AVS 49th International Symposium
    Applied Surface Science Tuesday Sessions
       Session AS-TuP

Paper AS-TuP4
Multivariate ToF-SIMS Image Analysis of Patterned Protein Surfaces

Tuesday, November 5, 2002, 5:30 pm, Room Exhibit Hall B2

Session: Topics in Applied Surface Science
Presenter: B. Wickes, University of Washington
Authors: B. Wickes, University of Washington
D.G. Castner, University of Washington
Correspondent: Click to Email
Novel biomaterial surfaces are being developed to specifically interact with their biological environments. These surfaces are patterned with multiple species of biomolecules to generate regions of differing bioactivity. The chemical structure of these surfaces must be characterized at high spatial resolution. Static Time-of-Flight Secondary Ion Mass Spectrometry (ToF-SIMS) imaging offers a modality for simultaneously visualizing the spatial distribution of multiple surface species. Because ToF-SIMS images yield a full mass spectrum at each pixel, it is possible to use characteristic mass fragments to identify and differentiate regions of different chemistry with a spatial resolution of approximately one micron. However, these datasets can be challenging to analyze because of their large size, complex chemistries and low ion counts per pixel. The combination of spectral data denoising and multivariate image analysis provides a convenient method to process ToF-SIMS images. Wavelet filtering followed by Principal Components Analysis (PCA) was applied to ToF-SIMS images. For example, the raw image data from a patterned poly(ethylene oxide) (PEO)-protein surface showed contrast in over 50 peaks; the resulting PCA model compressed the contrast from the raw data into two variables describing the variation between the protein- and PEO-regions, and the background and a defect region, respectively. Applying PCA to filtered image matrices removes user bias in peak selection and allows use of the full mass spectrum at each pixel. It highlights the peaks important in the chemical image and yields a new set of variables identifying the chemistries responsible for the image contrast.