A mathematical model which describes the process when a spherical metallic particle is exposed to an oxidizing atmosphere, has been formulated. We refer to the oxidation process in a generic sense, and the model can be used for the formation of oxides, nitrides and sulphides in general. The oxidation process is modeled for oxides which form n-type semiconductors with the defects in the form of interstitial cations and electrons (e.g., Al2O3) in terms of mass and energy transport phenomena.It accounts for the release of metal cations and electrons at the metal-oxide interface, the transport of these species through the oxide layer due to concentration and eleclric field gradients, and the formation of the metal oxide at the outer interface. As a result a hollow oxide sphere develops as the metal is converted 10 oxide. If the energy released by the oxidation reaction exceeds the heat losses, self-heating will occur. Such a rise in temperature accelerates the release of cations and electrons from the metal into the oxide as well as the transport of cations. This can eventually lead to the total combustion of the particle. The model is solved numerically by the Galerkin Finite Element Method. The effects of several process variables are investigated for the aluminum-oxygen case study: initial passivating oxide layer thickness, average particle size, storage temperature and oxidant partial pressure, and heat losses. For certain limiting conditions, a simplified model is developed, and an approximate analytical treatment is presented.