Automated identification of components in a chemical mixture utilizing multi-wavelength resonant-Raman spectroscopy and a Pearson correlation algorithm

In complex environments, the ability to identify the constituent chemicals within a mixture is extremely important. By utilizing a Pearson correlation algorithm to compare sets of multi‐wavelength resonance‐Raman signatures, we demonstrate the automated identification of chemicals within a mixture. Applying a linear mixture model, we are also able to estimate the fractional volumetric abundances contained therein. The multi‐wavelength resonance‐Raman signature used for identification is obtained by illuminating the unknown mixture with a series of 21 sequential laser wavelengths. This signature is then compared with the signatures of a set of known chemicals. By maximizing the Pearson correlation coefficient between the signature of the mixture and a weighted superposition of the signatures of the pure chemicals, we are able to determine the mixture components with 100% accuracy. The linear superposition of the selected chemicals, which minimizes the least squares distance between the signatures of the mixture, and its mathematical recreation determines the corresponding fraction, by volume, of each chemical within the mixture. Copyright © 2012 John Wiley & Sons, Ltd. By utilizing a Pearson correlation algorithm to compare sets of multi‐wavelength resonance‐Raman signatures, we demonstrate the automated identification of component chemicals within a mixture. Applying a linear mixture model, the superposition of the selected chemicals, which minimizes the least squares distance between the signatures of the mixture and its mathematical recreation determines the corresponding fraction, by volume, of each chemical within the mixture.