Supplementary Information

Transcription

1 Supplementary Information Size-Tuning of WSe 2 Flakes for High Efficiency Inverted Organic Solar Cells George Kakavelakis 1,, Antonio Esau Del Rio Castillo 2,, Vittorio Pellegrini 2, Alberto Ansaldo 2, Pavlos Tzourmpakis 1, Rosaria Brescia 3, Mirko Prato 3, Emmanuel Stratakis, 4 Emmanuel Kymakis 1*, and Francesco Bonaccorso 2 * 1 Center of Materials Technology and Photonics & Electrical Engineering Department, Technological Educational Institute (TEI) of Crete, Heraklion Crete, Greece. 2 Istituto Italiano di Tecnologia, Graphene Labs, Via Morego , Genoa, Italy. 3 Istituto Italiano di Tecnologia, Nanochemistry Department, Via Morego , Genoa, Italy. 4 Institute of Electronic Structure and Laser (IESL), Foundation of Research and Technology (FORTH), Heraklion, Crete, Greece. These authors contributed equally to this work. *Address correspondence to kymakis@staff.teicrete.gr and francesco.bonaccorso@iit.it

2 S.1 Atomic Force Microscopy The size of the PC 71 BM domains in PTB7:PC 71 BM blend is estimated by atomic force microscopy (AFM). A pristine PTB7:PC 71 BM blend film, prepared in the same way of the full cells, is imaged in tapping mode and the tapping phase image is acquired (Figure S1a). The domain area is estimated by image-j software measuring 100 random domains. The domain area distribution is shown in Figure S1b. The average domain area is 220 ± 80 nm 2 and it is very close to the average size of sample 2 (240 nm 2 in average). Figure S1: a) AFM phase image of a PTB7:PC 71 BM blend film; b) domain area distribution. S.2 X-ray photoelectron spectroscopy We analyze the surface of our WSe 2 samples by means of X-ray photoemission spectroscopy (XPS) to get elemental and chemical information from the outer 5 to 10 nm of the solid surface. 1 We studied the chemical composition of the three samples as well as the starting material, i.e., bulk WSe 2, see Figure S2. The spectra collected on samples 2 and 3, (reported in panels Figure S2 c-h) can be fitted by using the same doublets used for the starting material, thus indicating that the same chemical species are present. Also the ratios between WSe 2 and WO 3 remain similar as in the starting material (15% ± 3% of WO 3 ). A noticeable increase of the oxide component, likely due to a surface oxidation process occurring on WSe 2 during the exfoliation process, is instead detected on sample 1.

3 The data collected on Se 3d and W 4f peaks are reported in Figure S2a-h together with the results of the best fit procedure (see Methods, Characterization of WSe 2, for experimental details). The starting material is characterized by a single Se 3d doublet (blue profiles in Figure S2a), having the Se 3d 5/2 component centered at 54.8±0.2 ev, in agreement with values reported in literature for WSe 2. 2 The best fit of the W 4f region (panel S2b) required the use of two doublets, thus indicating that W is present in two different oxidation states in the bulk WSe 2 before the LPE procedure. Figure S2: XPS spectra collected on the starting material (a, b) and the three WSe 2 flake samples: (c, d) sample 1; (e, f) sample 2; (g, h) sample 3. Panels (a), (c), (e) and (g): data collected on Se 3d peaks. Panels (b), (d), (f) and (h): data collected on W 4f peaks. Experimental data are represented by black markers. Blue lines refer to the Se and W components related to WSe 2. In particular, the most intense doublet (blue profiles, accounting for 83% of the whole W content of the sample) has W 4f 7/2 component at 32.6±0.2 ev and W 4f 5/2 component at 34.7±0.2 ev, and it is assigned to W (IV) in WSe 2. 3,4 The second doublet (green profiles, accounting for 17% of the W content) has W 4f 7/2 component at (36.2±0.2) ev and W 4f 5/2 component at 38.4±0.2 ev: this W doublet can be assigned to a small amount of WO 3. 3 A broad W 5p 3/2 peak (related to WSe 2 ) is also observed in the range of the oxide peaks, as detailed in the Methods section. The spectra collected on samples 2 and 3, (reported in panels S2 c-h) can be fitted by using the same doublets used for the starting material, thus indicating

4 that the same chemical species are present. Also the ratios between WSe 2 and WO 3 remain similar as in the starting material (15% ± 3% of WO 3 ). A noticeable increase of the oxide component is detected on sample 1. In fact, by analyzing the data reported in panel S2d (W 4f peaks measured on sample 1), we calculated a WO 3 content of 61% (with WSe 2 accounting for the 39% of the total W content of the sample). In parallel, a new doublet is needed to fit the data collected on sample 1 for Se 3d peaks. The new doublet (orange profiles in panel 2b) has the Se 3d 5/2 component centered at 55.6±0.2 ev. The observed position is in the range reported for elemental Se, 2 indicating that the new component is likely due to a surface oxidation process occurring on WSe 2 during the exfoliation process. S.3 Optimization Experiments In order to optimize the photovoltaic performances of the WSe 2 -based OSC, it is key to investigate the influences of different concentration of the WSe 2 flakes in the WSe 2 -based ternary blends. For this purpose, we prepared three batches of the ternary blend by separately adding diverse WSe 2 samples (sample 1, sample 2 and sample 3), in different concentrations (1.5%, 2.0%, 2.5% v/v), in the as-prepared binary (PTB7:PC 71 BM) dispersions, as reported in the Method section of the main manuscript. The results obtained for the three different samples at the WSe 2 concentration of 1.5, 2.0 and 2.5% v/v are summarized in table S1. For all the three different samples, the best photovoltaic performances are achieved with a WSe 2 concentration of 2.0% v/v.

5 Table S1. Optimization experiments of the concentration of WSe 2 in PTB7:PC 71 BM BHJ WSe2 Sample area/thickness WSe 2 Calculated J sc ratio (nm) Concentr. (%) J sc (ma cm -2 ) (EQE) Voc (V) FF (%) η (max.) * Reference No ± ± ± ± (8.22) 1 ~ ± ± ± ± (8.37) ± ± ± ± (8.57) ± ± ± ± (8.41) 2 ~ ± ± ± ± (8.85) ± ± ± ± (9.45) ± ± ± ± (9.01) 3 ~ ± ± ± ± (8.61) ± ± ± ± (8.91) ± ± ± ± (8.65) S.4 Bimolecular recombination kinetics In Figure S3 is reported the measured short-circuit current density (J sc ) of PTB7:PC 71 BM BHJ OSCs with and without the different WSe 2 flakes plotted against light intensity. In Figure S3 it is also reported the corresponding fitted power law, which determines the exponential factor n of Eq. 4 of the main text. As shown in Figure S3, n for the PTB7:PC 71 BM, PTB7:WSe 2 (sample1):pc 71 BM-, PTB7:WSe 2 (sample 2):PC 71 BM- and PTB7:WSe 2 (sample 3):PC 71 BM- based devices are 0.870, 0.884, and 0.893, respectively.

6 Figure S3. Measured J sc of PTB7:PC 71 BM BHJ OSCs with and without the different WSe 2 flakes plotted against light intensity (symbol) and corresponding fitted power law (line), which determines the exponential factor n of Eq. 4 of the main text. References (1) Briggs, D.; Grant, J. T. (Eds.). Surface Analysis by Auger and X-ray Photoelectron Spectroscopy, IM Publications, Chichester, (2) NIST X-ray Photoelectron Spectroscopy Database, Version 4.1 (National Institute of Standards and Technology, Gaithersburg, 2012); (3) Yu, X.; Prévot, M. S.; Guijarro, N.; Sivula, K.. Self-Assembled 2D WSe 2 Thin Films for Photoelectrochemical Hydrogen Production. Nat. Commun. 2014, 6, 10. (4) Wang, H.; Kong, D.; Johanes, P.; Cha, J. J.; Zheng, G.; Yan, K.; Liu, N.; Cui, Y. MoSe 2 and WSe 2 Nanofilms with Vertically Aligned Molecular Layers on Curved and Rough Surfaces. Nano Lett. 2013, 13,