In this work, we present an original numerical approach developed to evaluate injection performances of a new injection system (More s.r.1.(R)) designed for siderurgic applications. The system exploits a supersonic jet of oxygen to inject carbon particles into the slag. A precise characterization of the injection process by experimental analysis is extremely difficult and costly because of the complex chemico-physical mechanisms controlling transport, burn-out and devolatilization of carbon particles inside the oxidizing, high temperature environment of the electric arc furnace. In this work, we use numerical simulations to test and characterize injector performances for conditions corresponding to a 120 ton capacity electric furnace. We exploit the best available, state of the art numerical techniques to characterize the fluid-dynamics and chemico-thermal environment seen by carbon particles, which we couple to ad hoc research tools (Lagrangian tracking routines and complex chemistry schemes) to reproduce carbon consumption due to thermally and chemically controlled kinetics. These data are used to analyse the factors controlling injector performances, to identify a most critical configuration of the injector in the furnace and to obtain a conservative estimate of the injection yield of carbon particles. The performances of the injection device are evaluated in two different geometries and for three different types of carbon particles. Numerical results confirm that the supersonic injection promotes high injection yields: (i) by decreasing drastically the residence time of carbon particles inside the furnace and (ii) by modifying the hot reacting environment seen by carbon particles. (c) 2007 Elsevier Ltd. All rights reserved.