Proceedings of nanoGe Fall Meeting19 (NFM19)
DOI: https://doi.org/10.29363/nanoge.nfm.2019.164
Publication date: 18th July 2019
Nanoscale morphology is an essential feature of bulk-heterojunction organic solar cells. In the past, disordered geometries have helped to push efficiencies beyond what early bilayer strategies could achieve, by providing a compromise between interface processes and transport along percolation pathways. [1] It is therefore desirable to acquire a fundamental understanding of the relationship between the heterojunction's nanoscale morphology and the transport of exciton or free carriers in these devices.
In this context, I use numerical simulations to investigate the influence on transport of various morphologic traits such as the tortuosity of the conduction pathways, the presence of dead-ends or the fraction of disconnected domains. In this contribution, I show how much can be learned from applying drift-diffusion models to a few representative test cases. I argue that these can form a basis for a better understanding of structure-property relationships in organic solar cells, as depicted in the attached figure. For instance, in open-circuit conditions, the surface-to-volume ratio of the conduction pathways is a sufficient geometrical descriptor, whereas a quantitative definition of tortuosity is needed to describe short-circuit conditions. In the future, a strong knowledge of these basic cases could provide the missing link between extensive morphology information (e.g. as obtained by electron microscopy or X-ray diffraction) and interpretation in terms of transport and device performances.
This is embedded in a parameter extraction procedure to quantitatively interpret characterization data, and infer morphologic information. To this end, I use a Bayesian framework [2] to aggregate information from multiple techniques. This makes the inference more robust [3], and takes the algorithm presented in reference [4] one step further. I demonstrate with simple examples why this should be done routinely to avoid dramatic interpretation errors.
Possible applications include the use of non-destructive techniques on fully grown devices (e.g. I-V curve, TPC, TPV, CELIV, …) as a way to compare active layer depositions or identify degradation mechanisms.