Mitigating dead‐time effects during multivariate analysis of ToF‐SIMS spectral images

Abstract

a †ToF-SIMS spectra are formed by bombarding a surface with a pulse of primary ions and detecting the resultant ionized surface species using a time-of-flight mass spectrometer. Typically, the detector is a time-to-digital converter. Once an ion is detected using such detectors, the detector becomes insensitive to the arrival of additional ions for a period termed as the (detector) dead-time. Under commonly used ToF-SIMS data acquisition conditions, the time interval over which ions arising from a single chemical species reach the detector is on the order of the detector dead-time. Thus, only the first ion reaching the detector at any given mass is counted. The event registered by the data acquisition system, then, is the arrival of one or more ions at the detector. This behavior causes ToF-SIMS data to violate, in the general case, the assumption of linear additivity that underlies many multivariate statistical analysis techniques. In this article, we show that high-mass-resolution ToF-SIMS spectral-image data follow a generalized linear model, and we propose a data transformation and scaling procedure that enables such data sets to be successfully analyzed using standard methods of multivariate image analysis.

Exploiting spatial‐domain simplicity in spectral image analysis

“…[30] This data set is denoted Dataset A in the reference, and the total-ion image is shown in Fig. 1.…”

“…Some of the low-mass data were severely limited by detector dead-time effects, and the entire data set was corrected for this according to a published procedure. [30] Our initial expectation is that the sample can be described by three pure components: the two acids and aluminum. Clearly, something unexpected occurred in the palmitic acid region, however, as there is a distinct difference between the inner and outer portions of the spot as shown in Fig.…”

“…Prior to multivariate analysis, each data matrix was scaled to account for nonuniform noise, either Poisson- [32] or binomially distributed, [30] as appropriate. SVDs were computed using a freely available code based on the Lanczos algorithm.…”

“…Additional details about the MCR-ALS calculation are presented in the Appendix. After completion of the various matrix factorizations, the derived components were inversely scaled [30,32] to return them to the physical domain for interpretation.…”

Image and Spectral Processing for ToF-SIMS Analysis of Biological Materials

“…MCR was carried out using the MCR-ALS gui (Roma Tauler and Anna de Juan, University of Barcelona, http://www.mcrals.info/) accessed through the Imagegui. PCA and MCR were carried out using the Poisson scaled 57,58) and mean centered data. PCA scores were used as initial estimates for the MCR analysis.…”