Review of Chemometrics Applied to Spectroscopy: 1985-95, Part I

Abstract

Chemometrics can generally be described as the application of mathematical and statistical methods to 1) improve chemical measurement processes, and 2) extract more useful chemical information from chemical and physical measurement data. Recent advances in this discipline have led to a new breed of analytical tools– microprocessor controlled “intelligent” instrumentation. and data analysis systems. Chemometrics represents in the broadest sense the penetration of serious mathematical science into the realm of the chemist. Prior to the introduction of a formal subdiscipline of chemistry, termed chemometrics, chemists applied as much mathematics as they had available for research and problem solving, but in general, the use of statistical experimental design and data analysis were relegated to the engineer or the specialist statistician; chemists did chemistry, and life was simple. The current trend in scientific thinking across disciplines involves a multivariate approach. The world is not so simple as was once postulated, in fact it is common for scientists to approach problems realizing that there are deeply hidden relationships between variables that can be wrestled from an experiment only by the use of newer data analysis techniques; thus chemometrics becomes a necessity, not the luxury it once might have been. The purpose of this three-part review is to summarize the use of chemometrics in spectroscopy over the past 10 years. A survey of basic chemometric resources and techniques as applied to mass spectrometry, ultraviolet-visible spectrophotometry and infrared spectroscopy (Near-IR and mid-IR) will be reviewed. General techniques will be covered in Part 1 of this review, with a special emphasis on sample selection for calibration (teaching) sets and qualitative techniques. Part 2 of this review will cover quantitative analysis; with part 3 covering the more advanced chemometric methods, in greater detail.