A New Algorithm for Baseline Subtraction for Spectra

In spectroscopic characterization studies of materials, due to various reasons, there may be on many occasions, baseline shifting. This is very pronounced in case of infrared (IR) spectroscopy of heavily compounded carbon black rubbers. For such a shifted baseline, or for any other uncorrected baseline, the exact characteristic peak heights of materials in a spectrum, which is necessary for quantitative analysis, cannot be obtained. To obtain the exact peak heights, a spectrum needs to be modified through baseline creation on and subsequent subtraction from the original spectrum. In the present work, a newly developed numerical algorithmic method is devised to bring such a modification. Instead of a smooth baseline fitting over the entire range of data, the algorithm focuses on finding a baseline separately for each peak. It is assumed that at the base of each peak, which mathematically consists of two local minima, the baseline is almost a straight line. Therefore, an array of paired points is separately fitted at the base of all peaks and these points are then joined by straight lines to obtain the dataset of the new baseline. The ordinate values of the new dataset are then subtracted from those of the original dataset to obtain the dataset of the modified spectrum.