“…This could be a problem for nanometer-sized apertures, since 10 nm of glass are dissolved in 28 h. In fact, the interaction between water and the silica network is not limited to the surface, but can also occur behind the glass-water interface ͑Scholze, 1988͒, resulting in a loose, swelled, and unstructured layer at the surface of glass that is denoted as a silica gel. Compared to bulk glass, this gel layer has an enhanced ion mobility, lower electric resistance, different Si-O bond lengths, and faster dehydration ͑Wikby, 1974; Yaroshchuk, 2001a͒. The long-term stability and controllability of glass is therefore limited, but can be improved with surface pretreatment ͑Hench and Clark, 1978; Dilmore et al, 1979͒ or Schwarz and Schonhoff, 2002;Huebsch et al, 2004;Lulevich and Vinogradova, 2004;Poptoshev et al, 2004;Sui and Schlenoff, 2004;Lee et al, 2006͒.…”