Development of a Nitrogen Atmosphere Pretreatment Chamber for the Determination of Trace Amounts of Carbon in High-Purity Iron by the Infrared Absorption Method after Combustion in an Electric Resistance Furnace

The characteristics of high-purity metals have been widely investigated. It has become apparent that 99.999% (5N) iron differs from 99.95% (3N5) iron in their physical and chemical properties.' A new technique for the determination of trace elements is required so that the impurities in high-purity iron can be estimated more accurately. When trace amounts of gas elements (C, S, 0, N, H) are determined, methods for a decrease in the blank values are needed to obtain better reliability of the analytical values. For the determination of carbon in metal samples, an infrared-absorption method after combustion in an oxygen atmosphere has been used. When an electric resistance furnace is used for combustion, samples can be subjected to combustion under any temperature conditions independent of the type or shape of samples and different carbon species can be distinguished by analyzing the time-dependent peaks. However, the structure of the electric tube furnace allows gaseous carbon species in the air, which are represented by carbon dioxide or hydrocarbons, to enter the combustion furnace when a sample is inserted (the authors have called "air contamination").2 Carbon species on the surface of samples exert a serious influence on the analytical values.3,4 In the determination of trace amounts of carbon, these carbon species have to be considered as blank values. A method in which samples are baked before the determinations and a method in which additional absorption peaks are separated6 have been reported for decreasing the blank values. The authors made a nitrogen atmosphere chamber in front of the sample-introduction entrance of the electric resistance furnace, and investigated the performance of this chamber. When this chamber is used, only factors that originate from the apparatus can be removed in various blank factors. Then, the carbon values on the surface and in the bulk of samples can be respectively observed.