Communications
ADVANCED MATERIALSfact, the devices could not be operated in air owing to significant degradation as a function of oxygen and/or water partial pressure. Whereas storage in air hardly changed the I-V characteristic (e.g., 40 % reduction of conductivity after 3 weeks for EC4T) and luminance or output power, operation in air led to a rapid degradation. For example, the output power decayed by a factor of five within 10 s for EC6T and within 20 s for EC7T; EC4T and ECST could not even be measured. Thus, we operated the LEDs in an inert N, atmosphere, which provided relatively stable conditions such that EC7T and EC6T showed little degradation (e.g., 25 Yo) after 15 min of operation at 2 mA. The reason for the rapid degradation in air (and also the slight degradation in a nitrogen atmosphere) is most likely oxidation of the Al electrode, which increases the barrier height and hence reduces the injection probability of electrons. The question remains why different oligothiophenes cause different degradation behavior. At present we can think of two possible answers: a) a different degree of dissociation or follow-up reactions of the oligothiophenes, perhaps at the inner electrode surface, in oxidizing atmosphere and in the presence of a current and/or electric field, and b) different diffusion velocities of 0, and/or H,O through the thin molecular film. The thermal molecular motion induced by the current may be responsible for rapid diffusion in operation while the different film morphologies (see above) may have caused the differences amongst the oligomers. More work is needed to give a clearer answer and to solve the problem of poor LED stability.Finally, we should briefly comment on the output power, which was measured by a calibrated laser power meter and is also displayed in Figure 2 for EC6T (dash-dotted curve, right hand scale). The light emission starts at voltages as low as 2.5 V (for a 160 nm thick EC6T film) and reaches a maximum around 8 V, as can be derived from the inset, which shows the power conversion (not quantum) efficiency curve. This curve varies relatively little (factor of two) as a function of voltage but has only a modest maximum of 4 x
YO.(The value for EC7T is even smaller by a factor of four; the value for ECST is markedly higher but could not be measured with sufficient accuracy.) These values have not been corrected for geometric effects, for internal reflection and absorption, and for internal energy conversion processes. If estimated corrections are taken into account the values become considerably higher (by about a factor of 100) and the quantum efficiency may be 1 0~2 -1 0 -3 % . This is still very low and far from being satisfactory, in particular in view of applications. We have some evidence that the low efficiency has various causes, such as film morphology, electrode material with too high a work function, interface bonding and reactions, and the physical properties of the molecules. We believe that it can be improved by orders of magnitude if materials and preparat...