www.afm-journal.de FULL PAPER © 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 2744 www.MaterialsViews.com wileyonlinelibrary.com Adv. Funct. Mater. 2011, 21, 2744–2753 Dan Credgington,* Rick Hamilton, Pedro Atienzar, Jenny Nelson, and James R. Durrant* 1. Introduction With power conversion efﬁciencies (PCEs) approaching 8% in laboratory-scale devices, [1] bulk-heterojunction (BHJ) organic photovoltaics (OPVs) comprising phase-separated blends of electron-donating and electron-accepting organic species have begun to demon- strate their potential as a route to ﬂexible, low cost microgeneration. The underlying physical processes determining the opera- tion and efﬁciency of BHJ solar cells have been shown to be substantively different from those of their inorganic counter- parts, and many questions on the origin of device behaviour are still not fully resolved. One aspect of organic solar cell performance which has been studied extensively in recent years is the origin of the open-circuit voltage ( V OC ), the voltage developed across an illuminated cell’s electrodes when no external current ﬂows between them. [2–11] While it is clear that the choice of active layer materials has a signiﬁcant effect on open-circuit voltage for a given cell, [8,10,12,13] of similar importance are the detailed processing conditions [14] and choice of cell structure, including the selection of suitable device electrodes. [6,11,15] These can lead to consid- erable variations in V OC , particularly when the optimal processing conditions for a new material are still under investigation. Here, we demonstrate that variations in open-circuit voltage arising from differ- ences in fabrication conditions for a large set of BHJ solar cells fabricated from a single polymer:fullerene pair are predominantly determined by variations in non- geminate recombination losses in the device. Signiﬁcant vari- ations in the recombination coefﬁcient and charge densities in the photoactive layer lead to concomitant variations in open- circuit voltage between cells fabricated from nominally identical active layer materials. Several approaches have previously been explored to explain the origin of V OC in BHJ solar cells, with the simplest being the MIM (metal–insulator–metal) model. In this interpretation, at open-circuit, electron and hole polarons generated in the device drift to opposite electrodes under the inﬂuence of a built-in ﬁeld, until sufﬁcient charge accumulates to negate that ﬁeld (or rather, reduce it sufﬁciently such that residual drift is balanced by back-diffusion). Under these conditions, V OC scales with the difference between the work functions of the two electrodes. MIM behaviour has been observed in BHJ solar cells for the Non-Geminate Recombination as the Primary Determinant of Open-Circuit Voltage in Polythiophene:Fullerene Blend Solar Cells: an Analysis of the Inﬂuence of Device Processing Conditions The physical origin of the open-circuit voltage in bulk heterojunction solar cells is still not well understood. While signiﬁcant evidence exists to indicate that the open-circuit voltage is limited by the molecular orbital energies of the heterojunction components, it is clear that this picture is not sufﬁcient to explain the signiﬁcant variations which often occur between cells fabricated from the same heterojunction components. We present here an analysis of the variation in open-circuit voltage between 0.4–0.65 V observed for a range of P3HT/PCBM solar cells where device deposition conditions, electrode structure, active-layer thickness and device polarity are varied. The analysis quantiﬁes non-geminate recombination losses of dissociated carriers in these cells, measured under device operating conditions. It is found that at open- circuit, losses due to non-geminate recombination are sufﬁciently large that other loss pathways may effectively be neglected. Variations in open-circuit voltage between different devices are shown to arise from differences in the rate coefﬁcient for non-geminate recombination, and from differences in the charge densities in the photoactive layer of the device. The origin of these differences is discussed, particularly with regard to variations in ﬁlm micro- structure. By separately quantifying these differences in rate coefﬁcient and charge density, and by application of a simple physical model based upon the assumption that open-circuit is reached when the ﬂux of charge photogenera- tion is matched by the ﬂux of non-geminate recombination, we are able to calculate correctly the open-circuit voltage for all the cells studied to within an accuracy of ±5 mV. DOI: 10.1002/adfm.201100225 Dr. D. Credgington, Dr. R. Hamilton, Prof. J. R. Durrant Centre for Plastic Electronics Department of Chemistry Imperial College London London, SW7 2AZ, UK E-mail: d.credgington@imperial.ac.uk; j.durrant@imperial.ac.uk Dr. P. Atienzar, Prof. J. Nelson Centre for Plastic Electronics Department of Physics Imperial College London London, SW7 2AZ, UK