SUPPORTING INFORMATION Hierarchical Self-Assembly of Gold Nanoparticles into Patterned Plasmonic Nanostructures Cyrille Hamon 1, Sergey Novikov 1, Leonardo Scarabelli 1, Lourdes Basabe-Desmonts 2,3 Luis M. Liz-Marzán 1,3,* 1 Bionanoplasmonics Laboratory, CIC biomagune, Paseo de Miramón 182, 20009 Donostia San Sebastian, Spain 2 BIOMICs Research Group, Lascaray Ikergunea research center, Univ Basque Country, Euskal Herriko Unibertsitatea UPV EHU, Vitoria, Spain 3 Ikerbasque, Basque Foundation for Science, Bilbao, Spain *Corresponding author s e-mail: llizmarzan@cicbiomagune.es 1
Details on Gold Nanorod Functionalization Figure S1. TEM characterization of Au nanorods size distribution. a) Representative TEM image of as synthesized nanorods with an average length of 55±6 nm and an average width of 17±2 nm, aspect ratio of 3.2. b) UV-Vis spectra of CTAB GNRs in water. c,d) Histograms of the width and length of the GNRs in a). Statistics were performed over 200 particles using a semi-automated procedure in the open access software Image J. 2
Figure S2. UV-Vis spectra of a GNR colloid before (black) and after (red) functionalization with MUDOL. A slight red shift of 5 nm was determined. GNRs of 55±6 nm length and 17±2 nm width were synthesized using the method described in ref. 1. After synthesis, CTAB was exchanged with MUDOL (see Experimental Section in main text). Upon ligand exchange, a red-shift of 5 nm was noticed (Figure S2). Close packed structures were formed after drying colloids with high nanorod concentration (1.5 mm Au 0 ), whereas no standing superlattice formation was observed from concentrations below 0.15 mm Au 0, in agreement with the literature. 2 In this case CTAB-GNRs possess a positive surface charge (ζ = +40 mv), inducing highly repulsive forces. On the other hand, the zeta potential of MUDOL rods was significantly lower (ζ = +10 mv) and self-assembly into standing supperlattices occurred even at low GNR concentrations (i.e. below 1.5 mm). The self-assembly of CTAB and MUDOL capped GNRs was compared by TEM at a particle concentration of 1.5 mm. Evaporation was allowed to proceed for 4-6 hours in a homemade setup (Figure S3). 3
Figure S3. Comparison of the self-assembly of MUDOL-GNR and CTAB-GNR by drop casting on TEM grids. a) Home-made setup used to slow down the evaporation of the solvent. The GNR colloids were drop casted (10 µl) on a TEM grid inside a petri dish, which was then carefully placed in a beaker that was half filled with water. Finally, the beaker was sealed with a watch glass. Samples were completely dried after 4-6 h, as compared to 2 h in air. b,c) TEM images at different magnifications for GNR-MUDOL. The inset in c) shows a homogeneous standing supperlattice. d,e) TEM images at different magnifications for GNR-CTAB. The inset in e) shows non-uniform organization of the nanorods but with certain local order. 4
Since MUDOL-GNRs showed better organization into standing superlattices, they were used in all experiments. Nanorods stabilized with 0.1 mm MUDOL and residual CTAB (0.5 mm) remained well dispersed for weeks. This CTAB concentration was required for colloidal stability, since even dilution with an equal volume of water led to fast aggregation, as indicated by a double plasmon band with two peaks centered at 635 nm and 800 nm (Figure S4a). This was attributed to side-to-side and end-to-end aggregation of the GNRs and the corresponding plasmon coupling. When the initial CTAB concentration was recovered, complete disassembly was observed (Figure S4b). This reversible aggregation has been previously reported by Xie et al. 3 and supports the replacement of CTAB by MUDOL even if the positive surface charge of the MUDOL-GNR (ζ = + 10 mv) indicates that some residual CTAB remains on GNR surface. 4 Figure S4. UV-Vis spectra of GNR colloids upon addition and removal of CTAB. a) 0.25 mm CTAB and 0.05 mm MUDOL: fast aggregation occurred. b) 0.5 mm CTAB and 0.05 mm MUDOL: a stable dispersion was obtained. 5
Figure S5. PDMS template characterization. a-d) Optical microscopy images of a typical PDMS template. a) General view of the template. The area comprised between four crosses was 40 10 3 µm². b-d) Higher magnification images showing in detail the different cavity shapes: circles (b), drops (c), and squares (d). The dimensions of the different cavities and distances between them are indicated. e) Sketch showing that all cavities had a fixed depth of 4.8 µm. 6
Figure S6. Multiscale SEM characterization of patterned substrates obtained by PDMS templated self-assembly. Patterned substrates were obtained from colloidal solutions of nanorods and spheres with dimensions of 150 30 nm (a-c) and 17 nm (d-i), respectively. Images were selected so as to represent supercrystals from cavities with different size and shape: a-c) Circles of 12 µm in diameter; e,f) Squares of 20 µm in diameter; d,g-i) Squares of 6 µm in diameter. 7
Figure S7. Representative TEM image of pentatwinned GNRs with average length of 150 nm and average width of 30 nm, aspect ratio 5. The inset is a magnified image highlighting the sharpened ends of a pentatwinned GNR. 8
Figure S8. Characterization of GNR supercrystals obtained from colloids of different concentrations. a-h) SEM images of patterned substrates obtained by PDMS templated selfassembly on a glass slide. The initial GNR (55 17 nm) concentrations (expressed as Au0 molar concentration) were 15 mm (a-c); 75 mm (d-f); and 375 mm (g-i). When the GNR concentration was 500 nm, a monolayer of GNRs was observed outside the limits of the cavities. 9
Figure S9. SEM images at different magnifications of the structures obtained by drying a GNR (55 17 nm) suspension between a glass slide and a PDMS mold displaying circular holes. The sample was imaged on the glass slide after removing the mold. The initial concentration of the GNR dispersion (expressed as Au 0 molar concentration) was 15 mm. Discrete GNR supercrystals are observed instead of a continuous line, which was however observed at higher nanorods concentration. 10
Figure S10. Determination of interparticle spacing. a-d) TEM images at different magnifications of the structures obtained by drying a GNR (55 17 nm) dispersion between a TEM grid and a PDMS mould displaying circular holes. The plots in d) correspond to intensity profiles where the separation distance between lying (red) and standing (green) nanorods was determined to be 2 nm in both cases. e-g) Multiscale AFM characterization. e) AFM topography map of a supercrystal obtained from cavities displaying drop geometry. f) AFM characterization of the surface at high magnification. g) Profile corresponding to the grey line in f). An interparticle distance of 2 nm was determined, in good agreement with TEM measurements. 11
Figure S11. Removal of organics with O 2 plasma. SEM images of nanostructures formed on a glass slide, after removing the mold and cleaning with oxygen plasma at 0.4 mbar O 2 and 200 W, with increasing etching times: 0, 0.5, 2, 4 min for a-d. The optimized etching time was found to be 2 minutes. After 4 minutes, reshaping of nanorods into spheres was noticed. 12
Figure S12. SEM multiscale characterization of a patterned substrate after SERS measurements. The sample was first cleaned with oxygen plasma (2 min, 0.4 mbar O2, 200 W) and then with UV/ozone for 1h. Crystal Violet (10-6 M in ethanol) was drop casted and dried under ambient conditions. After SERS measurements, the sample was cleaned with UV/Ozone for 1h prior to SEM characterization. 13
Figure S13. Comparison of crystal violet SERS spectra measured on supercrystals obtained with different gold nanoparticle shapes: 17 nm diameter spheres (red), 55 17 nm single crystal nanorods (blue), 150 30 nm pentatwinned nanorods (black). Crystal violet at a concentration of 10-6 M in ethanol was dropcasted on the substrate and dried under ambient conditions. The acquisition time was 500 ms and the laser power was 0.15 mw, with an excitation wavelength of 633 nm. Spectra were vertically shifted for the sake of clarity. The total number of counts is indicated by black arrows corresponding to the intensity of the peak at 1626 cm -1. References 1. Scarabelli, L.; Grzelczak, M.; Liz-Marzán, L. M. Tuning Gold Nanorod Synthesis through Prereduction with Salicylic Acid. Chem. Mater. 2013, 25, 4232 4238. 2. Ming, T.; Kou, X.; Chen, H.; Wang, T.; Tam, H.-L.; Cheah, K.-W.; Chen, J.-Y.; Wang, J. Ordered Gold Nanostructure Assemblies Formed By Droplet Evaporation. Angew. Chem. Int. Ed. 2008, 120, 9831-9836. 3. Xie, Y.; Guo, S.; Ji, Y.; Guo, C.; Liu, X.; Chen, Z.; Wu, X.; Liu, Q. Self-Assembly of Gold Nanorods into Symmetric Superlattices Directed by OH-Terminated Hexa(ethylene glycol) Alkanethiol. Langmuir 2011, 27, 11394-11400. 4. Hamon, C.; Postic, M.; Mazari, E.; Bizien, T.; Dupuis, C.; Even-Hernandez, P.; Jimenez, A.; Courbin, L.; Gosse, C.; Artzner, F.; et al. Three-Dimensional Self-Assembling of Gold Nanorods with Controlled Macroscopic Shape and Local Smectic B Order. ACS Nano 2012, 6, 4137-4146. 14