A chemicurrent is a flux of fast (kinetic energy տ 0.5؊1.3 eV) metal electrons caused by moderately exothermic (1؊3 eV) chemical reactions over high work function (4؊6 eV) metal surfaces. In this report, the relation between chemicurrent and surface chemistry is elucidated with a combination of top-down phenomenology and bottom-up atomic-scale modeling. Examination of catalytic CO oxidation, an example which exhibits a chemicurrent, reveals 3 constituents of this relation: The localization of some conduction electrons to the surface via a reduction reaction, 0.5 O2 ؉ ␦e ؊ 3 O ␦ ؊ (Red); the delocalization of some surface electrons into a conduction band in an oxidation reaction, O ␦ ؊ ؉ CO 3 CO 2 ␦ ؊ 3 CO2 ؉ ␦e ؊ (Ox); and relaxation without charge transfer (Rel). Juxtaposition of Red, Ox, and Rel produces a daunting variety of metal electronic excitations, but only those that originate from CO2 reactive desorption are long-range and fast enough to dominate the chemicurrent. The chemicurrent yield depends on the universality class of the desorption process and the distribution of the desorption thresholds. This analysis implies a power-law relation with exponent 2.66 between the chemicurrent and the heat of adsorption, which is consistent with experimental findings for a range of systems. This picture also applies to other oxidationreduction reactions over high work function metal surfaces.heterogeneous catalysis ͉ hot electrons ͉ metal surfaces ͉ surface science ͉ transition metals T he significance of electronic excitations on metal surfaces was probably first recognized within the photoelectric effect (1, 2) and has been reaffirmed ever since. Photoelectrons are emitted into the vacuum from metal surfaces upon exposure to light, the energy quantum of which exceeds an exit thresholdthe work function. Just as light ejects photoelectrons that map the surface, chemical reactions of exothermicity exceeding the work function eject exoelectrons (3) that map the chemistry on low work function (Շ3 eV) metal surfaces. Low work function metal substrates (e.g., groups IA, IIA, and IIIA) are too reactive to be practical for catalysis unlike their noble or late transition d-metal counterparts, which support rich chemistry, often of industrial significance. These more interesting substrates have work functions of 4Ϫ6 eV (4), too high for exoelectrons to emerge. But electronic excitations in metals are omnipresent: Any surface movement invokes, by virtue of Anderson's orthogonality catastrophe, sub-work function electronic excitations (5) that stay inside the metal-unseen unless looked for-and ultimately dissolve in the sea of thermal conduction electrons, phonons, and photons.Experimental evidence of internal metal excitations associated with chemistry on high work function metal surfaces has emerged only recently. A pattern in quenching of NO molecules ro-vibrationally excited by 2.9Ϫ3.8 eV near an Au(111) surface (6) alludes to the involvement of sub-work function electronic excitations. Involvement of the Au(111) surface...