The Role of Surface Plasmon Resonance of Gold Nanoparticles for the Enhancement of Second Harmonic Generation of Nonlinear Chromophores

Abstract

A significant enhancement of the second harmonic generation (SHG) from second-order nonlinear optical (NLO) dyes in solution has been evidenced when these dyes were grafted onto the surface of spherical gold nanoparticles (AuNPs), as compared to their NLO response without AuNPs. The length, as well as the nature of the spacers between the gold particle and the chromophore, is shown to play an important role, a benzene linker being more favorable to AuNP-4-dimethylamino-N-methyl-4-stilbazolium tosylate (DAST) interactions, as illustrated by a higher exaltation of the NLO response for DAST molecules linked to AuNP via a phenyl ring as compared to their equivalent with a CH₂–CH₂ link.

1. Introduction

During the past decades, a great effort has been devoted to the synthesis of new materials with large second-order optical nonlinearities because of their potential use in applications such as high bit rate long-distance optical communications, and more generally, optical information processing [1]. The main advantages of organic second-order nonlinear optical (NLO) materials are their high NLO activity, chemical stability, and easy processability [2]. Among organic NLO materials, 4-dimethylamino-N-methyl-4-stilbazolium tosylate (DAST) is one of the most studied molecules. Due to its high activity, this organic salt has already been used for high-sensitive electric field sensors [3] and terahertz (THz) wave generation [4].

Gold nanoparticles (AuNPs) are attracting a great deal of attention due to their optical properties related to surface plasmon resonance (SPR), which depends, not only on the metal itself and on its environment, but also on the size and shape of the particle [5,6]. SPR is related to optically induced collective excitations of conduction electrons. A large enhancement is expected for electromagnetic fields through SPR [7]. Surface functionalization of gold nanoparticles with various ligands results in the generation of nanocomposites exhibiting optical properties with great potential for biological or medical applications [8,9], with the development of ultrasensitive detection and imaging methods, or regarding physical properties, such as magnetism [10] and linear and nonlinear optical properties [11]. It was demonstrated that the NLO activity of small metallic nanoparticles can be enhanced by deviation of the shape of the nanoparticles from a perfect sphere [12].

Various experimental studies have reported on the exaltation of quadratic NLO properties of molecular units by gold nanostructures. Exaltation of second harmonic generation (SHG) from ultrathin dye layers deposited on fractal gold surfaces has been reported [13]. A second harmonic generation (SHG) signal of individual molecules in the presence of very small AuNPs (1 nm) has been evidenced in the context of biological membrane imaging [14]. Large hyperpolarizability (β) values have also been reported for AuNPs in a solution using harmonic light scattering (HLS), a powerful technique to measure the β value of molecules or particles dispersed in solution [7,15,16,17,18,19]. Theoretical studies on SHG properties of gold nanoparticles have been implemented by several authors [20,21].

In this experimental work, highly nonlinear DAST derivatives have been grafted onto the surface of gold nanoparticles by two different linkers and their NLO response was studied and compared.

2. Results

2.1. UV–Visible Spectroscopy

Absorption spectra for all compounds were recorded in N,N-dimethylformamide at room temperature, with all samples converted as hexafluorophosphate salts. Spectral data of nongrafted DAST derivatives 1, 3, 4, sets of nanoparticles-dye associations Au-3 and Au-4, and the dye/Au-SPh mixture Au-1 (see their structure in the Materials and Methods) are presented in Table 1 and typical spectra are shown in Figure 1.

Electronic absorption spectra of molecules 1–4 showed an intense and broad band in the visible region, corresponding to the π(–NMe₂) → π* (pyridinium) intramolecular charge transfer (ICT) excitation from the –NMe₂ electron donating groups to the pyridinium acceptor moiety. Absorption maxima for the three ligands were centered at 485 nm (1, ε = 34,200 L·mol⁻¹·cm⁻¹; 3, ε = 28,915 L·mol⁻¹·cm⁻¹; 4, ε = 27,100 L·mol⁻¹·cm⁻¹), which were slightly shifted compared to the maximum observed for DAST (469 nm). These bathochromic shifts of the ICT band were in line with results previously reported in the literature for DAST derivatives bearing aromatic groups in the acceptor moiety [22]. It was attributed to an improvement in the accepting ability of the pyridinium moiety by a stronger electronic deficiency. Study of the electronic absorption spectra of all Au⁰ samples revealed full information about the final nanocomposites. For Au-3 and Au-4, the band detected in the spectral range 370–600 nm was resulted from a double contribution, one around 461 nm for Au-3 and 474 nm for Au-4 and the second one, with a similar intensity compared to the first band, around 541 nm for Au-3 and 533 nm for Au-4. The first band corresponded unambiguously to ICT band of the chromophores and the second band to the surface plasmon band (SPB) of Au-NPs. A hypsochromic shift of the ICT band and a bathochromic shift of the SPB were then detected after covalent linkage of the chromophores to Au-NPs. For Au-1, the ICT band was not shifted and more intense due to the overlap with the SPB of gold nanoparticles. This result was consistent with a physical mixing of a chromophore with NPs, with the overall observed spectrum corresponding to the addition of the ICT band of the chromophore and the SPB band of the gold particles. For Au-3 and Au-4, only the ligands grafted onto the Au⁰ surface were detected and the ICT band was nearly as intense as the SPB band. A similar result has recently been obtained for other organic molecules [23]. A comparison of the SPB position for Au⁰-S-Tolyl without chromophores (521 nm) with the position observed for Au-4 (533 nm) and Au-3 (541 nm) confirmed the anchorage of the chromophores onto the Au⁰-NPs surface. The strong interaction between the chromophores and the nanoparticles was also evidenced by the shift of the ICT bands, especially for Au-3. It was certainly due to the close proximity of chromophore 3 to the surface.

2.2. Transmission Electron Microscopy

Transmission electron microscopy (TEM) images of all samples were obtained from the samples analyzed by HLS. No specific aggregation was observed. The TEM images showed an average core diameter of 4 ± 1 nm (Figure 2), consistent with the expectations from the experimental process. The size, shape, and dispersity of gold nanoparticles Au-1, Au-3, and 4 were similar to those observed for the initial dodecylamine-coated nanoparticles.

2.3. Nonlinear Measurements

We have compared the NLO response of the Au-3 and Au-4 composites with a simple mixture of the chromophore and AuNPs without covalent linkers or electrostatic interactions (sample Au-1). The results are displayed in Table 1.

3. Discussion

In spite of the fact that the plasmon resonance wavelength of AuNPs lies far from that of the second harmonic signal at 820 nm, we evidenced a clear exaltation factor (by a factor of 3 for Au-4) of the hyperpolarizability of stilbazolium derivatives when they were grafted to a gold nanoparticle. The nature of the linker has a significant influence on β value. In spite of the larger distance (about 5.8 Å vs. 3.8 Å for Au-3) separating AuNP to the nitrogen atom of the pyridinium moiety in DAST molecule in the case of Au-4 as compared to Au-3, the phenyl group favored the interaction of the chromophore with gold NP compared to the CH₂–CH₂ linker. This behavior does not seem to reflect the expected exponential decrease of I(2ω) after increasing the distance between the dye and the gold surface, as evidenced in Reference [24]. However, our configuration differs from that of [23], as in our case the dye-gold surface distances are very small, and the spacers between the DAST derivative and gold surface are strongly different. In the case of Au-4, the π-electrons of the conjugated benzene ring seem to interact significantly with both pyridinium and AuNP electrons, resulting in a higher interaction between DAST and electrons of Au-NP, and hence to a higher exaltation of the NLO response of the DAST derivative, prevailing over the distance effect. A weaker, however, detectable magnification was also observed for a simple mixture of stilbazolium ligand and Au-NPs functionalized with thiophenol (Au-SPh).

4. Materials and Methods

Materials: All reagents were of high-purity grade and used without further purification and purchased from Sigma-Aldrich Chimie (Lyon, France), ACROS Organics (Fisher Scientific, Illkirch, France), and from CLAL (Paris, France).

4.1. Synthesis of AuNPs

AuNPs were synthesized by radiolysis [25]. Radiolysis has proven to be a very efficient technique for the synthesis of nanoparticles of different aspect ratios, controlled size, and morphology. The effect of the interaction of high-energy radiation such as gamma rays with a solution of metals ions induces ionization and excitation of the solvent and leads to the formation of molecular and radical species throughout the solution. The gamma-irradiation source was a ⁶⁰Co gamma-ray source of 7000 Curies with a dose rate of 1.75 Gys⁻¹ (6300 gyh⁻¹) at University Paris Sud. The deposited energy into the irradiated medium is expressed in Grays, 1 Gy = 1 J·kg⁻¹ (for aqueous solutions, 1 Gy corresponds to 1 J·L⁻¹).

HAuCl₄ and the stabilizer were added to a stirred aqueous solution containing 0.1 M of 2-propanol (added to scavenge HO· radicals). The solutions were bubbled with N₂ before irradiation. Au complexes were homogeneously reduced by solvated electrons and alcohol radicals.

4.2. Functionalization of AuNPs with DAST

A highly active NLO chromophore derived from 2 has been grafted onto the surface of spherical gold nanoparticles by two different linkers and has also been mixed with spherical AuNPs without covalent linker or electrostatic interactions. When the two types of linkers were synthesized, molecule 2 was transformed into a stilbazolium derivative by grafting a mercaptoethyl ethyl group in 3 and a mercaptobenzyl group in 4 (Figure 3). The presence of the SH moiety allowed the functionalization of dye molecules onto the gold surface. Chromophores 3 and 4 were prepared starting from 2 [26] in three and two steps, respectively.

For the chromophore and AuNPs mixture with no specific interactions, we used 1 as this chromophore exhibits a structure similar to 4. It was synthesized in one step from 2 and p-bromomethyl toluene.

In order to obtain gold nanoparticles with similar sizes and shapes to compare NLO coefficients, only one assay of dodecylamine-capped AuNPs was prepared in cooled dichloromethane (DCM) and used during the postfunctionalization step. Monodisperse gold nanoparticles were synthesized in DCM by reduction of a tetrachloroaurate salt (AuCl₄⁻) with sodium borohydride in the presence of the stabilizing agent (dodecylamine), according to the procedure of Brust and coworkers [27]. Subsequently, thiol-functionalized gold nanoparticles Au-3 and Au-4 were obtained through ligand exchange reaction of the stabilizing agent with molecules 3 and 4 previously dissolved in N,N-dimethylformamide (DMF). This exchange was facilitated by the weak stabilization ensured by the amine and a stronger affinity of the thiol group for the gold surface. In the third set of Au⁰-NPs (Au-1), Au⁰ surface was first capped with thiophenol by ligand exchange followed by addition of 1. This addition was done in such a way that the amount of 1 in the assay was similar to that observed for samples Au-3 and Au-4. The final step of the preparation consisted of precipitation of the nanoparticles and their redispersion in DMF, except for Au-1 for which the solvent was evaporated and the residue dissolved in 5 mL of DMF. The final structures of the samples are shown in Figure 4.

4.3. NLO Experimental Set-Up

The harmonic light scattering (HLS) method at λ = 1640 nm was used for β measurements, because the SH wavelength at 820 nm lies far from 2-photon resonance with the absorption spectra of DAST and Au-NPs. The fundamental beam was emitted by an optical parametric oscillator (OPO) pumped at 355 nm by a frequency-tripled Nd³⁺: YAG ns laser at a 10-Hz repetition rate. The incident laser beam was passed through a half-wave plate in order to continuously change the incident fundamental intensity between two crossed Glan polarizers. A small part of the incident beam was taken by a glass plate and sent onto a highly nonlinear NPP (N-4-nitrophenyl-prolinol) powder used as a reference frequency doubler. The emitted second harmonic signal was detected by a photomultiplier. The NLO measurement set-up is represented in Figure 5.

AuNPs functionalized solutions with NLO-dye were placed in a 1 cm long fluorescent cell, at the focus of the incoming fundamental beam. Harmonic incoherent emission was collected in a 90° configuration using two convergent lenses, detected by a photomultiplier, sampled and averaged using a Stanford boxcar integrator, and processed by a computer. The fundamental beam was focused into the fluorescence cell using a 5-cm focal length converging lens. Solutions were preliminary cleaned through 0.45 µm Millipore^® filters in order to remove most microscopic particles which could otherwise induce breakdowns in the presence of the focused IR laser beam.

Variation of the scattered second harmonic intensity from the solution is recorded on the computer as a function of the reference second harmonic signal provided by the NPP powder (I_NPP^2ω), which scales like the square of the incoming fundamental intensity I(ω). For a two-component (solvent and analyte) solution, the HLS intensity I(2ω) is given by

I(2ω) = G < N_s β_s² + N_a β_a² > I² (ω) e^(−Nsαl)(1)

where G is a parameter that accounts for the instrument and local field factors, N_s and N_a is the number of solvent and analyte molecules per unit volume, β_s and β_a is the molecular hyperpolarizability of the solvent and analyte molecule, α is the molar absorption coefficient of the solution, l is the path length of the optical cell, and I(ω) is the intensity of the incident light. To obtain β_a values, it is necessary to work in two steps. First, the second-order intensity of the solvent only is measured as a function of (I_NPP^2ω), giving a slope p_s. The same experiment is then carried out with the solution of analyte, resulting in a slope p_a. Knowing β_S value for the solvent, β_a is determined by calculating the ratio of the two slopes.

5. Conclusions

In this article, we have evidenced the positive role of plasmonic effects in the enhancement of the microscopic nonlinear response of chromophores in the proximity of gold nanoparticles, in spite of the fact that the fundamental and harmonic frequencies are remote from nanosphere plasmonic resonance. These NLO nanostructures offer interesting perspectives in the domain of ultrasensitive detection and imaging methods. Further studies will focus on the influence of the different sizes of AuNRs on the NLO response and functionalization of AuNRs with different NLO dyes.

Author Contributions

Synthesis of chromophores, functionalization on gold nanoparticles, UV–visible spectroscopy: F.D., A.G., E.D., C.R.M.; nonlinear optical measurements: A.S.B., K.H.-T., I.L.-R.; synthesis and TEM characterization of gold nanoparticles: H.R. writing and editing: I.L.-R., C.R.M.

Funding

This research was funded by C’NANO IDF, CNRS, the ANR agency (AURUS program), the PRES UniverSud and the PHOREMOST European Network of Excellence for financial supports.

Conflicts of Interest

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Figures and Table

Figure 1. UV–visible absorption spectra of molecules 1 (continuous line), 3 (dotted line), 4 (thick, continuous line), dye-functionalized nanoparticles Au-3 (dotted-dashed line) and Au-4 (dashed line), and the Au-1 sample made of a mixture of 1 and Au-SPh without link between dye 1 and gold nanoparticle (double-dotted-dashed line).

Figure 2. TEM image of Au-SPh nanoparticles; scale bar 30 nm.

Figure 3. 4-dimethylamino-N-methyl-4-stilbazolium tosylate (DAST) derivatives with linkers.

Figure 4. Structures of Au-3 and Au-4 functionalized gold nanoparticles (AuNPs) (top) and composition of the Au-1 sample (bottom).

Figure 5. A schematic diagram of HRS measurement at 1.64 µm. F: Filter; M: Mirrors; A: Attenuator; GP: Glass plate; L: Lens; P: Polarizer; λ/2: Half wave plate; PMT: Photomultiplier tube; NPP (N-4-nitrophenyl-prolinol): Reference material.

Table 1

Maximum absorption wavelength λ_max and hyperpolarizabilities λ_max per dye unit for dyes 3 and 4, dye–nanoparticle associations Au-3 and Au-4, nonfunctionalized gold nanoparticles Au-SPh, and Au-1 mixture without link between dye 1 and Au-SPh. Hyperpolarizability measurements were performed at 1.64 µm.

Samples	λmax (nm)	Β (×10−30 esu)
3	483	790
4	483	840
Au-3	461, 541	1800
Au-4	474, 533	2380
Au-1	482, 526	1580
Au-SPh	526	260