I would like to be able to overlay a variety of Theoretical IR spectrum with a single Experimental IR spectrum and be able to determine the best matching theoretical spectrum. Is there a way to do this quantitatively on excel, sigmaplot, or any other plotting program?
Attached is an example spectrum where the black line represents the experimental spectrum. The red and blue areas/lines represent two different theoretical spectra. Both provide reasonably good matching, so I would like to quantitatively determine which matches the best.
you can use the sigma fit or you can make a statistical correllation (fit a line)
The easiest way to compare the two fits I would suggest is: calculate the square of the difference between fit and actual spectrum for all points of the spectrum and sum up all these squared differences. The fit with the smallest sum of squared differences would be the better fit.
Please pay attention on the following paper as well, where we have performed a vaidation of a quantitative approach based on IR-spectroscopy accounting for exactly contribution of various mathematical methods for processing of spectroscopic patterns to mentioned above point (a) and (b): 1. Validation of reducing-difference procedure for the interpretation of non-polarized infrared spectra of n-component solid mixtures, B. Ivanova, D. Tsalev, M. Arnaudov, Talanta, 69, 2006, 822-828
Mr. Batoon,
The applicability of any theoretical approach, including chemometric dataprocessing of the spectroscopic patterns is evaluated accounting for: (a) the accuracy towards the band/frequency position, meaning are they are shifted significantly or not; and (b) accuracy towards integral intensity, meaning is it is influenced significantly. Because of the intensity ratio of pairs of bands is constant under given experimental conditions, including phases of the substances.
Please find an example, dealing with prediction of theoretical IR-spectrum of MgSO4 in the solid-state, predicted by the quantum chemical approach using crystallographic coordinates.
The intensity ratio of experimental pairs of bands is 0.3029, while the theoretical one is 0.2489. But it is important to highlight here that the experimental IR-spectrum has shown a broad IR-band, which perturbes the pure integral intensity of the used modes as well as there has obtained an artifact at 1159 cm-1. In this example the chemometric (mathematical approach) requests the involvment of this artifact in order to ensure a higher r2 value (0.99967) or a higher level of confidence. So that in this example, exactly, should take into consideration not only the complexity of the experimental data, but chemometrics carried out as a source of error, determining the integral intensity.
Given that in your case it seems qualitatively only that the first method can be regarded as more accurate one than the second approach.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication