A convenient model has been developed which qualitatively describes the heat and oxygen mass transport within a TGA sample crucible. The major conclusion from this model and from experimental observations was that for exothermal reactions, heat- and mass-transfer limitations can significantly influence the results of thermogravimetrical analysis. Therefore, it is crucial to use small, diluted samples, in particular for quantitative thermo-analytical measurements. It is recommended that the occurrence of heat- and mass-transport limitations be assessed by variation of the sample mass by at least a factor of ten. The model provides an explanation for the double-peak-shaped curves which are often observed when studying the catalysed oxidation of a model soot. The first peak is attributed to a thermal runaway reaction, which occurs when the sample is relatively large and when it is not diluted with silicon carbide. It is shown that the occurrence of the two peaks is caused by a combination of heat- and mass-transport limitations within the TGA sample.