Structure of the water ice surface studied by x-ray absorption spectroscopy at the O K-edge

Vapor-deposited ${\mathrm{H}}_2\mathrm{O}$ ice films grown between 38 and 150 K under ultrahigh vacuum conditions have been investigated using near-edge x-ray absorption fine structure (NEXAFS) spectroscopy at the oxygen K-edge, in conventional mode—which is bulk sensitive-, and using the photon-stimulated desorption mode (PSD-NEXAFS), which is surface sensitive. By recording simultaneously those two signals, we have evidenced the differences between the surface and bulk electronic and atomic structures, for both amorphous porous ice condensed at 40 K and crystalline ice condensed at 150 K. We have also followed the bulk and surface evolutions of an amorphous ice film annealed from 38 to 147 K. A steep change in the local atomic structure of the bulk is observed, likely related to the high-density amorphous $\mathrm{ice}\to \mathrm{low-density}$ amorphous ice phase transition between 38 and 55 K. We have shown that the surface of crystalline ice is well ordered, but this order is different from that of the bulk. We have evidenced that the ${\mathrm{H}}_2\mathrm{O}–{\mathrm{H}}_2\mathrm{O}$ intermolecular distance at the surface of ice is always longer than in the bulk, and that this difference increases with temperature, as the thermally induced reordering of the surface proceeds. SCF-Xα multiple scattering calculations allow us to figure out those structural differences, both in the bulk and at the surface of amorphous ice, but further calculations are necessary for crystalline ice. We have shown that the PSD-NEXAFS signal is sensitive to the surface morphology that changes with temperature because of the micropores collapse. We have used a model [E. Vichnevetski, A. D. Bass, and L. Sanche, J. Chem. Phys. 113, 3874 (2000)] that quantitatively describes the effect of the surface porosity on the ion yield. The surface of ice at 38 K is well described by a network of vertical cylindrical pores of 20 Å of diameter, separated by 6 Å, collapsing when annealing the film. This model also properly accounts of the peculiar temperature evolution of the PSD signal at the $\mathrm{O}{\text{1s}}^{\text{−1}}{\text{4a}}_1^{\text{+1}}$ excitation, and therefore, allows to establish the relation between the PSD-NEXAFS signal and the surface porosity.