Self-cleaning antireflective optical coatings – Supporting Information – Stefan Guldin 1† , Peter Kohn 1 , Morgan Stefik 2 , Juho Song 3 , Giorgio Divitini 4 , Fanny Ecarla 5 , Caterina Ducati 4 , Ulrich Wiesner 3 & Ullrich Steiner 1⋆ Further experimental details Material fabrication. The poly(isoprene-b-ethylene oxide) (PI-b-PEO) BCPs were synthesised by anionic polymeri- sation as described in reference 1 . Two different polymers were used: BCP-34 with a molecular weight of M n = 34.4 kg mol −1 , 28 wt% PEO, PDI = 1.05, and BCP-92 with M n = 91.9 kg mol −1 , 31.2 wt% PEO, PDI = 1.09. The polymer characteristics were determined by gel permeation chromatography and NMR. TiO 2 nanocrystal synthesis was carried out in a similar method as reported elsewhere 2 . The synthesis was modified to compatibilise the nanoparticles with the solvent mixture used to dissolve the polymer and sol. The resulting TiO 2 crystallites mix with the sol and are thereby incorporated into the highly porous network during film deposition. In short, the glassware was cleaned, dried, connected to a Schlenk line and evacuated before the remaining water was removed by heating. After stabilisation of a nitrogen atmosphere, the following chemicals were consecutively added to the flask under vigorous stirring: 5.75 ml absolute ethanol, 1 ml TiCl 4 , 19.2 ml benzyl alcohol, and 0.23 ml 1,3-propane diol. The solution was then heated to 80 ◦ C and stirred for 12 hours. The solute was subsequently precipitated in diethyl ether (1:10 volume ratio) and centrifuged at 3500 rpm for 10 minutes. The resulting wet precipitate was dried for 2 hours in ambient conditions and was then redissolved in an azeotrope solvent mixture of toluene (72.84 w%) and 1-butanol (27.16w%). To maintain consistent concentrations of TiO 2 nanocrystals in the azeotrope solution (20 μg per ml), a fraction of the precipitate was fully dried and heated to 350 ◦ C aside to reveal the weight content of TiO 2 . The silica-based sol was prepared as follows: 2.8 g (3-glycidyloxypropyl)trimethoxysilane (≥ 98%, Aldrich) and 0.32g aluminum-tri-sec-butoxide (97%, Aldrich) were mixed with 20mg KCl (TraceSELECT, Fluka) and promptly placed into an ice bath. In a first hydrolysis step, 0.135 ml of 10 mM HCl was added dropwise in 5 s intervals at 0 ◦ C and stirred for 15 min. After warming up to room temperature, 0.85 ml 10 mM HCl was further added dropwise. The sol was then stirred for 20 min before adding to the hybrid solution. Depending on the TiO 2 content, the mixing ratios were as follows: For 0w% TiO 2 loading, 50 mg polymer was dissolved in 1.2 ml of azeotrope solvent before 75 mg silica-based sol was added. For 25w% TiO 2 loading, 50 mg polymer was dissolved in 0.7 ml of azeotrope solvent and 0.5 ml of TiO 2 solution was added prior to the addition of 56 mg sol. Solutions with 37.5 w% TiO 2 loading consisted of 50 mg polymer in 0.5 ml of azeotrope solvent, 0.7 ml of TiO 2 solution and 47 mg silica-based sol solution. The 50 w% TiO 2 solution was prepared by adding 50 mg polymer to a mixture of 0.6 ml azeotrope solvent and 0.94 ml TiO 2 solution before 37.5 mg sol was added. The hybrid solutions were then stirred for 1 hour and further diluted as needed for film deposition. Hybrid films were deposited onto pre-cleaned glass slides by spin coating (2000 rpm, 20 s). The cast films were annealed on a hotplate by gradually increasing the temperature to 200 ◦ C (180 min linear ramp, 30 min dwell time). In a final step, the organic components of the hybrid films were removed by reactive ion etching in oxygen plasma (Diener MRC 100, 30min, 0.4mbar). In the light of moving from glass to flexible substrates we have subsequently worked on reducing requirements for temperature annealing and plasma treatment. A thermal annealing at 130 ◦ C (15 min linear ramp, 5 min dwell time), followed by an exposure of 10 min to oxygen plasma led to identical results in terms of optical performance and self-cleaning for flexible ARCs on PET foil (Melinex, Dupont), one of the most widely used flexible substrate in the optoelectronic industry. Indeed it has been previously been shown that hydrolysis and condensation of the aluminosilicate network is complete after annealing at 130 ◦ C 3 . By carefully studying the effects of our processing on neat PET foils we could further exclude deteriorating effects of oxygen plasma treatment on the foil itself for etching times of 25 min and below (see Figure SI12). Spin coating on flexible substrates was carried out by attaching the substrate on an aluminium sample holder that was custom-built. We note that durability against these processing conditions are a prerequisite when moving to other substrate materials. Electron Microscopy. Scanning electron microscopy was carried out with a Leo Ultra 55 with a field emission source at an acceleration voltage of 3kV. Transmission electron microscopy was accomplished with a FEI Tecnai F20-G2 FEGTEM with a 200 kV field emission gun. 1 Department of Physics, University of Cambridge, J. J. Thomson Avenue, Cambridge, CB3 0HE, UK. 2 Department of Chemistry and Biochemistry, University of South Carolina, Columbia, 29208, USA. 3 Department of Materials Science and Engineering, Cornell University, Ithaca, New York, 14853, USA. 4 Department of Materials Science and Metallurgy, University of Cambridge, Charles Babbage Road, Cambridge, CB3 0FS, UK. 5 CSM Instruments, Rue de la Gare, 2034 Peseux, Switzerland. † Now at: Institute of Materials, École Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland. ⋆ e-mail: u.steiner@phy.cam.ac.uk