Proceedings of International Conference on Hybrid and Organic Photovoltaics (HOPV22)
DOI: https://doi.org/10.29363/nanoge.hopv.2022.015
Publication date: 20th April 2022
Organic and hybrid electronic devices are applied in various areas ranging from light-emitting diodes and solar cells to field-effect transistors and photodetectors. In common to all those applications is the use of amorphous organic layers which intrinsically contain significant amount of trap states. Depending on the nature of such trap states, they may modify the charge carrier transport or lead to parasitic recombination pathways which affects both the efficiency and the lifetime of the device – usually in a negative way. Despite their relevance, the characterisation of traps in complete devices remains challenging. An electrical technique that allows to potentially determine all relevant parameters of trap states (i.e. density, depth and capture rate) at once and in a full device is the thermally stimulated current (TSC) technique. During this experiment, the device is cooled down to temperatures below 100 K and the trap states are filled by either light irradiation or electrical bias. Afterwards, the temperature is ramped linearly to at least room temperature while the current is recorded. Various analytical formulas, which make use of different parts of the current signal, are available to determine the trap parameters from this current signal.
In this contribution, we aim to test the validity range of the most popular analytical formulas for TSC. We use drift-diffusion simulations to generate synthetic TSC data with varied trap parameters, but also other microscopic quantities, such as the charge carrier mobility, and device parameters, such as the layer thickness. We then apply the analytical expressions and compare the extracted trap parameter with the simulation input. It is found that the “initial rise” method yields reliable trap energies in a large range of tested parameters while the simplest method, that considers only the peak temperature, is off by at least 10%. In addition, the determined trap density is always below the simulation input. Only half of this factor can be explained with electrostatic arguments. Thanks to the electrical profiles obtained from the simulation we can explain this discrepancy by incomplete emptying of trap states. In order to minimize this effect, it is advisable to apply a reverse bias voltage during the temperature ramp.