Quantitative Fiber Mixture Analysis by Scanning Electron Microscopy

Abstract

Labeling textile blends requires identification and quantification of their fibrous components. Blends of specialty animal fibers with sheep's wool are of special, practical importance; for these the light microscope is the traditional tool of analysis. To investigate the actual applicability of light microscopy for analyzing such blends as an alternative to the scanning electron microscope (SEM), we analyzed in detail the results of round trials conducted in the seventies. The results confirm that light microscopy, in general, is neither an objective nor a reproducible method for analyzing wool/specialty fiber blends. Though there was substantial progress with subsequent round trials, the data suggest that there is a fundamental statistical limit to the pass/fail rate, i.e., the ratio of correct versus incorrect analyses in a round trial that can be achieved by light microscopy. Even allowing for generous error limits, this effect leaves an intolerable element of chance for the correctness of analysis. Such performance is in pronounced contrast to that of the SEM method, where round trials have shown that laboratories that perform well reach analysis errors for specialty fiber/wool blends that are within or close to the natural error limits of microscopic analyses.