distinct satellite peaks were observed on both sides of a narrow central peak at the laser frequency as predicted by theory, 1,2,3,4 indicating a strong collective scattering effect between the electrons and the ions. The widths of the satellite lines were greater than the values predicted by theory. The discrepancy is ascribed to spatial variations in the electron density in the volume of the observed plasma. The intensities and frequencies at-which they occur these peaks also vary with the current of the plasma jet in a manner consistent with theory. The scattered intensity of the central peak was measured approximately and it agrees with theoretical prediction. Some indication of perturbation of the plasma by the incident laser light has also been observed. 1. E.E; Salpeter, Phys. Rev. 120, 1528(1960). 2. J.A. Fejer, Can. J. Phys. 38, 1114(1960). 3. J.P, Dougherty and D.T. Farley, Pro. Roy. Soc. (London) A259, 79(1960). 4. M.N. Rosenbluth and N. Rostoker, Phys. Fluids 5., 776(1962). / Science, Faculty of / Physics and Astronomy, Department of / Graduate

In this thesis an attempt is made to understand certain experimental results in optical excitation of impurities in semi-conductors. These results are: i) the temperature- and concentration-dependence of the halfwidths of the absorption peaks and ii) the changes in the absorption lines when a compensating impurity is added. These effects are explained by studying the influence (on the neutral absorbing impurities) of the internal fields produced by the ionised impurities. For boron-doped silicon the first effect is due to the inhomogeneity in the field and the second to the quadratic Stark effect produced by the field. An analysis is also presented of the effect on the spectrum of a uniform applied field. / Science, Faculty of / Physics and Astronomy, Department of / Graduate

The polarization of light from helium atoms excited by the impact of low energy electrons has been measured for the spectral lines 2³P - 2³S (10,829 Å) and 3³P - 2³S (3889 Å). An electron beam carrying a current of 10μA was directed into helium gas at a pressure of 4 x 10⁻³ torr or less. Polarization was measured as a function of electron energy in a range from the excitation threshold (approximately 23 electron volts) to 50 e.v. For the 2³T - 2³S line, this work represents the first reported measurement of this type. There is special interest in the value of polarization near the excitation threshold. The theoretical threshold polarization for both lines studied in this thesis is 36.6%. In the experiment of this thesis, the observed polarization of the 2³P line rises to 21% near threshold, and by means of a curve fitting procedure may be extrapolated to 32 ± 6%. The polarization of the 3³P - 2³S line rises to 11% and may be extrapolated to 15 ± 3%. / Science, Faculty of / Physics and Astronomy, Department of / Graduate

Photoelectric mixing in a photodiode is used In this work as a statistical spectroscopic tool. A number of experiments were performed to determine the fluctuation spectrum generated by this process and the statistical properties of the light which might be deduced from the data. Due to practical limitations in attainable temperatures, blackbody sources were not able to produce an observable mixing above shot noise. Experiments were also carried out using line spectra from gas discharge lamps. The best source available was a 300 W Xenon lamp, emitting lines in the red, which under optimum conditions produced excess current fluctuations equal to 60% of shot noise. The observation of photoelectric mixing losing a gas laser source has already been reported in the literature, but the Gaussian distributed electric field model usually applied does not fit the experimental results. A new model was proposed in this thesis which considers the laser light as a narrow band of coherent light embedded in a relatively broad band of spontaneous light. Mixing between the signal and the spontaneous emission was considered to be the only observable effect due to experimental limitations. This model appeared to fit the data and gave some information about the statistical properties of the laser beam. / Science, Faculty of / Physics and Astronomy, Department of / Graduate