Influence of the pH value of a colloidal gold solution on the absorption spectra of an LSPR-assisted sensor

In recent years, gold nanoparticles and other noble metal nanoparticles have attracted much attention from researchers, particularly with reference to the electronic and optical properties of their colloidal solutions.^1–3 Localized surface plasmon resonance (LSPR) nanosensors based on metal nanoparticle arrays are effective platforms for the detection of biological targets. Selecting optimized nanoparticle structures can improve the sensitivity of LSPR-based sensors.⁴ In many studies, the desired architecture consists of a two-dimensional array of nanoparticles supported on a substrate.^5–10 However, relatively few studies have investigated the factors that govern the nanostructure within this geometry. The optical properties of noble metal nanoparticles are determined by the size, shape, dielectric properties, and local environment of the nanoparticle.^13–15 Grabar's group^11,12 reported that the assembly time had an effect on the particle coverage and the interparticle spacing, but the effects of other factors have not been reported.

We perform a detailed study of the kinetics of the binding of colloidal gold particles to (3-aminopropyl)-trimethoxysilane (APTMS)-coated transparent quartz substrates. Colloidal gold sols were prepared via the citrate reduction of HAuCl₄ in aqueous solution. We aimed to change the relative rates of the two independent nucleation and growth processes of the gold particles by changing the relative amounts of reactants. The pH value is an important environmental factor, especially in chemical and biomedical systems. The pH value affects the dissociation of functional groups on the surface of self-assembled monolayers, and the attachment of gold particles to the substrates; consequently, changes in the pH result in changes in gold particle coverage.

The fabrication of the self-assembling gold surfaces consisted of two major steps: (1) Preparation of gold colloids in pH solutions varying from 4.0 to 10.8. (2) Self-assembly of the gold nanoparticles on transparent quartz substrates.

HAuCl₄·3H₂O (Alfa, 97% purity), trisodium cirate dihydrate (Alfa, 99% purity) and APTMS (Alfa, 97% purity) were used to prepare gold colloids. Substrates were 330-μm thick blank AT-cut quartz plates with 1 inch diameter. Both surface of the quartz plates were carefully mechanically polished to obtain smooth surface.

Gold colloids were prepared by citrate reduction of HAuCl₄ in aqueous solution. Aqueous HAuCl₄·3H₂O (1% g/mL, 0.6 mL) was added into deionized water(60 mL) in a round-bottom, scratch-free flask, that had been washed in aqua regia and rinsed in water. When the solution was heated to boiling, the solution of citrate ion (trisodium citrate dihydrate, 1% g/mL, 2 mL) was injected. The solution changed from yellow to red.

Acidic buffer solutions (pH 3.5–7.0) were prepared by combining different proportions of trisodium citrate dihydrate and citric acid in aqueous solution. Highly alkaline pH buffer solutions (pH > 10) were prepared by adding sodium hydroxide to an aqueous solution.

To investigate LSPR properties of our gold particles, the gold sols prepared at different pH values need to remain stable. We chose citric acid and trisodium citrate dihydrate to prepare the acidic pH buffer solutions because: (1) The ions adsorbed on the surface of gold particles are OH⁻¹ and citrate ions. Therefore the buffer solution does not introduce new species. (2) Citrate reduces the impact of counter ions on the double ionization layer, and enhances the stability of colloidal gold sols. Our tests showed that gold sols with a pH of 3.5 that were adjusted using the buffer solution remained stable for 6 h, and only then did the color of the sol change to blue; in contrast, gold sols with a pH of 3.5 that were adjusted using an aqueous HCl solution immediately aggregated. The pH value of the gold sols changed significantly after the buffer solution was added. For example, the pH value of the solution increased to 3.9 after gold sols (8 mL) were added to a buffer solution (0.5 mL) of pH 3.5, because the low concentration of the buffer solution led to a weak buffer capacity.

We monitored the changes in the maximum LSPR extinction curve by varying the particle diameter, particle size spacing, and molecular adsorbates on the gold colloids. Figure 1 shows electron micrographs of gold nanoparticles with three different sizes; these particles were prepared using reductant (trisodium citrate dehydrate)/gold ratios of 1.4:1, 3.5:1, and 4.6:1.^16,17 Figure 1(a) to 1(c) show gold particles with diameters of 16, 20, and 30 nm, respectively. The particle size increases when the amount of trisodium citrate dehydrate increased.

The influence of the pH of the gold sols on the particle coverage was investigated using AFM. The gold particles were distributed homogeneously on the surface of crystal plates. The coverage was measured as the number of gold particles in an area of 2 μm × 2 μm. Different regions on the surface of the crystal plate showed approximately the same coverage value. The experimental procedure was as follows: The pH value of the gold sols was adjusted to 4.0, 4.5, 4.8, 5.4, 6.4, and 10.8 by adding the pH buffer solution. The crystal plates were placed in gold sols with different pH values for 4 h. After 4 h, only the gold sols in the buffer solution of pH 10.8 aggregated and turned blue. The quartz plates were then removed from the sols/buffer solutions and cleaned with distilled water. The surface morphologies of crystal plates that were dipped in the gold sols with a reductant/gold ratio of 4.6:1 were observed using AFM. Figure 2 shows that the particle coverage varies significantly. When the pH value was increased from 4 to 8, the particle coverage decreased sharply. For pH values greater than 8, the particle coverage approaches 0. When the particle coverage reaches a maximum value, the factors affecting the particle coverage are the repulsive forces between particles, and the attractive interactions between particles and the substrate.¹² 

The pH value affects the binding affinity between the gold particles and functional groups. It also affects the electrostatic repulsion between the gold particles, and consequently affects the coverage. Some papers have reported the covalent binding of gold nanoparticles with amino groups. The amino groups provide a lone electron pair more easily in an alkaline solution than in an acidic solution, and consequently the gold particles assemble more easily on the substrate in the presence of an alkaline solution. However, our experimental results showed that it was difficult for the particles to assemble on the substrate in alkaline solutions; the particles assemble more easily on the substrate in acidic solutions. When the pH value was lower than 6, most of the amino groups were protonated, providing a maximum surface charge, and the particle coverage approached a maximum level. When the pH value was higher than 10, the amino groups were not dissociated, and the particle coverage was close to 0. In conclusion, we believe that the combination of the gold nanoparticles with the amino groups occurred via electrostatic interactions.

Figure 3 shows the absorption curves of the substrates from gold sols with different reductant/gold ratios at pH 6.4 and pH 4.8. The absorption peak intensity of the 16-nm gold nanoparticles is significantly higher than that of the 20-nm gold nanoparticles. The absorption peak intensity of the 30-nm gold nanoparticles is similar to that of the 20-nm gold nanoparticles. This dependence is governed by several factors. For very large metal particles, the number of nucleation centers provided by the gold sols is extremely low.¹⁸ With increasing particle size, the decreasing surface area may not have been large enough to absorb the same proportion of the wavelengths transmitted from the solution with the same intensity.¹⁹ Smaller particles aggregate more easily, and the absorbance peak red-shifts.

Figure 4 shows the in-solution, normalized absorption spectra for quartz substrates supporting assemblies of nanoparticles of three different particle sizes. In all three graphs, the absorption peaks showed a clear red-shift. These results demonstrated that the gold sols with a higher pH produced a larger red-shift. According to the classic Mie theory, the variations in the absorption peak position, the half peak width, and the peak intensity are affected by three factors: (1) the nanoparticle's size; (2) the conduction electron density of the particle surface plasma; and (3) the dielectric constant of the metal particles.^20,21 With increasing pH value, the electrophoretic migration rate increased. The particle surface adsorbed more negative ions, leading to stronger electrostatic repulsion between the particles. As a result, the particle spacing increased, and the particles did not aggregate easily. In contrast, the nanoparticles aggregated easily in the acidic solutions, and the absorption peak showed an obvious red-shift. Figure 4(a) shows that there was no significant absorption peak in the substrates produced using the gold sols with a pH of 10.8, because the particle coverage was close to 0 in the alkaline solutions.

This study compared the absorption spectra of LSPRs. By comparing the optical properties of substrates covered with gold nanoparticles of different sizes and with different particle coverage values, we showed that the optical properties of the gold particles depend on nanometer-scale architecture of the substrate surface. The pH of the colloidal gold sols significantly affects the particle coverage. In acidic solutions, most of the amino groups were protonated, and the gold nanoparticles combined with the amino groups more easily via electrostatic interactions. The absorption peak intensity varies with the nanoparticle size. The small nanoparticles showed greater absorption peak intensities. With increasing particle size, the decreasing surface area may not have been large enough to absorb the same proportion of the wavelengths transmitted from the solution with the same intensity. The results demonstrate that the pH of the gold sols determined the gold particle coverage, and also affected the red-shift of the LSPR band on the quartz substrates. This study suggests that the particle size and particle coverage are both important factors affecting the optical properties of LSPR-assisted sensors, and that rational nanometer-scale architectures will improve the sensitivity of LSPR-assisted sensors.

The authors thank the National Natural Science Foundation of China (Grant No.61271099), Natural Science Founation of Tianjin (12JCZDJC20400), and National Research Foundation for Doctoral Program of Higher Education of China (20120031110031) for their financial support to complete this work.