Figure 1: Overview over the used molecules. Molecular structures and energetic gap with respect to the vacuum are shown for the materials 6T, 5T and 3T. Data is obtained by CV measurements in solution and taken from ref. [1],[2] and [3]. Figure 2: 2D GIWAXS plots of neat films and blends. a, 6T; b, 5T; c, 3T; d, 3T:6T; e, 3T:5T. The blends have a volume ratio of 1:1. While there are no peaks of 3T visible in the 3T:6T blend, in the 3T:5T blend the out-of-plane p-p stacking peak (labeled C) is still present. The data files are in edf format and can be opened for example with SAXS Program Package or EDFExplorer. Figure 3: Comparison of the 3T dimer and 6T monomer. a,b, Geometry of the 3T dimer (a) and the 6T monomer (b). c,d, Electrostatic potential for the 3T dimer (c) and the 6T monomer (d). Presented are iso-surfaces for an absolute value of the potential of 0.01 V and a positive sign (red) or negative sign (blue). The grid in the background is a guide for the eye; the distance between two grid points is 10 Ĺ. The data files contain the source code for rendering with POV-Ray. Figure 4: Results of the photoelectron spectroscopy measurements. a, IEs of 3T and 5T as function of the amount of 5T in the blend. b, IEs of 3T and 6T as function of the amount of 6T in the blend. c,d, Electron affinity and calculated single-particle gap for the pure materials 3T and 6T and for 3T in blends depending on the amount of 6T in the blends. The single-particle gap is defined as difference between the measured IE and EA. In all four graphs, an arrow indicates the difference between the intrinsic layers. In addition, in (a) and (b) an arrow indicates the distance between the materials in the blend. In (d) an arrow also indicates the increase in the gap of the 3T during the transition from the pure layer to the blend, and another arrow the reduction with increasing 6T content. Figure 5: Comparison of the simulated and experimental changes of the ionization energy and single-particle gap. a, Blending-induced changes of IE in experiment (red dots: 3T, filled green diamonds: 6T) and simulation (pink circles: 3T, open green diamonds: 6T). Simulated data points are at mixing ratios of 0.0, 0.52, and 1.0. b, Comparison of the gap energy of 3T from experiment (red dots) and from simulations that include a varying dielectric constant (pink circles). The experimental data is determined with PES and also shown in Fig. 4c. c, Dielectric constant (black circles, scale to the left) and molar volume (grey squares, scale to the right) for different amounts of 6T in the blend. The molar volume is normalized to the neat phase of 6T, while ‘molar’ refers to 3T dimers and 6T monomers as species. Supplementary Figure 1: Schematics of the orientation of the molecules in the neat films of 3T, 5T, and 6T including the characteristic stacking distances obtained from the data below. Supplementary Figure 2: Intensity profiles for the various neat and blend films. a,c, Out-of-plane; b,d, in-plane. The values of the peak positions and widths of the marked peaks are presented in Supplementary Table 1. Supplementary Figure 3: UPS spectra (gray circles) and fits (solid lines) for mixed blends: a,b, 3T:6T; c,d, 3T:5T. The left panels show the high binding energy cut-off (HBEC) and the right panels the HOMO region. Please note, that we have chosen the binding energy with respect to the Fermi energy EF for the HBEC to visualize the trend of the work function (wf), whereas we plotted the HOMO region versus the binding energy with respect to the vacuum energy Evac to directly show the behavior of the IE. The spectra of the HOMO region are fitted using two Gaussian peaks for each material and Tougaard backgrounds (black lines). The peaks belonging to 3T are drawn in red, the peaks of 6T are green and the peaks of 5T blue. For fitting the blends, the distance of the two peaks belonging to one material and ratio of the peak areas are kept constant. This means, the free fitting parameters are the distance between the different materials, the ratio of the total areas of the materials, and the peak widths. Supplementary Figure 4: IPES spectra of 3T, 6T, three mixing ratios and ITO, which is used as substrate. The mixing ratios are obtained from XPS data. The black vertical lines mark the onset positions. A rough description of the determination of the onset is given in experimental section of the main text. Supplementary Figure 5: Determination of the optical gap as described in ref. [4]: a, 6T; b, 3T. Supplementary Figure 6: Spectra of the optical density for the pure materials and three molar mixing ratios: a, 3T; b, 4:1; c, 2:1; d, 1:1; e, 6T. The black lines represent the measured spectra, the red dashed lines the fits for the 3T peaks and the green dotted lines the fits for the 6T peaks. The cumulated fit results are drawn as dash-dot line. Fig. f shows the position of the 3T peak with higher energy in the blends and the higher energetic 6T peak for pure 6T. Supplementary Figure 7: Current-voltage data for solar cells containing blended layers as donor layers. a, Current density j as a function of applied voltage V for planar heterojunction solar cells with the investigated blends as donor layer. b, Open-circuit voltage Voc as function of the amount of 6T in the donor layer. It decreases constantly from 1.06 V for pure 3T to 0.82 V for pure 6T. Supplementary Figure 8: Illustration of the monopole-quadrupole interaction. a, Schematic representation of the surface geometry. The excited molecule (orange) is located at the surface in z-direction, resulting in two nearest neighbors in x-direction and only one nearest-neighbor in z-direction. The red plus/blue minus represent more positive/negative quadrupole moments of 3T dimers with respect to 6T monomers (not the actual quadrupole moments). b, Change of the experimental ionization energy (filled diamonds) and monopole-quadrupole interaction energy (open diamonds) with respect to the blending ratio. Supplementary Figure 9: Dependence of the difference of the electrostatic interaction energy DeltaPhi=Phi_EA+Phi_IE for 3T in dependence of the content of 6T in the blend. The largest DeltaPhi is shifted to zero energy. [1] Fitzner, R. et al. Dicyanovinyl-substituted oligothiophenes: Structure-property relationships and application in vacuum-processed small molecule organic solar cells. Adv. Funct. Mater. 21, 897–910 (2011). [2] Fitzner, R. et al. Interrelation between crystal packing and small-molecule organic solar cell performance. Adv. Mater. 24, 675–680 (2012). [3] Wynands, D. et al. Organic thin film photovoltaic cells based on planar and mixed heterojunctions between fullerene and a low bandgap oligothiophene. J. Appl. Phys. 106, (2009). [4] Vandewal, K., Benduhn, J. & Nikolis, V. C. How to determine optical gaps and voltage losses in organic photovoltaic materials. Sustain. Energy Fuels 2, 538–544 (2018).