Determination of ultra-low carbon and nitrogen contents in steel: combustion versus electrical spark source optical emission spectrometry for steelmaking process control

Abstract

Electrical spark optical emission spectrometry (spark-OES) was evaluated for carbon and nitrogen trace (<30 µg g−1) analysis in steel in comparison with traditional combustion methods associated with infrared detection. The development of an easily practicable process control method is described, including steel sample quality requirements, sample preparation optimization, and spectrometer operating conditions definition. In comparison with traditional combustion methods, spark-OES can decrease analysis delays by 2 min, with a corresponding reduction in the number of potential sources of error, and reduced analytical costs. Concerning the analytical figures of optimized spark-OES, limits of detection of 1 µg g−1 and 4.5 µg g−1 for C and N, respectively, have been demonstrated, with a precision in the range 1–2 µg g−1 for both elements at a level of 25 µg g−1 in steel. These figures are comparable to those obtained by use of combustion methods. Mean analytical differences between spark-OES and combustion analysis were evaluated to 0.6 µg g−1 and 1.8 µg g−1 for C and N, respectively, typically within both methods’ scatter. A successful transfer of ultra-low carbon spectrometric determination to five industrial process control laboratories has been carried out and is discussed. The organization of a round-robin for C determination in these laboratories resulted in an inter-laboratory scattering of 2.6 µg g−1.