Authors' response

The comments by Pei and Heeger on our paper published in Nature Materials1 call for a clarification of the mechanism of the operation of light-emitting electrochemical cells (LEECs).

The simplest configuration of an organic light-emitting diode (OLED) consists of an organic semiconductor layer sandwiched between two electrodes, the anode and the cathode. The anode is selected to have a high work function in order to minimize the barrier for hole injection to the highest occupied molecular orbital of the organic layer (the equivalent of the valence band). Equivalently, a low-work-function cathode is required for efficient electron injection to the lowest unoccupied molecular orbital of the organic layer (the equivalent of the conduction band). Charge injection creates oxidized and reduced forms of the organic semiconductor molecules near the anode and cathode, respectively, and the different optical properties of these species have been used to measure charge densities in devices. Some of the injected charges recombine to produce light emission, and the efficiency of OLEDs is critically dependent on the ability of the electrodes to inject electrons and holes in a balanced manner2. Low-work-function metals such as calcium make great cathodes to most organic semiconductors, but are not air-stable. When a high-work-function metal such as gold is used as the cathode, the large energy barrier prohibits electron injection3, and such devices are called 'hole only'. By definition, an operating OLED is not at electrochemical equilibrium.

A way to work around energy barriers and improve charge injection involves the introduction of mobile ions to the organic semiconductor film. These devices are called light-emitting electrochemical cells, after the first report of a device using a semiconducting polymer layer with ions introduced in it4. On the application of a bias, mobile ions inside an organic film respond by accumulating near the electrodes (double-layer formation). This ion redistribution screens the electric field in the bulk and leads to interfacial electric fields that enhance charge injection. The ions may redistribute further to accommodate changes in the local electric field caused by the injected electrons and holes, but electric fields remain present at the contacts to maintain charge injection. The bulk remains essentially field-free until the applied bias becomes too high for ions to be able to completely screen it. This essential physics, which is the basis of the electrodynamic model5,6, is missing from the electrochemical model, which postulates that no electric field is present at the contacts4. Potential drops at contacts can in principle be accounted for by introducing an 'overpotential' into the electrochemical model, but such generalizations do not help understand the operation mechanism of LEECs.

The question is how significant are these interfacial electric fields in LEECs? The answer depends on the size of the energy barriers, the applied bias, and the details of the ions' distribution near the electrodes. The latter is hard to predict, as it depends on the availability of 'free volume' near the electrodes. Our paper1 reported the first direct measurement of electric-field distribution in an LEEC and found that, at steady state, interfacial fields are indeed significant. The measurements were carried out in LEECs with two high-work-function electrodes, so a larger electric field was observed near the cathode and a smaller one near the anode, as predicted by the electrodynamic model. At 5 V applied across the 7.5-μm device, the electric field near the cathode was 50 kV cm−1, whereas the bulk remained essentially field-free. Only when the bias reached 120 V (and the field near the cathode reached 800 kV cm−1) were the ions unable to completely screen the applied field, and a field began to appear in the bulk.

The spatial resolution of our electrostatic potential measurement is 200 nm. As we determine the lateral electric field by numerically differentiating the measured electrostatic potential7, this finite resolution necessarily results in an electric field that peaks at least 200 nm away from the metal electrode. Given an additional ±250 nm uncertainty in the location of the electrode — determined from measured capacitance — we can locate the peak of the electric field to within only approximately ±450 nm. Figure 4 in the paper1 shows the field maxima to be 450 nm and 220 nm away from the cathode and the anode, respectively. The measured locations of the electric field maxima are thus entirely consistent with the hypothesis that the electric field peaks right at the contacts.

The material of choice for our work1 was the ionic transition metal complex [Ru(bpy)3]2+ with PF6− counter ions. The exact mobilities of electrons and holes in this material are not known, as the presence of ionic conduction makes the usual techniques, such as time-of-flight, challenging to apply (the same holds for any semiconductor with mobile ions). However, the mobilities must be substantial to support electroluminescence, which has been extensively studied in LEECs made from [Ru(bpy)3]2+(PF6−)2 and its derivatives over the past decade8. We would thus expect that the observed interfacial fields are not a peculiarity of the [Ru(bpy)3]2+(PF6−)2 material. Indeed, in a recent publication9, the Ginger group at the University of Washington reported scanning Kelvin probe measurements of potential profiles in LEECs made using a poly(p-phenylene vinylene) with mobile ions. Most of the potential in their study was also found to drop near the cathode.

In conclusion, direct measurements of electric-field distribution in LEECs made with ionic transition metal complexes1 and, more recently, conjugated polymers9 show high electric fields near the electrodes. The electrochemical model fails to account for these interfacial electric fields as it ignores important aspects of the physics of metal–semiconductor contacts.