Quantitative Spectroscopy of Gas Phase Reactions at Elevated Temperatures

Sammanfattning: Gas phase reactions at elevated temperatures are versatile tools for material manufacturing and energy conversion. They are of great commercial interest, but they often entail hazardous impacts on the environment. There is consequently a pronounced need for the further develpment of processes utilising these types of reactions. Spectroscopic methods, especially when combined with lasers, are ideally suited for this type of development, which is demonstrated in this thesis for diamond chemical vapor deposition in a thermal radio frequency plasma. Detailed theoretical studies and simulations for this type of environment have been reported in the literature, and experiments were carried out in order to assess these theories. Rayleigh thermometry and atomic hydrogen concentration mapping with two-photon laser-induced fluorescence were conducted in the reaction zone of the diamond deposit. Measured layer thicknesses and atomic hydrogen concentrations were found to be almost uncorrelated to each other, an observation which disagrees with predictions from the aforementioned theories. The theories also fail in predicting the growth rate and quality of diamond deposits obtained in our studies. Possible reasons for both theoretical shortcomings are discussed. A growth inhibiting process, which is promoted by highly excited atomic hydrogen, is proposed in order to explain the poorly predicted growth. The highly excited atomic hydrogen originates from a self-sustained secondary discharge that was discovered in the vicinity of the substrate. Established spectroscopic methods do not meet all the demands for development of gas phase reactions at elevated temperatures. Novel coherent methods hold potential in this respect, and one particular technique, polaristion spectroscopy, has been investigated in order to increase its applicability. One drawback of polarisation spectrsocopy is its mathematically cumbersome theory. A simple model of the signal generation process was therefore proposed and checked against detailed numerical simulations and against experiments. This model provides excellent predictions of saturation curves for the case of homogeneous line broadening, and an extension to the case of inhomogeneous line broadening is possible. The progress reported renders polarisation spectroscopy more applicable, and it also indicates the possibility of deriving a new theory that is appropriate and practical at the same time. Quantitative spectroscopy requires evaluation schemes that provide the maximum precision and accuracy attainable from measurement data. An evaluation scheme based on a proper statistical treatment of the measurement process is proposed in this thesis, and for the first time applied to development with pulsed lasers. These laser systems exhibit pulse-to-pulse intensity fluctuations that entail a statistical correlation of signal and laser intensity. Common evaluation schemes do not account for this correlation, which results in estimates of measurands featuring poor accuracy and precision. In contrast, the estimators derived in this thesis provide consistent estimates with excellent precision. The results obtained are extremely encouraging for further development of quantitative laser spectroscopy in this respect.