A Novel Physical Approach for Cationic–Thiolate Protected Fluorescent Gold Nanoparticles

Knowledge on the synthesis of cationically charged fluorescent gold nanoparticles (Au NPs) is limited because the electrostatic repulsion between cationic ligands on the surface of NP hinders the formation of small Au NPs (usually less than ca. 2 nm) during nucleation in solvents. We herein propose a novel methodology for a synthesis of water-dispersible, cationic–thiolate protected fluorescent Au NPs by the sputtering of Au into liquid matrix containing thiolate ligands. By controlling mercaptan concentration the size and photophysical characteristics of Au NPs were directly controlled, resulting in near IR fluorescence with a 0.9% of absolute quantum yield. Cationically charged fluorescent metal NPs are promising, especially in biological fields, and this work provides a novel methodology towards the synthesis of a new series of functional metal NPs.

Scientific RepoRts | 5:15372 | DOi: 10.1038/srep15372 ethoxylate 19 . Liquid matrixes in group (a) resulted in very small (~1.3 nm) fluorescent Au NPs, while those in group (b) formed rather large Au NPs (2-5 nm). The difference in their size was attributed to the coordination property of mercapto groups in group (a) that can prevent the coalescence of Au NPs inside or on the interface of liquid matrices. Moreover, this technique can be applied for various conductive materials such as Ag 22 and Cu 23 . From these results, it is hypothesized that we can prepare metal NPs of various sizes of the order of a single nanometer, and thus, can control the physical properties of NPs by matrix sputtering method.
In this paper, we present a first synthesis of cationically charged fluorescent Au NPs with precisely controlled diameters using the matrix sputtering method. While previous approaches have only used liquid matrices, our new approach uses both a liquid matrix and a solid organic molecule with mercapto groups. The organic molecule cannot be used as the matrix because it is in the solid form at room temperature with non-melting property (it decomposes at 173 °C without melting, see TG-DTA spectra shown in Fig. S1). The stabilizing molecules can dissolve in a liquid matrix and prevent the coalescence of Au NPs inside or on the interface of the liquid matrix, and their concentration successfully controlled the diameter of Au NPs of the order of a single nanometer. To the best of our knowledge, little is known about the synthesis of water-dispersible fluorescent Au NPs using cationic mercaptan ligands except for only a few reports 18 , because the electrostatic repulsion between cationic ligands on the surface of NP hinders the formation of small Au NPs (less than 2 nm) during nucleation in solvents. Cationically charged fluorescent Au NPs are promising especially in biological fields, for example, in bio-electronic devices and in targeted drug delivery 27,28 . The method presented here will provide a novel methodology for the synthesis of a new series of functional metal NPs.

Results and Discussion
The liquid matrix used in this work was diglycerol (DG, b.p. = 412 °C at 760 mmHg) in order to dissolve hydrophilic ligands for this study. The cationic stabilizer molecule, thiocholine chloride (TC, Fig. 1), was synthesized by the hydrolysis of commercially available acetylthiocholine iodide and then the counter ion was exchanged to Cl − according to a method described in a previous paper 29,30 . TC is solid at room temperature, and using TG-DTA analysis decomposition was observed at 173 °C instead of the observation of a melting point (Fig. S1). Thus, the use of liquid matrix (DG) is required for the sputtering synthesis of TC-stabilized Au NPs. The experimental setup for the sputtering synthesis is illustrated in Fig. 1. The concentration of TC in DG was set at 0, 0.01, 0.1, and 0.5 M, and TC completely dissolved in DG even at 0.5 M concentration. The thiol group of TC has a good affinity to Au due to the formation of a coordination bond between Au 0 and the lone pair of electrons of sulfur, or a covalent bond of Au-SR. It is expected that the steric hindrance by the coordination of TC suppresses the growth of Au NPs during the sputtering deposition. Judging from TC's salt molecular structure, it is expected that TC has almost no vapor pressure and thus the stabilization for nanoparticles should be a liquid process. This mechanism is rather different from our previous case, where α -thioglycerol ligand can coordinate to nanoparticles in a gas phase due to its relatively low boiling point 20 . It should be noted that Au thin film did not form on the liquid matrix, while it is usually deposited on solid surfaces for the purpose of SEM measurement. After the sputtering of Au into the matrix, extinction spectra of Au NPs suspension in DG were measured in a quartz cell with 1 mm optical path just after the sputtering preparation without further purification or dilution (Fig. 2). Broad absorption peaks around 520 nm, which are characterized as the localized plasmon absorption of Au NPs, were observed for 0 M and 0.01 M samples. On the other hand, the absorption spectra did not show any plasmon absorption peak for the 0.1 M and 0.5 M samples. Fundamentally, surface plasmon absorption originates from the vibration of free electrons on the surface of NPs resulting in an absorption peak corresponding to the vibration frequencies. Therefore, the generation of plasmon absorption needs a certain particle size (over ca. 2.4 nm for Au NPs, as reported in a previous paper 31 ). As the number of atoms in each particle decreases, the energy band gap becomes wider according to the quantum size effect. Then, plasmon absorption is not observable in such small particles. Judging from the extinction spectra in Fig. 2, the size of Au NPs would decrease with an increase in the TC concentration, and those prepared with 0.1 M and 0.5 M TC should be very small, and thus, no plasmon absorption was observed. The obtained Au NPs prepared in 0.1 M and 0.5 M TC were water-soluble, as judged by the addition of 3 mL of water into 300 μ L of DG dispersion. Figure 3 shows the representative transmission electron microscopy (TEM) images and size-distribution histograms of Au NPs prepared at 0 and 0.5 M TC. The average sizes of Au NPs prepared at 0 and 0.5 M of TC were 6.7 ± 3.2 and 2.0 ± 0.7 nm, respectively. This difference in their particle sizes reflects the change in plasmon absorption, as shown in Fig. 2. The change in particle size as a function of TC concentration is shown in Fig. 4. From Fig. 4, it is obvious that a higher TC concentration produces smaller Au NPs. This tendency can be simply explained by the collision probability between TC and Au NPs inside and at the interface of DG. Thus, this result clearly indicates that the size of Au NPs is controllable by the concentration of the thiol-stabilizer in the liquid matrix.
The absence of plasmon absorption and very small particle sizes led us to investigate the fluorescence property of Au NPs stabilized by TC. The excitation wavelength was set at 300 nm. Au NPs prepared at 0 and 0.01 M TC, which show plasmon absorption in Fig. 2, did not fluoresce. On the other hand, those prepared at 0.1 and 0.5 M TC showed fluorescence in the near IR region, as shown by the red line in  Fig. 5. The excitation maximum was observed at 326 nm; therefore, the Stokes shift of our Au NPs can be calculated to be 347 nm (1.96 eV = 1.58 × 10 4 cm −1 ). The large Stokes shift of Au NPs has been reported in our previous matrix sputtering synthesis 18 , and the reason was assumed to be the stabilization of the d band and/or the destabilization of the sp-conduction band in the Au NPs that usually results in a blue shift in the absorption (excitation) spectrum.
The diameter of current TC stabilized Au NPs is considerably larger than usual fluorescent Au nanoclusters synthesized by chemical-reduction method [32][33][34][35] . Typically, Au NPs with ca. 2 nm of diameter do not fluoresce by a conventional chemical reduction method. Only the very small Au clusters such as Au 8 , Au 25 etc. fluoresce in different wavelength regions according to their core sizes. However, TC stabilized Au NPs with a 2 nm of diameter showed near IR fluorescence with a relatively large Stokes shift. This diameter is significantly larger than that reported previously for fluorescence Au clusters. By a comparison with known Au clusters prepared by a chemical reduction method, the photophysical characteristics of our TC stabilized Au NPs are similar to those of a Au 6 complex (excitation spectrum from 300 to 420 nm and fluorescence spectrum from 600 to 850 nm) reported in literature 34 . It should be noted that the polarity of thiolate ligands sometimes affects the fluorescence property of Au clusters 35 . The fluorescence quantum yield of the synthesized Au NPs prepared at 0.1 and 0.5 M TC was measured using the absolute method. Both samples showed 0.9% quantum yield, and this value is similar to those of gold nanoparticles and nanoclusters prepared by various synthetic methods including sputtering and chemical reduction methods [32][33] . These photophysical characteristics (fluorescence and excitation  wavelengths, fluorescence quantum yield) remained even when the solution was diluted in water by 10 times (v/v).
In conclusion, we presented a novel synthesis of water-dispersible, cationic thiolate-stabilized fluorescent Au NPs with precisely controlled diameters by the sputtering of Au in DG containing TC. While only liquid matrix has been used in previous sputtering methods, both liquid matrix (DG) and organic molecules with mercapto groups (TC) are used in our new approach. The size and photophysical characteristics of Au NPs could be directly controlled by controlling the mercaptan concentration, resulting in a near IR fluorescence with 0.9% quantum yield. Cationically charged fluorescent Au NPs are promising, especially in biological fields [27][28] ; therefore, this work will provide a novel methodology for the synthesis of a new series of cationic, functional metal NPs.

Methods
Diglycerol and acetylthiocholine iodide were obtained from Aldrich. Thiocholine chloride was synthesized by the hydrolysis of acetylthiocholine iodide and then the counter ion was exchanged to Cl − (Organo Amberlite IRA400JCL) according to a method described in a previous paper [28][29] . The purity of TC was confirmed by FT-IR and 1 H-NMR. Au target (99.9%) was supplied by Tanaka Precious Metals (Japan).
The experimental procedure for sputtering synthesis is described below. Firstly, to remove volatile substances, DG and TC were dried under vacuum at 100 °C for 2 h under stirring. DG (12.8 g = 10 mL) and the corresponding amount of TC was put in a glass petri dish with a diameter of 6.5 cm, and was horizontally positioned against the sputtering target. The concentration of TC in DG was set at 0, 0.01, 0.1, and 0.5 M (corresponding to 0, 15.6, 156, and 779 mg of TC, respectively), and TC completely dissolved in DG even at 0.5 M concentration. Au was sputtered using a current of 30 mA under Ar at a pressure of 20 Pa. The distance between the surface of DG and the surface of the gold target was 25 mm. Sputtering was carried out for 30 min at room temperature under stirring at 100 rpm. Excitation spectrum was observed at 673 nm. Fluorescence spectrum was observed by the irradiation at 300 nm. For a comparison purpose, the extinction spectrum under the same condition is shown by a black line.