As demonstrated by recent STM [1] and LEED [2] experiments the platinum (110) surface undergoes, at carbon monoxide submonolayer coverages, a phase transition from the 1 x 2 “missing-row” (reconstructed) state to the 1 x 1(bulk-like) state under specific temperature and partial-pressure conditions. The catalytic oxidation reaction CO + 1/2 → CO2 drives a microfaceting instability [3] [4] of the Pt(110) surface which ends up in a regular sawtooth profile with a period ≈ 200 Å, along the [110] direction.

We introduce the idea that the rather extensive Pt mass transport, as involved in the process, could be energetically assisted by the reaction itself. Energy and momentum-balance considerations lead us to expect an energy ≲ 0.5 eV to be transferrable to thesubstrate. This should efficiently contribute to initiating the “scraping”process that leads to the microfaceted pattern.

A simple model for nucleation and growth of facets is presented (see ref. 5), yielding characteristic times of order minutes (at T = 500 K), in fair agreement with experiment.

Independently of the structural/catalytic problem, adsorption of CO at submonolayer coverages on, e.g., Pt(110) might be of interest from a surfactantphysics point of view (see ref. 6 for a very recent study on layer-by-layer homoepitaxial metal growth).