Analysis of high resolution FTIR spectra from synchrotron sources using evolutionary algorithms

Room temperature Fourier transform infrared spectra of the four-membered heterocycle trimethylene sulfide were collected with a resolution of 0.00096 cm(-1) using synchrotron radiation from the Canadian Light Source from 500 to 560 cm(-1). The in-plane ring deformation mode (nu(13)) at similar to 529 cm(-1) exhibits dense rotational structure due to the presence of ring inversion tunneling and leads to a doubling of all transitions. Preliminary analysis of the experimental spectrum was pursued via traditional methods involving assignment of quantum numbers to individual transitions in order to conduct least squares fitting to determine the spectroscopic parameters. Following this approach, the assignment of 2358 transitions led to the experimental determination of an effective Hamiltonian. This model describes transitions in the P and R branches to J' = 60 and K-a' = 10 that connect the tunneling split ground and vibrationally excited states of the nu(13) band although a small number of low intensity features remained unassigned. The use of evolutionary algorithms (EA) for automated assignment was explored in tandem and yielded a set of spectroscopic constants that re-create this complex experimental spectrum to a similar degree. The EA routine was also applied to the previously well-understood ring puckering vibration of another four-membered ring, azetidine (Zaporozan et al., 2010). This test provided further evidence of the robust nature of the EA method when applied to spectra for which the underlying physics is well understood. (C) 2015 Elsevier Inc. All rights reserved.