Whatever the method used for the synthesis of carbon nanotubes (CNTs), they always contain residual catalysts in variable amount. Many methods have been proposed in the literature to purify CNTs, but their efficiency strongly depends on the experimental conditions. Although the presence of residual catalysts in small amount is generally not a problem for many applications, this can become a critical issue when a high purity is required, typically for magnetic properties or for biomedical applications (because of the intrinsic toxicity of most catalysts). Quantification of the amount of residual catalysts is usually obtained by classical chemical analysis, which requires a preliminary digestion (complete mineralisation) of the CNT samples. In this work, we systematically compared 3 different digestion protocols and optimised one, reaching 100% dissolution within a very limited time (1 h) together with the requirement of only a few milligrams of sample, and safe experimental conditions. This method can be easily transferred for use in research laboratories, making accessible the quantitative analysis of CNT samples, and has been validated following ISO/IEC 17025:2005 for linearity, specificity, intermediate precision, limits of detection and quantification.