A method has been developed to investigate the sensitivity and photometric accuracy of a Fourier transform infrared (FT-IR) spectrometer by statistically analyzing a set of spectra of a 75 μm thick film of polyethylene terephthalate (PET). These tests were applied to three commercial FT-IR spectrometers. Although the photometric accuracy of all three instruments was generally better than 0.1%, small artifacts that could not be removed by signal averaging were often observed where strong bands should have a transmittance of less than 0.01%. Binding of the moving mirror of one of the spectrometers appeared to degrade its performance. When the standard deviation of the noise on 100% lines was calculated in the conventional way, none of the instruments gave efficient signal averaging after 25 scans had been coadded, because of small slopes in the baseline. By calculation of the noise level by successive difference, signal averaging was improved, but even with this approach the noise level on one of the instruments did not decrease after 100 scans had been averaged.
Bryan T. Bowie and Peter R. Griffiths, "Measurement of the Sensitivity and Photometric Accuracy of FT-IR Spectrometers," Appl. Spectrosc. 54, 1192-1202 (2000)