Surface chemistry of lanthanum chromite powders II. Multivariate data modelling of the isoelectric point by the use of surface composition data achieved from X-ray photoelectron spectroscopy measurements

 Electrophoretic mobility measurements and X-ray photoelectron spectroscopy (XPS) analysis have been performed on several lanthanum chromite powders with different dopants. Principal component analysis of both deconvoluted XPS data and differentiated overall XPS spectra showed a clustering of the powders. From loading plots it was seen that high amounts of La and O on the surface gave highest isoelectric points (IEP). Partial-least-squares, multivariate response modelling was used to calibrate the IEPs from both deconvoluted XPS data and from differentiated overall XPS spectra. The best model was obtained when second-order differentiated overall XPS spectra were used, with an average predictive error of pH ± 0.25. This is promising considering that the IEP has been determined with an accuracy of pH_IEP± 0.3. When deconvoluted data was used, the average predictive error rose to pH ± 1.1. It is therefore an advantage to use multivariate data analysis which is a nonsubjective latent variable decomposing technique in contrast to deconvolution which is an even more time-consuming method for calibration of IEP values from XPS. Received: 8 October 1998 Accepted in revised form: 27 January 1999