Eliminating nonlinear Raman shift displacement between spectrometers via moving window fast Fourier transform cross-correlation

摘要

Obtaining consistent spectra using different spectrometers is of critical importance to fields that rely heavily on Raman spectroscopy. The quality of both qualitative and quantitative analysis depends on the stability of specific Raman peak shifts across instruments. Nonlinear drifts in the Raman shifts can, however, introduce additional complexity in model building, potentially even rendering a model impractical. Fortunately, various types of shift correction methods can be applied as preprocessing in order to address this problem. In this work, moving window fast Fourier transform cross-correlation is developed to correct nonlinear shifts in order to synchronize spectra obtained using different Raman instruments. The value of this method for synchronization is demonstrated using a series of Raman spectra of pharmaceuticals, as well as by comparison with the results obtained using an existing standard Raman shift scattering procedure. The results show that after the removal of shift displacements, the spectral consistency improves significantly, increasing the spectral correlation coefficient of the two Raman instruments from 0.87 to 0.95. The results demonstrate that the developed standardization method has, to a certain extent, reduced the instrumental systematic errors caused by measurement, while enhancing spectral compatibility and consistency through the simple and flexible moving window procedure.