Because of its structural specificity, rotational spectroscopy has great potential as an analytical tool for characterizing the chemical composition of complex gas mixtures. However, disentangling the individual molecular constituents of a rotational spectrum, especially if many of the lines are entirely new or unknown, remains challenging. In this paper, we describe an empirical approach that combines the complementary strengths of two techniques, broadband chirped-pulse Fourier transform microwave spectroscopy and narrowband cavity Fourier transform microwave spectroscopy, to characterize and assign lines. This procedure, called microwave spectral taxonomy, involves acquiring a broadband rotational spectrum of a rich mixture, categorizing individual lines based on their relative intensities under series of assays, and finally, linking rotational transitions of individual chemical compounds within each category using double resonance techniques. The power of this procedure is demonstrated for two test cases: a stable molecule with a rich spectrum, 3,4-difluorobenzaldehyde, and products formed in an electrical discharge through a dilute mixture of C2H2 and CS2, in which spectral taxonomy has enabled the identification of propynethial, HC(S)CCH.
