화학공학소재연구정보센터
Journal of Chemical Physics, Vol.105, No.16, 6844-6855, 1996
The Motion of Protons in Water-Ammonia Clusters
The dynamics of dusters (H2OH+ (n=1,2,3,4) interacting with an NH3 molecule has been studied by first-principles Born-Oppenheimer molecular dynamics (BOMD) simulations. These small clusters are chosen as prototype systems for studying the mechanisms of proton transfer at atomistic level. We focus on the fundamental steps of proton motion in molecular clusters, the dynamical consequences of proton affinities, and the interplay between proton motion and proton affinity in these systems. A characteristic feature of the motion, the forming and breaking of O-H bonds in H3O+ is analyzed in detail. The transfer process is found to be consecutive along a quasi-one-dimensional channel. The umbrella mode in NH3 can easily be excited to direct the lone pair of the ammonia molecule to the water clusters. The hydronium ion, however, reorients mainly via rotation. When NH3 reaches one terminal water molecule of a protonated water cluster, the system undergoes a series of intermediate states in which the mobile protons travel within the water clusters, H3O+ transients are formed as protons approach individual water molecules. The lifetime of the H3O+ transient is 8-20 fs, or 1-3 vibrational periods of the O-H stretch mode. Proton 3 transfer is observed for n=1, 2, 3, although for n=3 NH4+(H2O)(3) is in existence with NH3(H2OH+. For n=4, NH3(H2OH+ is the dominant statistical configuration. Vibrational spectrum of NH3(H2OH+ is analyzed in detail. The features of the spectrum can be used, in principle, to probe the proton motion in the transition state region reactions. In these calculations, the electronic charge distribution is calculated concurrently with the nuclear dynamics. An analysis of isocharge density surfaces gives qualitative and quantitative descriptions of the dynamics of electronic redistribution. The BOMD is performed in the framework of density functional theory with local spin density and generalized gradient approximations.