Two methods have been found by which unpaired electrons can be introduced into solid anthracene, namely doping with alkali-metals and photo-injection from a dry surface coating of rose bengal. Amongst the alkali-metals rubidium was the most effective.
In the solid, the alkali-metal anthracene complexes did not form solid solutions in anthracene, whereas in the melt, it was considered that the complexes were dissolved. The solid systems were frozen relatively rapidly so that the complexes were kept in a relatively dispersed state. The rate of freezing had no effect on the systems except for the CsAn system indicating that CsAn almost formed a solid solution.
In all the solid systems except CsAn, Heisenberg exchange and alkali-metal spin-orbit interaction were the main processes determining spin-lattice relaxation, g-shifts, line-shapes and -widths. In solid CsAn, motional narrowing augmented the effect of Heisenberg exchange.
In the melts, motional narrowing occurred and Heisenberg exchange was absent. Alkali-metal determined spin-lattice relaxation had less effect on the ESR signals than in the solids. Anthracene hyperfine interaction had more effect on the ESR signals in the melt than in the solid.
The exchange frequency in the solid and in the melt was estimated to be about 5 × 108 anthracene molecules per second.
The presence of spin correlation indicated that a chemical reaction took place to some extent between the alkali-metal and anthracene. Prior to the chemical reaction the electrons were simple donated doublets. Some other effects of the chemical reaction are recorded.
Trapping of electrons in an organic system could be studied by ESR by dispersing a potassium acceptor complex in solid anthracene.
The trap depth correlated with the electron affinity difference between the acceptor molecule and the anthracene molecule. This was deduced from the observation of a positive temperature dependence of the height of the ESR line in solid KAnDBA only and in molten KAnPent only.
A considerable increase in the degree of electron delocalization occurred on melting, which is consistent with the lower dependence of the line-width on spin-lattice relaxation in the melt than in the solid reported in Part 1.
A similar chemical reaction occurred in the systems of Part 2 as occurred in Part 1.
The ESR behaviour of three component systems made from other substances is predicted on the basis of the work done in this Part.
Excessive thickness of the dye coatings, poor adsorption of the dyes, association of dye molecules, the presence of iodine atoms in the dye molecules and the rigidity of the dye molecules were the factors that produced the low quantum yield from a dry coat rose bengal spectrally sensitized anthracene and the absence of observable sensitization by the other dyes.
The action spectrum effect of the sensitization was similar to the effect produced by dyes in aqueous solutions.
Trapping of charge carriers at the dye/anthracene interface was indicated by the relatively slow rise time of the transient photocurrents.
The low quantum yields were not due to space charge, or charge sheet limitation or due to an inadequate applied field.