In our group, we work on the surface modification of cellulose nanocrystals. During this work, we have encountered reproducibility issues when the same reactions were performed on nanocrystals from different hydrolysis batches, indicating a variable surface composition. Given the inherent purity of the nanoparticles themselves, this issue was believed to be due to the presence of adsorbed species at the surface of the nanocrystals blocking reactive sites. To investigate this in detail, nanocrystals from several batches were extracted with different solvents. The effect of these extractions on the surface composition of the nanowhiskers was investigated, followed by its effect on the Surface-Initiated RingOpening Polymerization (SI-ROP) of e-caprolactone. The extracted impurities were analysed by NMR ( 1 H and 13 C) and MS, showing a variety of adsorbed species which can be removed by solvent extraction.A Soxhlet extraction using ethanol before the reaction was shown to be the most effective in removing adsorbed low molecular weight organic compounds produced during the hydrolysis, resulting in improved reproducibility between reactions using nanocrystals from different batches, as confirmed by FTIR, elemental analysis and XPS. Extraction with ethanol should thus be performed before all reaction as these adsorbed species can be expected to interfere with all surface modification reactions.