Multivariate statistical analysis of non-mass-selected ToF-SIMS data

Cluster LMIGs are now regarded as the standard primary ion guns on time-of-flight secondary ion mass spectrometers (ToF-SIMS). The ToF-SIMS analyst typically selects a bombarding species (cluster size and charge) to be used for material analysis. Using standard data collection protocols where the analyst uses only a single primary bombarding species, only a fraction of the ion-beam current generated by the LMIG is used. In this work, we demonstrate for the first time that it is possible to perform ToF-SIMS analysis when all of the primary ion intensity (clusters) are used; we refer to this new data analysis mode as non-mass-selected (NMS) analysis. Since each of the bombarding species has a different mass-to-charge ratio, they strike the sample at different times, and as a result, each of the bombarding species generates a spectrum. The resulting NMS ToF-SIMS spectrum contains contributions from each of the bombarding species that are shifted in time. NMS spectra are incredibly complicated and would be difficult, if not impossible, to analyze using univariate methodology. We will demonstrate that automated multivariate statistical analysis (MVSA) tools are capable of rapidly converting the complicated NMS data sets into a handful of chemical components (represented by both spectra and images) that are easier to interpret since each component spectrum represents a unique and simpler chemistry. Copyright (C) 2008 John Wiley & Sons, Ltd.