The maximum temperature that a geotechnical bentonite barrier in a deep geological repository for radioactive waste can withstand while maintaining its integrity and meeting safety requirements is still an open question. Therefore, an international consortium set up an in situ heater test (HotBENT experiment) at the Grimsel Test Site (GTS) in Switzerland at relevant scales and gradients with temperatures ranging from 175°C to 200°C at the heater/canister surface. After dismantling (5 and 20 years, respectively), the identification of bentonite alteration processes of (clay) minerals has to be based on the comparison of data with reference values determined before the heating started. The experiment was set up using ~150 tons of two different clays (Wyoming and BCV from the Czech Republic) provided in different batches. The bentonites were used both as compacted bentonite blocks and as granular bentonite material (GBM). The determination of representative mineralogical and geochemical bentonite reference values must be based on a significant number of samples taken from all parts of the experiment, which is presented here. Most of the compositional variability was close to the accuracy of the methods used. However, chemical, mineralogical and exchangeable cation analyses showed that different raw materials were used to produce the BCV top blocks. The Wyoming bentonite used is similar to MX80 bentonite in that it is dominated by Na-rich smectite, but the HotBENT material contains slightly more feldspar and zeolite and slightly less smectite. Overall, 55 samples were analysed from different parts of the experiment, providing a statistical basis for post-excavation investigations.