There is a need for scientific research that evaluates the influence of important process variables on the scale up of supercritical technology. For supercritical fluid extraction (SFE), one of these variables is the extractor's bed geometry, which can be defined by the ratio of the bed height (HB) to the bed diameter (DB). A systematic study is needed to select suitable criteria that can be used to obtain similar extraction curves among beds with different geometries. In this study, maintaining a constant ratio of solvent mass to feed mass for two beds with 1-L volumes but different geometries (E-1: HB/DB = 7.1; E-2: H-B/D-B = 2.7) was confirmed as a successful scale up criterion. For constant values of the temperature, pressure and bed porosity, there is experimental evidence that the mass transfer rate is equal in the two beds when the solvent flow rate is high. When 0.6 kg of clove buds was packed in the beds, the extraction rates were 2.10 +/- 0.08 and 2.3 +/- 0.1 g extract/min for beds E-1 and E-2, respectively. However, when the solvent flow rate was lower, the extraction rates were 0.93 +/- 0.06 and 1.12 +/- 0.02 g extract/min for beds E-1 and E-2, respectively. This difference in behavior between the extraction beds is associated with the axial dispersion of the fluid, which is more pronounced when the H-B/D-B ratio is increased. Thin particles tend to compact in the beds with high HB/DB ratios, which shorten the solvent passage. Non-isothermal profiles and differences in chemical composition of the extracts were also observed: 17% more alpha-humulene and 9% more eugenol were extracted in E-1 and E-2, respectively. (C) 2013 Elsevier B.V. All rights reserved.