Iron oxide in various forms can chemisorb significant amounts of water which cannot really just be dried with a towel or removed with mild temperatures. That the voltage effect shows up with the (easily oxidized) iron deposition layer but not without one does not really rule out, on its own, that water is not involved.
EDIT: by the way, on a related note:
Why not just use 240 V AC mains? Two Fe plate electrodes etc? For LENR reactors all we need is Holmlid's catalysts and a large chunk of reactor ingredients rather than going down to nanoscales. Thin wires etc etc.will not build up sufficient reactant protons or neutrons for creating a critical mass.. I despair of the ignorance here in basic nuclear physics.
Speaking of those iron oxide catalysts, in this open access paper the reduction of iron oxide (+potassium in some form) under hydrogen at a few mbar and increasing temperatures was studied: Reducibility of potassium-promoted iron oxide under hydrogen conditions (iastate.edu)
It was found that a weight reduction in the material upon heating observed up to 300–550 °C was mainly due to loss of surface hydroxyls, i.e. chemisorbed water as –OH groups. While this is for hematite for which this effect will be strong, water chemisorption will also occur with other iron oxide types, e.g. FeO. Here is one of the first hits on a web search just as a starting point: Water adsorption on an iron oxide surface - ScienceDirect .