Sensitive detection of near-infrared (NIR) spectra of several organic liquids has been carried out by surface plasmon resonance (SPR) NIR spectroscopy. For all the liquids, 50- to 100-fold enhancements of the absorption peaks were obtained in the combination band region 4500–4000 cm<sup>−1</sup> using a gold film with a thickness of 14 nm. The SPR peak shows up as an unnecessary broadband peak or trend in an SPR-NIR spectrum, and it was difficult to separate it from the absorption signals. In order to remove the contribution of SPR from the raw SPR-NIR spectrum, the second-order derivative has been employed. The second derivative of the SPR-NIR spectrum was reasonably comparable to that of the corresponding transmittance spectrum. Two simple algorithms for sample identification from the second-derivative data have been proposed. One is similarity, which directly compares the second-derivative spectrum of an unknown sample with that of a known reference sample. The other is fitness, which is defined as a ratio of the common part of absorption peak wavenumbers of the sample and the reference. Although both methods are unfit for the identification of a minor component in a mixture, a major component can be definitely identified by choosing an informative wavenumber region. It was found that the wavenumber region 4250–4080 cm<sup>−1</sup> is especially useful for the identification of similar molecules such as normal alkanes.

PDF Article

Cited By

OSA participates in Crossref's Cited-By Linking service. Citing articles from OSA journals and other participating publishers are listed here.