Supporting Information. Chiral Plasmonic Films Formed by Gold Nanorods and Cellulose Nanocrystals

1 Supporting Information Chiral Plasmonic Films Formed by Gold Nanorods and Cellulose Nanocrystals Ana Querejeta-Fernández, Grégory Chauve, Myriam Methot, Jean Bouchard, Eugenia Kumacheva # * Department of Chemistry, University of Toronto, 80 Saint George Street, Toronto, Ontario M5S 3H6, Canada; FPInnovations, 570 St. Jean Boulevard, Pointe-Claire, QC H9R 3J9, Canada; Institute of Biomaterials & Biomedical Engineering, University of Toronto, 164 College Street, Toronto, Ontario M5S 3G9, Canada; # Department of Chemical Engineering and Applied Chemistry, University of Toronto, 200 College Street, Toronto, Ontario M5S 3E5, Canada *Corresponding author: ekumache@chem.utoronto.ca 1. Materials An aqueous 2.44 wt. % suspension of sulfate-functionalized cellulose nanocrystals (CNCs) at ph=3.05 was supplied by FPInnovations (Canada). Gold (III) chloride solution (99.99%, 30 wt.% in dilute HCl), NaBH 4 (98 wt.%) and L-ascorbic acid ( 99.0 %) (both Aldrich Canada); AgNO 3 (EMD), hexadecyltrimethylammonium bromide (CTAB, 99.0%, Sigma) and NaCl (ACP) were used as-received. 2. Sample preparation 2.1. Preparation of cellulose nanocrystals An aqueous suspension of CNCs was prepared by controlled sulfuric acid hydrolysis of a commercial bleached softwood kraft pulp according to a patent-pending procedure scaled up from the literature. 1

2 2.2. Synthesis of gold nanorods An approximately 515 ml-volume batch of gold nanorods (NRs) with dimensions 41 x 14 nm was synthesized by a modified version of the seed-mediated growth method. 2 First, a seed solution was prepared by mixing an aqueous solution of hexadecyltrimethylammonium bromide (CTAB) (5 g, 0.20 M), with deionized water (2 g) and HAuCl 4 solution (240 µl, 15 mm), which was followed by addition of ice-cold NaBaH 4 (1 ml, 10 mm). After stirring for 2 min, the seed solution was allowed to age for 60 min. The growth solution was prepared by mixing CTAB solution (268 g, 0.20 M), deionized water (200 g), AgNO 3 (8.5 ml, 4 mm), HAuCl 4 (25 ml, 15 mm) and ascorbic acid (6.2 ml, 78.8 mm). Gold NRs were grown in the growth solution upon addition of 5.0 ml of the aged seed solution. The mixture was heated at 27 ºC overnight. The resultant gold NR solution had a brown-reddish color. Gold NRs with dimensions 47 x 20, 33 x 11, 36 x 9 nm were synthesized using a protocol reported elsewhere 2 using varying amounts of AgNO 3 (16 to 60 µl per 1 µl of seed solution). 2.3. Preparation of composite NR-CNC films The as-synthesized colloidal solution of gold NRs with dimensions 41 x 14 nm (initial concentration of NRs, C NR, of 1.01 nm) was cleaned from excess of CTAB by 20 min centrifugation at 10,000 rpm at 27 ºC (Eppendorf centrifuge 5417R; 2.0 ml centrifuge tubes) and subsequent redispersion of the precipitate in deionized water. The process was repeated twice. By adjusting the initial and final volumes of the dispersion, six washed gold NRs colloidal solutions at NRs concentrations of C NR 2, C NR 4, C NR 8, C NR 16, C NR 40 and C NR 60 were

3 prepared. These solutions were used for the preparation of mixed NR-CNC solutions as follows. A NR solution (1 ml) was added drop-wise to a CNC suspension (5 ml, 2.44 wt.%) under ~20 min stirring. After that, the mixed colloidal solution (~6 ml) was poured into a 5.3 cm-diameter polystyrene Petri dish to form a ~4 mm-thick liquid layer. Solid films were obtained in two days after water evaporation. For the preparation of the CNC-NRs composite films with addition of NaCl, the mixed colloidal solutions were prepared by mixing 5 ml of 2.44 wt.% CNC suspension, 0.5 ml of NaCl solution (see salt concentration in Table S2), and 0.5 ml of the solution of gold NRs at the concentration of 16.16 nm. Films were dry cast in a manner described above. 3. Methods Transmission electron microscopy (TEM) imaging of NRs and CNCs was performed on a Hitachi H-7000 microscope operated at 100 KV and 75 KV. The sample of CNCs was prepared by adding a drop of a 0.04 wt.% aqueous suspension onto an ultrathin carbon film on holey carbon supported copper grid (Ted Pella) and rapidly withdrawing water using a filter paper. The sample for imaging gold NRs was prepared by removing excess of CTAB from the as-prepared NRs dispersion by two steps of centrifugation (10,000 rpm, 20 min, 27 ºC; Eppendorf centrifuge 5417R) and redispersing the precipitate in deionized water. The final NR solution was 16 times more concentrated than the initial solution. After placing a drop of the NR solution onto a Formvar-coated copper grid (Electron Microscopy Sciences), the water was quickly removed with a filter paper. Scanning electron microscopy (SEM) of CNC and CNC-NR films was performed on a Quanta FEG 250 environmental microscope. Imaging of the surface of the films was carried out

4 at room temperature without coating it with a conductive layer. Cross-sectional cryo-sem imaging was performed by mounting the film perpendicularly to the SEM holder at a temperature of ~-23.0 ºC. The film cross-sections were prepared by fracturing the films in liquid nitrogen and coating them with a conductive carbon layer. Control experiments confirmed that the cross-section of the films underwent no significant volume change when cooling down to - 23.0 ºC. Extinction measurements were carried out by ultraviolet-visible-near-infrared spectroscopy using a Cary 5000 spectrophotometer. We used either a 1 cm path length plastic cuvette for solutions of gold NRs, or a composite NR-CNC film mounted perpendicularly to the light beam. The spectra were recorded in the spectral range from 200 to 1200 nm; the data interval was 1.00 nm and the scan rate was 600 nm/min. Circular dichroism spectra of the NR-free and NR-CNC composite films were recorded on a Jasco J-710 spectropolarimeter. The films were mounted perpendicular to the beam path. The spectra were recorded in the spectral range from 200 to 800 nm. The acquisition conditions were as follows: the data interval was 2.0 nm, the scan rate was 100 nm/min, the slit width 500 µm and the sensitivity 1000 mdeg. 4. Characterization of cellulose nanocrystals and gold nanorods The electrokinetic potentials (ξ-potential) of gold NRs and CNCs were determined using a Malvern Nano ZS-ZEN3600 instrument. The ξ-potential of an aqueous CNCs suspension at 2.44 wt.% at ph= 3.05 was obtained as an average of 5 measurements using the index of refraction of crystalline cellulose of 1.54. 3 The value of ξ-potential of gold NRs (after one washing step) was determined as the average of 5 measurements by using the refractive index of gold of 1.426. 4

5 Dimensions of gold NRs and CNCs were determined as the average of the dimensions of at least, 200 NRs and 50 CNCs and, respectively. 5. Characterization of NR-CNC films Table S1 shows the concentration of gold NRs in the colloidal solution, in the mixed NR-CNC colloidal solution and in the corresponding NR-CNC films whose structural and spectral characterization is shown in Figures 3 and 4 of the main text, respectively. Table S1. Concentration of gold NRs in the colloidal solutions and in the corresponding cast films. C NR in NR colloidal solution (nm) C NRs in NR-CNC mixed colloidal solutions (nm) C NR in NR-CNC films (wt.%) 2.02 0.34 0.12 4.04 0.67 0.23 8.08 1.35 0.47 16.16 2.69 0.93 40.40 6.73 2.29 60.60 10.10 3.39 Table S2. Concentration of NaCl in suspensions and in the corresponding NR-free and NR-CNC films C NaCl-initial (mm) C NaCl in mixed NR- CNC solution C NaCl in NR-free CNC films (wt.%) / C NaCl in NR- CNC films (wt.%) C NRs in NR-CNC films (mm) (wt.%) 0 0 0 / 0 0 1 0.08 0.022 / 0.022 0.47 3 0.25 0.065 / 0.065 0.47 6 0.50 0.131 / 0.130 0.47 10 0.83 0.218 / 0.217 0.47

6 100 8.33 2.137 / 2.127 0.46 5.1. Preparation of the NR-CNC films When the mixed CNC-NR solution was allowed to settle down in a capped vial for over 24 h (Figure S1), phase segregation was observed: the bottom colorless phase with a chiral nematic organization of CNCs, and the upper isotropic phase containing NRs. In contrast, in the composite film, bottom-up phase segregation was not observed, which was confirmed by SEM imaging, as described below. Instead, partial lateral NR segregation occurred at the edge of films when NR content was from 0.47 to 0.93 wt.% (Figure S2g), presumably, caused by the drying pattern. All optical measurements were carried out in the center of the films. All of the cast films showed a brown-reddish color caused by NRs and strong angle-dependent iridescence, due to the photonic CNC matrix (Figure S2). Figure S1. Photographs of the NR-CNC mixed suspension at C NR =1.35 nm. (a) as-prepared suspension and (b) the same suspension after 24 h incubation

7 Figure S2. Photographs of the composite films at NR concentration in the film of (a) 0, (b) 0.12, (c) 0.47 and (d) 3.39 wt. %, taken with a black background (a-d) and white background (e-h), respectively. 5.2. Structural characterization of the NR-CNC films

8 Figure S3. SEM images of the surface of NR-CNC films at NR concentrations of (a) 0.47 wt.% and (b) 2.29 wt.%. The top and the bottom surfaces of the film exhibited similar structure. Local segregation of NRs took place with exposure of the films to the electron beam. Figure S4. SEM images of the cross-sections of the NR-CNC films at NR concentrations of (a) 0.12 and (c) 3.39 wt. %. (b, d) SEM images corresponding to the areas shown in (a, c), respectively, acquired with backscattered electron detector. The CNC matrix in (b) was partly deformed under electron beam. The effective refractive index, n eff, of the NR-CNC films was calculated using the volume fraction φ i and the refractive index n i of the component i as

9 n eff 2 = Σ φ i n i 2 Eq. S1, where the refractive indexes of CNCs and gold NRs were 1.54 3 and (1.840, 1.426 and 0.160 at λ of 318, 459 and 689 nm) 4 respectively and the volume fraction φ i was calculated as φ i = V i /V (V i is the volume of NR and CNC components and V is the volume of the film). Table S3. Refractive indexes of the composite films at varying contents of the NRs in the film C NR (wt.%) n eff (λ=318 nm) n eff (λ=459 nm) n eff (λ=689 nm) 0 1.540 1.540 1.540 0.12 1.540 1.540 1.540 0.47 1.540 1.540 1.540 0.93 1.540 1.540 1.540 2.29 1.541 1.539 1.539 3.39 1.541 1.539 1.538 Table S4. Helical pitch (P SEM ) and corresponding spectral position of the stop band λ SEM for the composite films with different NR content C NR P SEM * Std. dev. in P SEM λ SEM ** Std. dev. in λ SEM (wt.%) 0 551 19 848 29 0.12 495 28 762 43 0.47 446 29 687 45 0.93 457 26 704 40 2.29 392 31 603 48 3.39 368 38 566 58 * P SEM was determined by analyzing cross-sectional SEM images of the NR-CNC films. ** λ SEM was calculated using Eq. 1 and n eff at 459 nm. Table S5. Helical pitches (P SEM ) and the corresponding spectral position of the stop band λ SEM for the composite films containing NRs with different dimensions Series Aspect ratio of NRs P SEM * Std. dev. in P SEM λ SEM ** Std. dev. in λ SEM

10 A 2.3 414 31 638 48 B 3.0 440 50 678 77 C 4.0 447 54 688 83 * P SEM was determined by analyzing cross-sectional SEM images of the NR-CNC films. ** λ SEM was calculated using Eq. 1 and n eff = 1.540. Table S6. Helical pitches (P) and calculated wavelength of stop band (λ SEM * ) determined from the cross-sectional SEM images of the NR-CNC composite films containing different concentrations of NaCl C NaCl in NR-CNC films (wt. %) P SEM * Std. dev. P SEM λ SEM ** Std. dev. λ SEM 0 449 23 691 35 0.02 395 31 608 48 0.06 378 19 582 29 0.13 343 35 528 54 0.22 299 15 460 23 ~2.13 - - - - * P SEM was determined by analyzing cross-sectional SEM images of the NR-CNC films. **λ SEM was calculated using Eq. and n eff = 1.540. Figure S5. SEM image of the cross-section of the NR-CNC film prepared at the highest NaCl concentration (C NaCl =2.13 wt.%). An achiral nematic liquid crystalline structure is observed as vertical lines corresponding to the anisotropic layers of CNCs 5.3. Effect of CTAB concentration on the structure of NR-free and NR-CNC films To study the effect of CTAB concentration on the CNC host structure, we prepared a mixture from 5 ml of 2.44 wt.% CNCs suspension and 1 ml of NR-free supernatant collected after one

11 and two centrifugation steps of the original NR solution (red and blue spectra, respectively, in Figure S7a and b). The concentration of CTAB in the supernatant was higher after one centrifugation step. In CNC films prepared from these mixtures, the chiral nematic structure was completely destroyed at the higher CTAB concentration, as revealed by the absence of stop band in the extinction spectrum of the films (Figure S7a, blue line), as well as by the lack of the peak in their CD spectrum (Figure S7b, red line). For lower CTAB contents in the supernatant, that is, after two centrifugation cycles, the corresponding film exhibited a photonic band gap and CD activity (Figure S7 a and b, blue spectra), confirming the chiral nematic structure of the CNC film. Similarly, composite NR-CNC films did not exhibit chiroptical activity at a larger CTAB concentration, as revealed by the absence of CD peak (Figure S7, brown spectrum). Weak CD activity two orders of magnitude smaller was preserved for such films, because of the induced chirality by chiral glucose molecules (inset Figure S7d). On the contrary, when CTAB was removed by repeated washing steps, the hybrid film exhibited CD activity (Figure S7d, green line). However, excessive centrifugation (e.g. three cycles) led to the partial dissolution of NRs, as revealed by the change of color of the colloidal solution and the decrease in the intensity of the extinction peaks of the films (Figure S7c, green line). Thus in the present work, we used only two washing steps for gold NRs, in order to preserve the chirality of the films and integrity of gold NRs.

12 Figure S6. (a) Extinction and (b) CD spectra of NR-free CNC films cast from the mixture of 5 ml of 2.44 wt.% CNC suspension and 1 ml of the NR-free supernatant collected after a different number of centrifugation cycles for the as-prepared solution of gold NRs. (c) Extinction and (d) CD spectra of NR-CNC films cast from a mixture of 5 ml of 2.44 wt.% CNC suspension and 1 ml of NR suspension after a different number of centrifugation steps. Inset in (d) shows zoomed in brown spectrum. 5.4. Spectral characterization of NR-CNC films Figure S7. Zoomed in CD spectra of the NR-CNC films, corresponding to the dark-green, blue and dark-blue spectra shown in Figure 4b (main text). The legends indicate C NR (wt.%). At C NR

13 of 2.29 and 3.39 wt.% the CD spectra show the TLSPR plasmonic peak centered at 493 and 490 nm, respectively.. Table S7. Spectral position of the extinction and CD bands of the NR-CNC films at different NR concentrations. C NR (wt.%) λ ext-llspr λ ext-tlspr λ CD-MAIN PEAK λ CD-TLSPR 0 - - > 800-0.12 747 522 665-0.23 752 521 645-0.47 752 520 619-0.93 755 520 619-2.29 756 521 608 493 3.39-523 608 490 Table S8. Spectral position of the extinction and CD bands of the NR-CNC films loaded with NRs with different dimensions. Series Aspect ratio of NRs λ ext-llspr (solution) / λ ext-llspr (films) λ ext-tlspr (solution) / λ ext-tlspr (films) λ CD * A 2.3 669 / 711 513 / 522 620 B 3.0 750 / 788 513 / 539 640 C 4.0 846 / 916 516 / 550 663 * The spectral position of the strongest CD peak Table S9. Variation of the spectral position of the extinction LLSPR mode ( λ ext-llspr ), CD peak ( λ CD ) and SEM-determined stop-band ( λ P-SEM ) between NR-CNC films with different NR dimensions. Series λ ext-llspr λ CD λ P-SEM * A-B 77 20 40 B-C 128 23 10 * Taken from Table S5.

14 Figure S8. Zoomed in CD spectra of the NR-CNC films containing NRs with dimensions of (a) 33 x 11 nm and (b) 36 x 9 nm (Series B and C, respectively). The legends show C NR (wt.%). The CD spectra show the TLSPR plasmonic peak of NRs centered at ~480 nm. Table S10. Spectral position of the LLSPR and TLSPR modes for NR-CNC films containing NRs with different dimensions whose CD spectra is shown in Figure S8. NR aspect ratio LLSPR TLSPR Extinction CD Extinction CD 3 (series B) 788 619 539 490 4 (series C) 916 663 550 491

15 λ sb 900 800 700 600 500 400 0.0 0.2 0.4 0.6 C NaCl (wt.%) Figure S9. Variation in the spectral position of the photonic stop band plotted as a function of the concentration of NaCl in the NR-free CNC films. Figure S10. CD spectra of the NR-free CNC film (dashed line) and NR-CNC film (solid line), both prepared at C NaCl =2.13 wt.%. The NR-free CNC film exhibits a weak CD peak at 475 nm corresponding to the reduced long-range cholesteric order in the film and a CD signal in the UV spectral range (originating from the chirality of glucose repeat units). The corresponding NR- CNC film exhibits only a CD signal in the UV spectral range (centered at 225 nm), originating from the chirality of glucose repeat units. The absence of plasmonic CD peaks indicates that films lacking a long-range chiral nematic order do not induce plasmonic CD activity.

16 Figure S11. (a) Extinction and (b) CD spectra of the NR-CNC films prepared at C NaCl =0.22 wt.%. The legend in (a) shows C NR (wt.%). (c) Zoomed in CD spectrum shown in (b) at C NR =2.53 wt.%. The existence of two well-defined CD peaks at 483 and 585 nm correlate with the results shown in Figure 6 (main text). The color of spectra in (a-c) correspond to the same NR concentration in the films. Table S11. Spectral position of the extinction and CD peaks of the NR-free and NR-CNC films prepared at different NaCl concentrations C NaCl λ ext NR-free CNC films λ CD NR-free CNC films / λ CD NR-CNC films λ P-SEM NR-CNC films * (wt. %) 0 837 > 800 / 637 691 0.02 748 730 / 613 608 0.06 637 610 / 576 582 0.13 521 520 / 481 + 565 528 0.22 451 445 / 443 + 559 460 ~2.13 - - - * Calculated in Table S6. Figure S12 shows CD spectra of the NR-free and NR-CNC films, which were acquired at different angles of incidence of light. At the angle of incidence of ~10º the CD spectra presented a Cotton effect. With an increasing angle of incidence of light the CD spectra of both NR-free CNC films and composite films (C NR =0.47 wt.%) exhibited red-shift in the position of CD peak, in agreement with Eq. 1, thereby confirming thus that the CD activity arose from the long-range

17 order of the photonic matrix, rather than chirality of glucose molecules. The CD spectrum of the composite films at C NR =3.39 wt.% (the highest NR content) did not exhibit red-shift, due to the weak contribution of the photonic matrix to the hybrid CD signal for films with a largely destroyed chiral nematic structure. Figure S12. CD spectra acquired at different angles of incidence of light [(dashed lines) ~10º, (dashed-dot lines) ~45º, and (solid lines) 90º] of (a) NR-free CNC film and (b,c) NR-CNC films. Legend shows concentration of NR [(41 x 14) nm] in wt.%. In b) and c) the intensity of the CD spectra was multiplied by a factor of 20. References 1. Dong, X. M.; Revol, J. F.; Gray, D. G. Cellulose 1998, 5, 19-32. 2. Nikoobakht, B.; El-Sayed, M. A. Chem. Mater. 2003, 15, 1957-1962. 3 Woolley, J. T. Plant Physiol. 1975, 55, 172-174. 4. http://www.filmetrics.com/refractive-index-database/au/gold. Reference: Sopra Material Library 5. Berova, N., Nakanishi, K., Woody, R. W. Circular Dichroism: Principles and Applications, Wiley, New York, 2000.