Paper SS-ThA8
Wetting Monolayer Structure on Metals

Thursday, October 21, 2010, 4:20 pm, Room Picuris

The √37 and √39 rotated H₂O-molecule adlayers that form at low T on Pt(111) have been a mystery since 1997. [1] Their structures must optimize among : 1) forming as many H-bonds as possible, 2) allowing the maximum number of H₂O’s to lie flat, with O-atoms in atop sites, and 3) minimizing strain in bond lengths and angles. But, 1) rules out vacancy structures, 2) begs why rotation, leaving few O atoms in atop sites, would be preferred, and 3) is troublesome in the models proposed in Ref. 1, given their areal compression relative to ordinary ice.

Haq, et al. spun off a clue in an effort to interpret a spectroscopically observed metastable water monolayer on Ru(0001). Flat-lying molecules that have each other as neighbors can lie closer to the metal than in the conventionally imagined “ice-like” adlayer, where adjacent molecules’ dangling H-bonds force the flat molecules up. [2] Thus, they proposed a water layer on Ru(0001) comprising chains of flat H₂O molecules separated by molecules whose dangling H-bonds point toward the metal. Elaborating, we find that a checkerboard of compact flat-lying and H-down phases is even somewhat better bound, and this, together with new, telling STM images, has led us to plausible, low-energy structures for water on Pt(111).

STM images of these structures show √37 and √39 arrays of dark triangles separated by ribbons of H₂O-molecule hexagons. [3] We assign the triangles to di-interstitial molecular “defects,” wherein six flat-lying H₂O’s, with O atoms ~2.2Å directly atop Pt’s, anchor the water layer to the metal. In DFT calculations, these “defect” structures, containing three 5-member and three 7-member rings of H₂0 molecules, have lower energy than any purely hexagonal mesh. A competing explanation of the dark triangles as vacancy islands [4] is significantly less bound, because of its many dangling bonds. Remarkably, analogous 555777 structures on Ru(0001), never observed, are better bound than any purely hexagonal structure tried to date.

[1] A. Glebov, et al., J. Chem. Phys. 106, 9382(1997).

[2] S. Haq, et al., Phys. Rev. B73, 115414 (2006).

[3] S. Nie, et al, unpublished.

[4] S. Standop, et al., unpublished.

*Work supported by the DOE Office of Basic Energy Sciences, Div. of Mat. Sci. and Eng., under contract DE-AC04-94AL85000.