2007 年 93 巻 9 号 p. 583-590
The Spark OES-PDA (Pulse height Distribution Analysis) method has been widely used in the field of steel making processes, as a rapid composition analysis of elements, such as aluminum in molten steel, according to the chemical states respectively: acid soluble (Sol.), insoluble (Insol.) and the Total (Insol.+Sol.). When a sample surface is sparked, several thousands of elemental emission intensities are stored and transformed to the frequency distribution of PDA histogram. In literature, symmetric normal distribution of low intensity pulses on the histogram was identified to be the acid soluble (Sol.) component in the metal, whereas the high intensity pulses were attributed to the acid insoluble inclusions (Insol.).
In this study, the collapse process of inclusions on the sample surface was investigated using optical microscope, SEM-EDS, EPMA, and spark-OES. It was found that the number of inclusions decreased sharply before the stable discharge region over 500 pulses, because the inclusions were initially destroyed and dispersed finely into the metal matrix. As the result, the emission intensities of the symmetric normal distribution region are superimposed on those of finely dispersed inclusions (Fine Insol.) and the elements dissolved in the metal (Sol.).
It is proposed in this study that symmetric normal distribution of the PDA histogram should be reassigned to the Total (Insol.+Sol.).