Towards quantitative chemical imaging with ToF-SIMS

Methods for processing ToF-SIMS spectra and images have advanced significantly in recent years to include multivariate analysis methods. Multivariate methods can reduce noise, identify and quantify chemical components, and segment images into discrete chemical phases. To date, these methods have focused on the analysis of single images; however, quantitative or semi-quantitative methods for comparison of multiple images collected across multiple samples have lagged in development. This study evaluates simple noise reduction and scaling methods to facilitate semi-quantitative comparison of images collected across several samples with varying acquisition conditions. Down-binning, Poisson-scaling, and nearest-neighbor smoothing methods improved signal-to-noise in image datasets, with nearest-neighbor smoothing providing the greatest improvement. Image scaling methods including pixel-by-pixel (PbP) normalization and scalar multiplication were found to improve the relative quantification of images. While PbP normalization methods performed well for relatively flat samples, such methods were not suitable for samples with significant topography. Scaling methods using scalar multiplication of individual secondary ion images and histogram analyses facilitated semi-quantitative comparison of these samples.