Magnetic resonance studies of tris-(8-hydroxyquinoline) aluminum-based organic light-emitting devices

The electroluminescence (EL)-, electrical current density (J)-, and photoluminescence (PL)- detected magnetic resonance (ELDMR, EDMR, and PLDMR, respectively) of tris-(8-hydroxyquinoline) aluminum (Alq(3))-based organic light-emitting devices (OLEDs) and Alq(3) films is described. At low temperatures, a positive spin-1/2 resonance is observed, i.e., the changes in J, the EL intensity I-EL, and the PL intensity I-PL are positive (DeltaJ/J, DeltaI(EL)/I-EL, and DeltaI(PL)/I-PL>0). DeltaJ/J and DeltaI(EL)/I-EL are insensitive to the nature of the Alq(3)/cathode interface. They weaken with increasing T and become unobservable above 60 K. DeltaI(PL)/I-PL also decreases with T, but is still observable at 250 K. Since the resonances all have the same g value, similar linewidths, and a similar dependence on T and the excitation level (J or the laser power), they are all attributed to the same mechanism. That mechanism is either the reduction of singlet exciton (SE) quenching by a reduced population of polarons in the bulk of the Alq(3) layer ("the quenching mechanism"), or the enhanced formation of SEs from singlet polaron pairs at the expense of triplet excitons (TEs) ("the delayed PL mechanism"). However, the latter mechanism implies that the yield of SEs in Alq(3)-based OLEDs is greater than 25%. Due to evidence to the contrary, and other evidence which is inconsistent with the delayed PL mechanism, we conclude that the positive spin-1/2 resonance is due to the quenching mechanism. At Tapproximate to60 K, another spin-1/2 resonance, which reduces both J and I-EL (but is unobservable in the PL), emerges and grows with increasing T. This negative EDMR and ELDMR is sensitive to the buffer layer between Alq(3) and the cathode, and is attributed to the magnetic resonance enhancement of the spin-dependent formation of negative spinless bipolarons from spin-1/2 negative polarons at the organic/cathode interface. The increased trapping of injected electrons at the interface reduces J and consequently I-EL. However, at 295 K, the ratio \DeltaI(EL)/I-EL\ in Alq(3)/AlOx/Al devices to that in Alq(3)/CsF/Al devices is significantly lower than the ratio of \DeltaJ/J\ in these devices. Hence we suspect that other mechanisms, unidentified at this point, are also contributing to the negative ELDMR.