A new library searching technique is reported that relies on Fourier transforms of infrared (IR) absorbance spectra. Searching in the time domain is shown to be more tolerant to noise than searches in the spectral domain and fewer points are required to encode the unique characteristics of the individual spectrum. This decreases both search time and library storage requirements. Additionally, time-domain classification of IR spectra is demonstrated and the possibility of indexing the library for a binary search is discussed.
