Kinetics of mass transfer during nitrogen injection in industrial Fe—Cr—Ni—Mo‐melts

Experiments on the injection of nitrogen into Fe—Cr—Ni—Mo alloys in 30 t VOD‐Iadles have been carried out. The nitrogen contents obtained by nitrogen injection were between 0.05 and 0.17% for an initial nitrogen content of about 0.02%. Calculations of mass transfer associated with a rising nitrogen bubble indicate very high mass transfer rates for different nitrogen activities. The high efficiency of nitrogen gas measured in industrial practice agrees with these calculations. The specific interface area increases proportionally with the gas flow rate. According to reports in literature, nitrogen mass transfer can be described either as a 1st or 2nd order reaction. Both possibilities have been examined and the experimental results evaluated accordingly. In order to make possible a comparison with literature rate constants were first determined using comparable methods, i.e. on the assumption that the concentration difference is the driving force for mass transfer. The rate constants depend on the composition of the alloy. Assuming spherical cap bubbles with an equivalent diameter of 2.5 cm, the mass transfer coefficient was estimated to be about 5 to 8.10 −3 cm s −1 , using a 1st order reaction model; the apparent activation energy then has a value of about 100 kJ/Mol. If a 2nd order reaction model is assumed, the apparent activation energy has a value of about 125 kJ/Mol. Using the activity difference as the driving force for mass transfer, it was found that the kinetics of mass transfer in multi component alloys could be treated in a homogeneous manner. The mass transfer coefficient calculated in this way for a 1st order reaction lies between about 26 to 37.10 −3 cms −1 and the actual activation energy has a value of about 25 kJ/Mol. The evaluation of the experimental results using a second order reaction model gives an equally good correlation between the rate constants and the gas flow rate as that obtained using a 1st order reaction model. In this case the actual activation energies determined are slightly higher ‐ 40 kJ/Mol. However, the activation energies in both cases (i.e. with or without regard to the nitrogen activity) do not lie in the range of activation energies associated with a process controlled by interface reactions. Due to the low concentrations, it is not possible to make a definite statement about the order of the reaction. However, the activation energies determined fit better to a 1st order reaction model. Relations have been derived which, assuming a 1 st order reaction, and taking into account the composition of the alloy, enable prior calculation of suitable parameters for nitrogen injections in industrial melts. This allows a more precise process control.