AbstractFrom quantitative data reported by various investigators, it is concluded that the primary passive film on iron has the following properties: (1) 25–100 Å thick, (2) equivalent to 0.01 coulomb/cm2, (3) can oxidize CrO−2to CrO−−4, (4) follows the (Flade) potential relation —0.58 + .058pH(volts).All these properties cannot be accounted for by any of the known oxides of iron. In better accord with the facts is a proposed passive film composed of chemisorbed atomic and molecular oxygen supplemented perhaps by OH and H2O. The free energy of formation is calculated from potential data of Franck to be — 29,900 cal/mol adsorbed O in agreement with independent data for gaseous oxygen chemisorbed on iron surfaces (— 30,500 cal/mol adsorbed O). Thermodynamic calculations show that the film is unstable, moreso in acids than in alkalies.The same kind of film apparently forms on Cr—Fe alloys and on Cr. It is calculated that passivity for these metals is also unstable in acids or alkalies when metal ions or hydroxides are reaction products of the decomposed film. However, spontaneous formation of the adsorbed passive film can occur on exposure of Cr or>20% Cr—Fe alloys to water, contrary to the situation for iron or the low Cr‐Fe alloys. It is possible to show that the passive film on iron is able to oxidize CrO−2to CrO−−4as observed, but that this reaction is not thermodynamically possible for the passive film on Cr or on>12% Cr—Fe alloys.A passive film of chemisorbed oxygen fulfills the conditions of good electrical conductivity and freedom from pores. It is extremely thin (4 to 15 Å) in agreement with Tronstad's optical measurements for the passive film on stainless steels (10 Å). Tronstad's measurements for iron, corrected for a ferric nitrate film forming continuously by corrosion of passive iron in nitric acid, would probably also agree with this lower value. The chemisorbed film decomposes to Fe2O3more than 60 Å thick based on apparent area, which agrees with values reported by others