Rotating screw conveyors can be utilized in the application of heat to granular materials: the mechanical motion of the rotating screw provides a controlled means of transferring the product through different thermal sections. This is particularly interesting for thermochemical biomass reactor applications such as slow pyrolysis and torrefaction. To understand the transfer of heat to the biomass, it is necessary to predict the residence time distribution of the material within the rotating screw reactor. Experiments were performed using a benchscale (2.5 kg/h) rotating screw conveyor. In this work, it has been shown that an empirical mathematical model is sufficient for predicting the mean residence time of various materials (pine chips of different sizes, rice and sand) if the rotational frequency of the screw conveyor and the volumetric throughput rate are both known. The model parameters obtained in this work were found to be independent of material selection but are specific to the geometry of the screw reactor. The form of the mathematical model proposed in this work may be applied to similar applications. Unlike the mean residence times, the axial dispersion calculated from the experimental data showed greater variability and this precluded any predictive modeling. (C) 2014 Elsevier B.V. All rights reserved.