The three‐dimensional morphology of mixed organic layers are quantitatively measured using high‐angle annular dark‐field scanning transmission electron microscopy (HAADF‐STEM) with electron tomography for the first time. The mixed organic layers used for organic photovoltaic applications have not been previously imaged using STEM tomography as there is insufficient contrast between donor and acceptor components. Contrast is generated by substituting fullerenes with endohedral fullerenes that contain a Lu3N cluster within the fullerene cage. The high contrast and signal‐to‐noise ratio, in combination with use of the discrete algebraic reconstruction technique (DART), allows generation of the most detailed and accurate three‐dimensional map of BHJ morphology to date. From the STEM‐tomography reconstructions it is determined that three distinct material phases are present within the BHJs. By observing changes to morphology and mixing ratio during thermal and solvent annealing, the effects of mutual solubility and fullerene crystallization on morphology and long term stability are determined. This material/technique combination shows itself as a powerful tool for examining morphology in detail and allows for observation of nanoscopic changes in local concentration.

Here the influence that 1‐(3‐hexoxycarbonyl)propyl‐1‐phenyl‐[6,6]‐Lu3N@C81, Lu3N@C80–PCBH, a novel acceptor material, has on active layer morphology and the performance of organic photovoltaic (OPV) devices using this material is reported. Polymer/fullerene blend films with poly(3‐hexylthiophene), P3HT, donor material and Lu3N@C80–PCBH acceptor material are studied using absorption spectroscopy, grazing incident X‐ray diffraction and photocurrent spectra of photovoltaic devices. Due to a smaller molecular orbital offset the OPV devices built with Lu3N@C80–PCBH display increased open circuit voltage over empty cage fullerene acceptors. The photovoltaic performance of these metallo endohedral fullerene blend films is found to be highly impacted by the fullerene loading. The results indicate that the optimized blend ratio in a P3HT matrix differs from a molecular equivalent of an optimized P3HT/[6,6]‐phenyl‐C61‐butyric methyl ester, C60–PCBM, active layer, and this is related to the physical differences of the C80 fullerene. The influence that active layer annealing has on the OPV performance is further evaluated. Through properly matching the film processing and the donor/acceptor ratio, devices with power conversion efficiency greater than 4% are demonstrated.