Organic light-emitting diodes (OLEDs) are attractive for display applications because of their high brightness, low driving voltage, and tunable color. Their operating lifetimes, hundreds or thousands of hours, are sufficient for only a limited range of applications. The luminance efficiency decreases gradually as the device is operated (electrically aged), for reasons that are poorly understood. A prototypical OLED has the structure anode|HTL|ETL|cathode, where the HTL and ETL are hole- and electron-transporting layers, and the recombination and emission occur at or near the HTL|ETL interface. We find that the decreasing luminance efficiency is linearly correlated with an accumulation of immobile positive charge at the HTL|ETL interface, and the magnitude of the charge is comparable to the total charge at that interface when an unaged device is operated. A natural explanation of the connection between the two phenomena is that electrical aging either generates hole traps (and trapped holes) or drives metal ions into the device, and that either species act as nonradiative recombination centers. To estimate the accumulating immobile charge and determine its location, we use a variant of a recently introduced capacitance versus voltage technique. In the prototypical OLEDs described here, the HTL is a ca. 1000 Å layer of NPB, and the ETL is a 300−1800 Å layer of Alq3. A device with an additional “emission layer” (EML) of an anthracene derivative between the HTL and ETL, in which the electroluminescence spectrum is characteristic of the EML, behaved similarly. We surmise that the phenomena reported here may be common to a wider variety of OLED structures and compositions.

Enormous effort has been expendedIn the past toward the development of highly specific Ion-selective electrodes. However, we have further developed the case that the use of an array of sparingly selective electrodes can result In multicomponent analyses that are superior, In many cases, to those obtained when the electrodes are highly specific. The requirement for a successful analysis then shifts from selectivity to stability and reproducible characterization of the sensors.

General and simplified digital simulation schemes are used to generate current-time and component potential-time curves and carrier concentration profiles. For reversible Interfaces, the analytical transient Is a good representation. However, experimental I-t curves require Introduction of potential-dependent Interfacial Ion-transfer kinetics to reproduce shorttime parts of the transients. Attempts to fit whole transients globally, with independently determined parameters (e.g., concentrations and the dc diffusion coefficient), are partially successful. Very good fits can be made by adjusting membrane concentrations and Interfacial kinetic parameters, which are experimentally uncertain. A striking new result Is that best fits require Ion transfer coefficients far from 0.5 (asymmetric barriers). This effect Is not unexpected since the applied voltage appears asymmetrically across back-to-back diffusion layers, with most potential drop In the membrane phase.

No abstract

Publication costs assisted by the National Science FoundationAn algorithm for computer simulation of transient concentration profiles and both transient and ac electrical properties of conducting membranes has been developed and described. Applications include transport by several ions with different mobilities and valences in permselective membranes, totally blocked cells, and finite galvanic cells. The algorithm is based on the Nernst-Planck system of equations coupled with Poisson's equation, introduced via the displacement current; therefore, electroneutrality is not assumed and migration effects are included automatically. Simulations of transient and steady-state properties as well as ac impedances are compared with solutions obtained analytically. Conventional linearization of field equations is not required to obtain transient responses. However, conversion from time to frequency using the Fourier transform to obtain impedances is theoretically limited to dimensionless currents and voltages less than unity. We have demonstrated versatility of the algorithm by simulating a case of biological interest-the inductive behavior of the squid giant axon.