Research On Instrument Standardization Of Raman Spectra With Its Applications To Pharmaceuticals

Last few decades, Raman spectrum, as a molecular spectroscopy analysis technology, has witnessed a rapid development and wide application. Because of its simple, fast, efficient and non-destructive, which makes it more and more popular in our daily life. Raman spectrometers are becoming more and more intelligent and humane with the development of chemometrics and computer science and technology. In this study, by reducing the systematical errors of different spectrometers to correct the Raman shift displacements in order to improve the consistency of Raman spectra between different instruments and realize the sharing of spectra model established on one instrument, which called instrumental standardization, sometimes namely instrumental transfer as well. The detailed research of this study was done on the following contents:1) The several Raman shift standard chemical materials (e.g., cyanobenzene and cyclohexane, acetaminophen, etc.), recommended by National Institute of Standards and Technology were selected to correct the shifts, a large number of common commercial drugs (different manufacturers, agents, batches, etc.), has been collected from drug safety authorities and drugstores and so on. All the samples were measured on three Raman instruments, respectively, so that lay the foundation for subsequent algorithms research.2) As we known, the systematical errors between different spectrometers may cause the Raman shift drifts leading to decrease the consistency of Raman spectra. In this case, the continuous wavelet transform (CWT), Lorentzian function and polynomial fitting have been applied to the systematical errors to correct Raman shifts between instruments. This proposed method can obtain a curve to correct the shifts by using the identified peaks of standard chemical materials. As the result shown that this method can eliminate the influence of systematical errors to improve the consistency of Raman spectra significantly. Meanwhile, results of spectral library retrieval by a large number of real-world samples were verified the method effectiveness after instrument standardization.3) The moving window fast Fourier transform cross-correlation (MWFFT) was present to remove the systematical errors between different spectrometers when no standard chemical materials available. The usages of moving window strategy can estimate nonlinear shifts between spectra flexibly without peak detection to synchronize the peaks. As the result shown that this method can enhance spectral consistency for different instruments and increase compatibility with Raman spectral libraries as well. Moreover, the practicability of this method was well performed on pattern recognition.4) In order to differentiate the branded version and six generic drug manufacturers, Raman spectroscopy in combination with random forest method were used to distinguish the differences between the same drugs of different manufacturers efficiently. Especially, after optimization of parameters, the result shown that random forest (RF) characterized the differences very well. So one can draw a conclusion that the application of RF can make a great contribution to Raman spectroscopy for the drug fast screening efficiently and accurately.All in all, this study use different instrument standardization methods to eliminate the influence of systematical errors and improve the consistency of Raman spectra between instruments, meanwhile, its applications to pharmaceuticals camparing with the existed methods, shown that the robust, the significance and the practicability of the proposed method. Firstly, by using the Lorentzian function to identify peak shift more precisely. Secondly, polynomial fitting can correct the nonlinear shifts. Thirdly, the cross-correlation is present to calculate the similarity between the corresponding peaks, when no standard chemical materials available. Finally, moving window steps can fit the shifts more accurate. Moreover, with the development of chemometrics methods, the Raman spectroscopy can be efficient using in drug fast screening.