Rapid micron-resolution quantitative elemental mapping is possible at the University of Surrey using a combination of proton induced X-ray emission tomography (PIXE-T) and simultaneous on/off-axis scanning transmission ion microscopy-tomography (STIM-T). A preliminary analysis of hair was performed. However, experimental uncertainties lead to large errors in tomograms and this work focuses on identifying and reducing the sources of error in both tomographic and 2D mapping. The STIM-T counts per pixel are used to normalise the PIXE-T data for charge. However, the geometry of the collimator and the scattering foil affects the detection rate since the loss of protons in the collimator increases as energy loss increases due to scattering. Errors in the PIXE geometric efficiency are greater in mapping when the detector is close to the sample. Moreover when a ‘funny’ filter was used for PIXE-T the uncertainty in the efficiency was found to increase because the sample-filter distance changes during the experiment.