Calcium looping CO2 capture systems use CaO as a reversible sorbent of CO2. Therefore, the evolution of the CO2 carrying capacity of CaO-materials at increasing number of carbonation-calcination needs to be determined to assess sorbent performance. Thermogravimetric analyzers (TGA) are commonly used for this purpose, by simulating around a small batch of material the average cyclic conditions expected in the real system. Many variables have been reported to influence the results and we review in this paper the main observations and trends, which can at times be conflicting when diffusional effects are not ruled out from the experiments. Furthermore, in a selected number of tests on a typical limestone using four different TG equipment, we have detected that some design characteristics of the TGA apparatus can strongly affect the determination of the CO2 carrying capacities of the material. In particular, we note that the decay in CO2 carrying capacity is accelerated as the power density of the TGA oven increases. This effect is most pronounced in the first calcination cycle, and it seems to be linked to an additional shrinking of the particles taking place in the TG apparatus with the highest heating rates. The use of larger sample masses and/or larger particle sizes tends to reduce the error in the determination of CO2 carrying capacity curves at the expense of departing from differential conditions that are required to obtain kinetic information on the sample. (C) 2013 Elsevier Ltd. All rights reserved.