**3. Results**

Au10 clusters capped with d- or l-histidine were obtained according to a previously reported procedure [18]. The imidazole group of histidine plays a key role in AuNCs formation, as it serves as both a reducing agen<sup>t</sup> and a protecting ligand. The clusters showed an absorption edge at around 450 nm that rose very steeply below 320 nm (full line spectrum, inset of Figure 1), together with the typical fluorescence emission centered at 500 nm, as previously reported [19]. No significant spectroscopic differences were observed for the Au10 sample stabilized by the two different histidine enantiomers. The chiroptical properties of histidine AuNCs were investigated by circular dichroism spectroscopy, confirming that the chirality of the histidine ligands was transferred to the metal core [19]. When TPPS (5 μM) was added to the AuNCs solution, the electronic spectrum for both histidine enantiomers showed the presence of a porphyrin Soret band centered at 418 nm and four bands located

in the visible region at 517, 554, 587, and 642 nm (red spectrum, Figure 1). The positions of both the B and visible bands were bathochromically shifted compared to those of the free base TPPS in aqueous solution in the presence of histidine [51].

**Figure 1.** UV–vis spectra of tetrakis(4-sulfonatophenyl)porphyrin (TPPS)–Au10 in aqueous solution (red line): instantaneously upon HCl addition (pH 2) (green line), after 1 h (blue line), after 24 h (cyano line). The inset shows the crude Au10 NCs for the same experimental conditions in aqueous solution (black full line) upon HCl addition (pH 2) instantaneously (black dashed line) and after 24 h (black dotted line). Experimental conditions: [Au10] = 125 μM, [TPPS] = 5 μM.

Furthermore, the CD spectra of the two enantiomeric forms showed a signal in the porphyrin absorption region, indicating the occurrence of an asymmetrical perturbation resulting from the chiral AuNCs. The CD spectra (inset of Figure 2) performed for l- and d-His-AuNCs enantiomers showed signals in the porphyrin region that were correlated with the enantiomorphism of the stereocenter present on the AuNCs protecting ligand, with a negative and positive bisignate Cotton effect for l- and d-histidine, respectively. This suggests that chiral information was successfully expressed on the porphyrin chromophore at the supramolecular level. The low induced CD intensity was in line with the monomeric nature of the chromophore. Recently, it has been reported that no interaction between TPPS and histidine has been detected in aqueous solution, mainly due to electrostatic repulsions between the negatively charged porphyrins and the amino acids [51].

**Figure 2.** Circular dichroism (CD) spectra of Au10 in aqueous solution (black line), Au10–TPPS Jagg nanohybrid assemblies for l-(blue line) and d-histidine (red line) before (inset) and 1 h after HCl addition. ([Au10] = 125 μM, [TPPS] = 5 μM, pH 2).

This was confirmed by the absence of any CD signal induction in the porphyrin region. Nevertheless, as has been observed for other systems, the absorption band shifted and the induced

CD signal could be ascribable to electrostatic interactions or hydrogen bonds between the negative sulfonated groups of the macrocycle and the protonated amino groups of histidine, and/or to the localization of the porphyrin in a hydrophobic microenvironment due to the histidine surrounding the metal nanostructures [52–55]. Moreover, the formation of positively charged histidine oligomers with Cl− as a counter anion has been reported [18]. Therefore, it also cannot be excluded that electrostatic interactions between the positively charged histidine oligomers and the negatively charged sulphonate groups present in the periphery of the macrocycle were the origin of the spectral variations. It should be mentioned that, upon addition of the positively charged tetra N-methylpyridinium porphyrin (H2TMPyP<sup>4</sup>+) to the Au10 solution, the Soret band red-shifted to 427 nm with respect to the free base (422 nm) (ESI, red spectrum, Figure S1). This shift can be explained by an interaction of the carboxylic groups of histidine with the positively charged N-methylpyridinium groups at the meso position of the macrocycle. Since histidine Au10 clusters can be successfully prepared over a wide pH range (pH 2–12) [18], we decided to foster porphyrin aggregation by lowering the pH. When HCl (pH 2) was added to the TPPS Au10 cluster solution, the UV–vis spectra showed an almost instantaneous formation of diacid TPPS with a Soret band at 435 nm (green spectrum, Figure 1), which slowly interconverted into J-aggregates characterized by a narrow peak located at 492 nm. Furthermore, the electronic spectrum at the end of the aggregation process showed a very broad band at around 550 nm not ascribable to porphyrin features (blue spectrum, Figure 1). For comparison, a control experiment was performed in the same conditions in the absence of porphyrin. In this case, acidification of the Au10 cluster aqueous solution instantaneously induced the formation of a new band at 400 nm (dashed line spectrum, inset Figure 1) that broadened and red-shifted within 1 h to around 550 nm (ESI, red spectrum, Figure S2), and eventually it led to a general increase of the baseline within 24 h (dotted line, inset Figure 1). This effect is ascribable to the growth of AuNcs to form AuNPs, which show the typical plasmon resonance band [56]. The process was accompanied by a color variation of the solution from pale yellow to pink. It is reasonable to expect that upon acidification of the Au10 cluster solution, larger metallic structures were produced by aggregation of the clusters, which resulted from hydrogen bond formation between the carboxylic and protonated amino groups of adjacent units of the histidine residues covering the gold surface, as already reported for AuNPs stabilized by histidine [57]. In fact, in the case of heterocycle amines such as pyrrole or tryptophan, the occurrence of macroscopic segregation of the polymer due to the amine oxidation process and the metal phase has been reported as a function of the heterocycle amines/gold stoichiometric ratio [58]. After removing the solution from the cuvette, a purple deposit was observed on the quartz surface (ESI, Figure S3). The corresponding electronic absorption spectrum, obtained from washing the cuvette with water, exhibited a broad band that extended from ca. 550 nm to higher wavelengths very similar to that observed 24 h after HCl addition (inset Figure 1). This effect is ascribable to the adhesion of the metal nanoparticles to the quartz surface. A similar experiment was carried out at the same pH value, reducing the amount of gold clusters. In this case, the UV–vis spectra evolved from the initial 400 nm band to a broad feature extending from ca. 550 nm to the near-IR, but neither the 530 nm band due to AuNPs nor flocculated material were detected (ESI, Figure S4). This suggests the direct growth of small gold clusters into larger entities with nucleation on the surface, without the formation of larger metal colloidal suspensions in solution. This observation may result from slower kinetics related to a lower amount of Au10 clusters, thus improving the homogeneity of the deposited material onto the quartz surface. At a higher Au10 load, the formation of solid material could be ascribed to a faster growth process causing the formation of much larger metal structures that become unstable in solution and eventually precipitate. It is noteworthy that no metallization of the cuvette wall occured when Au nanoparticles synthesized by standard reduction with NaBH4 in the presence of histidine as capping reagen<sup>t</sup> were used under the same experimental conditions (ESI, Figure S5) [59]. In the presence of TPPS, a further broadening of the plasmonic band was observed over a period of 24 h, together with a drastic decrease of band intensities. This was due to the formation of a dark precipitate, as confirmed

by the electronic absorption spectrum recorded the day after, in which an increased baseline was evident (cyano spectrum, Figure 1).

The resonance light scattering (RLS) spectrum (black spectrum, Figure 3) of the Au10 clusters in aqueous solution in the absence of porphyrin displayed a Rayleigh scattering profile. After addition of TPPS, a well at 418 nm appeared due to absorption resulting from the presence of porphyrin in its monomeric form (green spectrum, Figure 3).

**Figure 3.** Resonance light scattering (RLS) spectra of Au10 in aqueous solution (black line), TPPS@Au10 in aqueous solution (green line), upon addition of HCl (pH 2) after 1 h (blue line) and after 24 h (red line). ([Au10] = 125 μM, [TPPS] = 5 μM).

Upon addition of HCl, the RLS spectrum of the sample showed a peak at 500 nm due to the presence in solution of large ordered J-aggregates, stabilized by a network of electrostatic and solvophobic interactions among porphyrins (blue spectrum, Figure 3). After 24 h, the Rayleigh scattering due to the presence of larger gold structures in solution formed by aggregation phenomena occurring in acidic medium, modulated by the absorbance of the sample, was the only detectable component present (red spectrum, Figure 3).

As the Au10 clusters can be stabilized by either l- or d-histidine, in order to confirm the involvement of the chiral capping agen<sup>t</sup> in the aggregated samples, we performed circular dichroism experiments. As previously described, chirality is transferred from the histidine capping agen<sup>t</sup> to the monomeric porphyrin. Upon acidification and aggregation, chirality induced by the two histidine enantiomers was observed, analogously to data already reported for similar systems in the absence of gold nanoparticles [51]. The CD spectra for solutions of both cluster types showed the presence of a weak bisignate Cotton effect, centered in the J-aggregate absorption region (490 nm). d- and l-histidine led to almost mirror image spectra, characterized by a negative and positive Cotton effect (Figure 2). Moreover, a strong light scattering contribution due to the presence of porphyrin aggregates and gold nanostructures in solution broadened the bands.

As already observed for crude AuNCs, after removing the TPPS–AuNP solution from the cuvette, a purple deposit was observed on the quartz surface. The corresponding electronic absorption spectra obtained for both enantiomers showed, in addition to the spectroscopic features of the AuNPs, the typical J-aggregate band at 491 nm (Figure 4).

The RLS signals confirmed the presence of nanostructured material attached to the quartz substrate (Figure 4, inset). The preferential adsorption of the porphyrin J-aggregates on the cuvette surface was further proven by the observation of a residual amount of diacid monomeric TPPS in solution. After transferring the solution into a new cuvette, only the spectrum corresponding to this species could be detected, with the Soret band at 435 nm and with no CD features detectable (ESI, Figure S6). Interestingly, the CD spectra of these films in the J-aggregate spectroscopic region showed an induced bisignate Cotton effect related to the configuration of the amino acid used in the synthesis of the metallic clusters (Figure 5). These dichroic signals cannot be ascribed to linear dichroism, as they do not depend on the cuvette orientation with respect to incident light (ESI, Figure S7). No CD signal was detected for the corresponding AuNP deposit grown from d- and l-histidine (Figure 5, inset).

**Figure 4.** UV–vis spectra of the TPPS@AuNPs deposit left on the cuvette surface, obtained after washing with water, for d- (**black line**) and l- (**blue line**) histidine samples. The inset shows the corresponding RLS spectra.

**Figure 5.** CD spectra of TPPS@AuNP co-deposit for d-(**black line**) and l-(**blue line**) histidine samples. The inset shows the corresponding AuNP deposit grown from d-(**black line**) and l-(b**lue line**) histidine.

*Raman and SERS measurements.* The co-deposition of J-aggregated porphyrins and gold nanoparticles prompted us to investigate the possibility of detecting surface-enhanced Raman scattering (SERS) effects. In order to quantify the SERS effect, nitrate was added to the solution and the intensity of the Raman band at 1053 cm<sup>−</sup><sup>1</sup> was taken as an internal standard. No significant spectral changes were introduced due to the presence of the internal standard. Figure 6 shows representative Raman spectra of the J-aggregates obtained with and without Au10 clusters for excitation at 514.5 and 785 nm. The spectra of the samples with (Figure 6a,c) and without (Figure 6b,d) Au10 clusters were obtained at the same point on the surface of each sample for the two excitation wavelengths. The presence of Au10 clusters did not alter the frequency or the relative intensity of the J-aggregate bands, as the spectra were very similar to those previously reported in solution [60]. This result confirms that co-deposition of the J-aggregated porphyrins and the gold nanoparticles generally does not alter their structural integrity, as was previously found for the deposition of the solution phase aggregates onto a gold substrate [46]. However, an intensification of the overall porphyrin spectrum was evident in the samples containing Au compared to those without Au. The effect was considerably more evident in the spectra obtained with 785 nm excitation (Figure 6c,d) than those obtained with 514.5 nm (Figure 6a,b). In fact, in the latter case, the resonance Raman intensification of the porphyrin bands due to the vicinity of the

excitation wavelength with the 490 nm aggregate band precluded an accurate evaluation of the SERS effect. In order to quantify the intensification of the porphyrin spectrum, we evaluated the intensity ratio (R) of the band of the internal standard and isolated bands of the aggregate. The spectra were collected on NO3 − crystals, which displayed both NO3 − and the J-aggregate Raman signals. Two isolated bands of the porphyrin were used, at 317 and 1232 cm<sup>−</sup>1, corresponding to the out-of-plane vibration involving motion of the pyrrolic hydrogens and to the totally symmetric C m-phenyl stretch, respectively [32]. Similar results were also obtained using the 993 cm<sup>−</sup><sup>1</sup> band of SO4 2−, however, in this case, the band of the internal standard overlapped with the band at 986 cm<sup>−</sup><sup>1</sup> of the J-aggregate (data not shown). In the presence of Au10 clusters, the J-aggregate spectrum was intensified compared to the 1053 cm<sup>−</sup><sup>1</sup> band of NO3 −.

**Figure 6.** Representative Raman spectra of porphyrin J-aggregates in the presence (**<sup>a</sup>**,**<sup>c</sup>**) and absence (**b**,**d**) of Au10 clusters with excitation at 514.5 and 785 nm. The spectra of each sample (**<sup>a</sup>**,**<sup>c</sup>**) and (**b**,**d**) were obtained at the same point on the sample surface for the two excitation wavelengths. The inset of spectrum (**d**) has been multiplied 10-fold. Experimental conditions: (514.5 nm): (**a**) average of twelve spectra with 10 s integration time; (**b**) average of eleven spectra with 10 s integration time; (785 nm): (**c**) average of seventeen spectra with 30 s integration time; (**d**) average of nine spectra with 30 s integration time. All the spectra have been normalized with respect to the band of nitrate at 1053 cm<sup>−</sup>1.

The SERS intensification is evident from the marked reduction of the R values observed in the presence of Au10 (Table 1). The scatter of the R785 values for the sample without Au was likely due to variation in crystal size of the internal standard; the bigger/thicker the crystal, the more intense the salt band.

*AFM investigations*. To obtain insight into the structures of the J-aggregates and of the gold entities grown on the glass surfaces, the morphology of the deposited materials was examined by AFM. The samples consisted mainly of small AuNPs, which were aggregates formed by several sub-nanometer AuNCs. Irregular agglomerates, due to clustering of these small objects, can be also detected with average height ca. 60 nm and diameter 130 nm (ESI, Figure S8). These were accompanied by flat and regular objects, probably single AuNCs, of heights in the tens of nm and mean widths of a few hundred nm. In addition, elongated TPPS J-aggregates formed on the glass surface, displaying lengths that ranged between 100 and 1000 nm, 3–4 nm height, and ca. 80 nm width (Figure 7 and ESI, Figure S9). These results are consistent with bundles of collapsed J-aggregates due to the solvent evaporation from the inner compartment of the nanotube [61].


**Table 1.** Comparison of the intensity ratio (R) of the 1053 cm<sup>−</sup><sup>1</sup> NO3 − band and the 317 and 1232 cm<sup>−</sup><sup>1</sup> bands of the J-aggregates obtained with 785 and 514.5 nm excitation.

**Figure 7.** AFM topography images (x gradient) of the deposits formed on a glass substrate immersed in a Au10 and TPPS solution at pH 2.0, showing Au nanostructures (dark arrows) with J-aggregates (white arrows). Experimental conditions: (**a**) 9 × 9 μm, Z range = 478 nm; (**b**) 2 × 2 μm, Z range = 145 nm.
