Pros and cons of using correlation versus multivariate algorithms for material identification via handheld spectroscopy

niques for rapid identification of unknown compounds (such as testing of fine chemicals, measurement of pharmaceutical ingredients, or authentication of drug compounds) is Raman spectroscopy. The economic and technical benefits of handheld Raman spectrometers have been well discussed in open literature, but one area where there is confusion for many users of this technology (novice and experienced alike) is in regards to the different statistical algorithms which are used for on board analysis of spectra. In this article, we will discuss the two most common mathematical representations used with handheld Raman spectroscopy as decision-making tools for spectroscopic data: Hit Quality Index (HQI) and significance level (p-value). Generally, HQI is the preferred method for library matching of unknown materials, and p-value is best suited for verifying the identity of a known material. Here we will discuss specific examples for the use of each tool. Library matching Library matching is a well-established method in spectroscopy for the investigation of unknown materials and is commonly used for identifica tion of materials from an FTIR, NIR or Raman spectrum. This is typically performed by crosscorrelating the measured spectrum of a material against a validated library of spectra of known materials. The degree of correlation (similarity) of each potential match is then quantified by a calculation of HQI defined by: