Glycine and its hydrated complexes: a matrix isolation infrared study.

The hydration of glycine is investigated by comparing the structures of bare glycine to its hydrated complexes, glycine.H(2)O and glycine.(H(2)O)(2). The Fourier transform infrared spectra of glycine and glycine.water complexes, embedded in Ar matrices at 12 K, have been recorded and the results were compared to density functional theory (DFT) calculations. An initial comparison of the experimental spectra was made to the harmonic infrared spectra of putative structures calculated at the MPW1PW91/6-311++G(d,p) level of theory. The results suggest that bare glycine adopts a C(s) symmetry structure (G-1), where the hydrogens of the amino NH(2) hydrogen-bond intramolecularly with the carboxylic acid C horizontal lineO oxygen. Also observed as minor constituents are the next two lowest-energy structures, one in which the carboxylic acid (O-)H group hydrogen-bonds to the amino NH(2) group (G-2), and the other where intramolecular hydrogen bonding occurs between the NH(2) and the carboxylic acid O(-H) groups (G-3). The abundances of these structures are estimated at 84%, 9% and 8%, respectively. The least favored structure, G-3, can be eliminated by annealing the matrix to 35 K. Addition of the first water molecule to G-1 takes place at the carboxylic acid group, with simultaneous hydrogen bonding of the water molecule to the carboxylic acid (C=)O and (O-)H. The results are consistent with the predominance of this structure, although there is evidence for a small amount of a hydrated G-2 structure. Addition of the second water molecule is less definitive, as only a small number of intense infrared modes can be unambiguously assigned to glycine.(H(2)O)(2). Anharmonic frequency calculations based on second-order vibrational perturbation theory have also been carried out. It is shown that such calculations can generate improved estimates (i.e., approximately 2%) of the experimental frequencies for glycine and glycine.H(2)O, provided that the potential energy surfaces are modeled with high-level ab initio approaches (MP2/aug-cc-pVDZ).