Raman and infrared spectra of polycrystalline samples of four structurally related sugars, namely D-fructose, L-sorbose, 5-deoxy-D-threo-hexulose (5-deoxy-D-fructose), and D-arabinose are reported. Three regions of the spectrum have been studied: the OH stretching region 3650–3100 cm−1, the CH stretching region 3000–2800 cm−1, and the "fingerprint" region 1600–200 cm−1. Raman spectra in the region 200–10 cm−1 have also been recorded. Bands in the OH stretching region have been correlated with the [Formula: see text] distances obtained from recent neutron diffraction studies. The results are discussed in relation to hydrogen bonding and relative sweetness of these sugars.Detailed assignments are given for the bands in the CH stretching region, and it is shown that the spectra are very sensitive to small structural changes in the molecule. Spectra are presented for the "fingerprint" region and some characteristic "group" frequencies are noted for ketohexoses and aldopentoses.

Can. J. Chem. 60, 1857 (1982. 5-Deoxy-D-threo-hexulose ("5-deoxyfructose", "5-deoxysorbose") has been prepared in six steps from D-fructose. The reaction of 2,3-0-isopropylidene-P-D-fructopyranose with sulfuryl chloride afforded exclusively 5-chloro-5-deoxy-2,3-0-isopropylidene-a-L-sorbopyranose in the key step. "5-Deoxyfructose" exists only in the *C,(D) pyranoid form in solution, and was found to be much sweeter than L-sorbose and nearly as sweet as D-fructose. Comments on this unexpected sweetness result are given.

Abinitio SCF LCAO-MO calculations at the STO-3G level have been performed on β-D-fructopyranose (1) and α-L-sorbopyranose (2) using crystallographic data as the geometrical input. Molecular properties of 1 and 2 are discussed in terms of orbital energies, total energy, ionization potentials, Mulliken population analysis, and electrostatic potentials, with a particular emphasis on the possible consequences of these features as regards the sweet taste of these two ketoses. No correlation was found, for example, with the electrostatic, point-charge distribution since the calculated hydrogen-bonding abilities would lead to the prediction of 2 being sweeter than 1. On the other hand, non-bonded overlap populations between oxygen and hydroxyl-hydrogen atoms reveal the presence of intramolecular interactions, which may have a determinant influence on the taste of these molecules and which could explain why D-fructose is much sweeter than its epimer at C-5, namely, L-sorbose.