PHOTOXIDATION OF WATER BY DYESTUFFS 77 



to bring about the oxidation of water. However, the primary process in 

 dyestuff solutions is excitation, and not an electron transfer from water 

 to dyestuff (as assumed above for the inorganic cations, Ce++++ and 

 Fe+++). In this case, the oxidation of water must be brought about by 

 a secondary electron transfer from water to the excited dyestuff ion. 

 Processes of this kind are known to occur between excited dyestuff ions 

 and other electron donors, as, for example, ferrous ions. As shown by 

 Weber (1931), Weiss (1935) and Rabinowitch (1940), excited thionine or 

 methylene blue cations oxidize ferrous ions, even though in the dark the 

 reaction proceeds in the opposite direction, in accordance with the posi- 

 tions of the normal oxidation-reduction potentials: 



light 



(4.17) Thionine + 2 Fe++ , leucothionine + 2 Fe+++ 



For example, at pH 3, the normal potential of the system thionme- 

 leucothionine is approximately - 0.3 volt, while that of the system 

 Fe+++-Fe++ is approximately - 0.75 volt. Nevertheless, in light, ferrous 

 ions are oxidized by thionine ions, and it takes the system several seconds 

 to come back to equilibrium in the dark. 



The slowness of the back reaction may be attributed to a peculiar 

 relation between AH and AF in reaction (4.17). The normal potentials 

 indicate that the free energy of this reaction is strongly positive; but 

 its heat effect probably is negative. The free energies of hydrogenation 

 of most organic systems, including thionine, are less negative than the 

 total energies— that is, the reduced state has a smaller entropy. The 

 relation is reversed in the case of the reduction of ferric ions by hydrogen. 

 Consequently, the reversal of reaction (4.17) is an endothermal reaction, 

 and as such cannot proceed with a high velocity. 



This more or less accidental circumstance is the explanation why, in 

 the thionine-iron system, the shift in the oxidation-reduction equilibrium 

 by light, which usually is hidden by rapid back reactions, becomes easily 

 observable, even though it remains transient. 



It can be asked whether, in the absence of ferrous ions, a reversible 

 reaction does not occur between dye and the solvent (even if with a 

 smaller quantum yield and with a more rapid back reaction). Such 

 "hidden" oxidation-reductions have been held responsible for the photo- 

 voltaic effect of dyestuff-coated electrodes by Audubert and coworkers, 

 Hoang Thi Nga (1935) and Stora (1935, 1936, 1937). In the same way, 

 the directly observable reversible reduction of thionine by ferrous ions 

 has been shown by Rabinowitch (1940^) to produce a strong photo- 

 galvanic effect. 



If oxygen is present in aqueous dyestuff solutions, one could expect 

 some of the leuco dye formed by the oxidation of water to be reoxidized 



