Mechanism of the Thermal Reaction Between Hydrogen and Oxygen

Abstract

Explosion limits and reaction rates of hydrogen and oxygen have been measured in spherical quartz and Pyrex vessels of varying diameter, clean and coated with various substances and for various temperatures, pressures, and mixture compositions, including addition of inert gases. Clean and B2O3-coated surfaces give rise to rapid and erratic reaction, indicating a surface chain-breaking efficiency ε≪λ/d (ratio of mean free path to vessel diameter); the reaction is self-accelerating, probably due to poisoning of the surface by H2O, which decreases ε. By coating with various salts such as KCl, BaCl2, K2B2O4, K2B4O7, and Na2WO4, the condition ε≫λ/d is established in the region between second and third explosion limits. The limits are farther apart, the rates much lower and not accelerating, and both rates and limits are reproducible and identical for the various salts. For K2B4O7, ε≃λ/d for small reaction rates. The chain-breaking mechanism on clean and salt-coated surfaces is discussed. With the elimination of the surface as a variable by coating with salts, data of explosion limits and rates can be used to derive the mechanism of the reaction. From a critical analysis of all imaginable reactions of the system and the present and earlier experimental results, the following mechanism is deduced. i.H2O2+M=2OH1.OH+H2=H2O+H2.H+O2=OH+O3.O+H2=OH+H6.H+O2+M=HO2+M11.HO2+H2=H2O2+H12.2HO2→ lim surfaceH2O2+O25.H+O2+H2O2=H2O+O2+OH7.HO2+H2O2=H2O+O2+OH13.H2O2→ lim surfaceH2O+12O214.H2+O2→ lim surfaceH2O2H,O,OH→ lim surfacedestructionThe surface destruction of H, O, and OH is important only at the first explosion limit and depends on the nature of the salt. Reaction 2 leads to chain branching by the formation of two new free valences. It is counteracted by reaction 6. H2O is a much more efficient third body than any other molecule investigated and, therefore, powerfully suppresses chain branching. Hence, explosions and reaction rates near the explosion limits are strongly inhibited by H2O, either added or formed during the reaction. If H2O is formed during the preparation of the mixture in the reaction vessel, the rates and limits measured may be very different from the values in the H2O-free mixtures, particularly near the junction of second and third explosion limits. An opposite effect is caused by the temperature rise in rapid reaction noticeable near the third limit above the junction. The superposition of both effects explains the deviation of the actual explosion boundary from that predicted by the isothermal explosion condition for H2O-free mixtures. In the range free from these influences, rates and explosion limits over wide ranges of all variables are well represented by the kinetic equations derived from the above mechanism after suitable numerical constants are introduced. A tentative equation of the course of the reaction in uncoated vessels assuming constant surface structure has been derived, which is similar to the empirical equation of Chirkov. The conclusion of Oldenberg and Sommers that the reaction is unbranched is shown to be invalid due to masking of branching by H2O in their experiments. Values of activation energies and rate coefficients of all reactions except 1 and 3 and concentrations of HO2 and H are given.