Development of methods for precise online-evaluation of Raman spectra
Spectroscopic methods as, e.g., Raman spectroscopy allow for the quantitative evaluation of the composition of multi-component mixtures. Therefore, Raman spectroscopy is an interesting option for measuring physical properties of multi-component systems as well as for process monitoring.
The basis for the quantitative analysis of Raman spectra is the proportionality between the intensity of one component in the mixture and the concentration of the same component. In multi-component mixtures, the components interact (e.g., through H-bonds). Therefore, the spectrum of each single component is a function of the composition of the mixture. Hence, a simple superposition of pure substance spectra may lead to quite poor results.
Instead of a simple superposition of pure substance spectra, the method of Indirect Hard Modeling (IHM) can be used. In IHM, models of the pure substance spectra are superimposed. These models are sums of parametrized peak functions (e.g., Gaussian, Lorentzian and Voigt peak functions) with parameters for shape, height and position of the peaks. To model the concentration dependent spectra, some peak-parameters are chosen to be adaptable. Thereby, for example the peak position of a moving peak is fitted within the superposition of the pure substance models.
By now, Indirect Hard Modeling is successfully used for more than 10 years. The method was developed in cooperation with the chair of Process Systems Engineering (now AVT.SVT) and the Chair of Technical Thermodynamics. A commercial implementation of IHM is available in the meantime.
The methods for spectra evaluation can be refined to expand the usability of Raman spectroscopy and explore new industrial applications. We work on calibration of reactive systems, robust evaluation of spectra with large fluorescence backgrounds and the even more precise modelling of mixture spectra.