During the production of stainless steel, a significant amount of by-products such as slags, dusts and mill scale are generated. These materials contain heavy metal compounds, which is why they generally have to be landfilled, although some of them have – compared to natural materials used in the construction sector – similar properties. Chromium is a widely used alloying element in stainless steel production and because of its high affinity for oxidation, chromium can be found in all oxidic by-products. Austria’s legislation limits the maximum chromium concentration in solid materials to 0.25 % for its application in road construction. The legislative limit for other construction sectors is even lower. This is the reason why slags from stainless steel production have to be treated in order to lower the chromium oxide concentration to less than 0.25 %. Treatment concepts can be divided into pyrometallurgical, hydrometallurgical and mechanical approaches. The most promising one is the carbothermic reduction of a liquid slag because it can extract metal inclusions as well as elements in oxidic form; hence, pyrometallurgical methods yield low final heavy metal concentrations in the treated materials when done properly. In order to achieve high recovery rates, all oxidic compounds have to be fully dissolved in a liquid slag. Thermodynamic calculations of the theoretical equilibrium between chromium and chromium oxide in a carbothermic reduction process showed that the process temperature has to be at least 1,600 °C. Additionally, the oxide activity and slag viscosity are assumed to have an influence on the achievable final concentration. Small-scale trials that were performed in an induction furnace pointed out that the final concentrations depend on the treatment duration, which is why the whole treatment process was assumed to be limited kinetically. Therefore, kinetic investigations were performed to get information about the necessary treatment time dependent on the process temperature. Since the process temperature has a major influence on the reaction rate, an exact control was necessary. For this task, a temperature controller was developed. The evaluation of trials in the melting lab led to the development of a kinetic model that describes the reaction rate for a specific chromium oxide concentration and temperature. This can be used to estimate the necessary time to reach a desired concentration for a given starting concentration. Experiments in the EAF were performed to get information about the influence of the furnace type on the reduction kinetics. While the chromium concentration course showed a similar course – this means the speed-limiting step of the reaction mechanism seems to be comparable – an exact comparison was not possible, because the furnace lacks an accurate temperature control system.