In vivo studies investigating the toxicity of carbon nanotubes (CNT) often rely on models to predict deposited dose, but the unusual properties of aerosolised CNT mean that a) existing models may not fully capture their behaviour, as there is little experimental data on CNT lung deposition with which to validate them, and b) some studies fail to characterise the aerosol sufficiently, leading to inaccuracies in model input parameters and thus potentially errors in the output.

Here we consider the second of these and investigate the response of a well-known particle deposition model to changing key aerosol parameters, using values covering the range expected for ‘broadly-spherical’ and ‘fibre-like’ CNT, showing that inaccuracies in these parameters, particle density in particular, could lead to large errors in estimated doses.