Laser induced emission spectrometry for rapid elemental analysis.

This thesis reports investigations into laser-induced emission spectrometry for rapid elemental analysis. An integrated laser and spectrometer system is configured, comprising a Q-switched Nd:YAG laser and an optical multichannel analyser, which enables the discreet monitoring of the laser-induced plasma produced by each laser shot. Novel applications are devised including the survey analysis of polymeric materials and the depth profiling of coated steels. A survey analysis of polymeric materials for twelve elements (Al, Ba, Ca, Cu, Fe, Mg, Pb, P, Sb, Sn, Ti and Zn) is reported. Results showed that element emission responses are dependent upon operating conditions, selective volatilisation of antimony can occur, and ablation characteristics, such as the rate of material removal and plasma lifetime, are very different compared to metals. With optimised operating conditions of low laser energy and sample positioned at the laser focal point, the limit of detection for antimony is 0.09 % mass/mass with precision of 1.8 (% relative standard deviation) using a carbon signal from the polymer as internal standard. Rapid discrimination between samples of poly (vinyl chloride) is demonstrated with a measurement time of 1 s. Data for the depth profiling of a range of coated steels using laser-induced plasma emission spectrometry are reported in detail for the first time. Coatings of zinc/nickel, tin, titanium nitride and chromium are examined. Depth profile signatures and crater shape are greatly influenced by operating conditions. Improved depth resolution and signatures are obtained using high laser energy and defocused laser radiation. Correlations are established between coating thickness and output parameters of the technique. Linear calibrations against coating thickness are achieved with good precision for replicate measurements (4 % relative standard deviation). Results showed that the technique can differentiate between tin coated samples with a difference in coating thickness of 0.02 pm, and can detect an ultra-thin chromium coating of 0.020 pm thickness. This performance and measurement times of 50 s indicate the rapid depth profile capability of the technique. The novel application of an artificial intelligence technique (artificial neural networks) to laser-induced plasma emission spectra is reported for the first time. Studies showed that artificial neural networks can rapidly discriminate between the emission spectra of different materials with a success rate of 100 %, to provide a new means of rapid data processing.