The Au nanoparticle dimers were produced through ligand-exchange of a nanoparticle surface using a toluene–water interface and a short sonication time, while the aggregation was dominantly formed when the SAM layered nanoparticles were ligand-exchanged in the single aqueous phase. As depicted in , the mixture of the 1-nonanethiol and the 1-mercapto-6-hexanol (MCH) in ethanol were first chemisorbed on Au nanoparticles overnight, resulting in the formation of the SAM on the whole surface of the Au nanoparticles. The MCH, which is a shorter ligand than the 1-nonanethiol, is known to create a relatively less stable SAM, therefore the mixed thiol ligands (with a 2:1 loading molar ratio of 1-nonanethiol to MCH) provide defect sites where the SAM molecules are to be replaced by aqueous tiopronin [24
]. Consequently, the SAM defects on the Au nanoparticles were ligand-exchanged with carboxylic acids (tiopronin) in water by sonication for 10 min. The ligand-exchanged Au nanoparticles were well dispersed in N,N′
-dimethylformamide (DMF) and then dimerized with an N,N′
-diisopropylcarbodiimide (DIC)-mediated amide bond formation. The gap size in a dimer should be less than 2 nm when ethylenediamine is used as a linker. A further increase in the particle–particle distance can be achieved by simply choosing longer diamine compounds.
The resultant assembly of the Au nanoparticles was visually confirmed by a transmission electron microscopy (TEM) (). As illustrated in these images, there are numerous assembled dimers and some unreacted monomers, but there are few higher-ordered assemblies (clusters). The overall dimerization yield was estimated utilizing the method by Shumaker-Parry and co-workers [13
]. Briefly, it was determined by counting at least 500 of the nanoparticles from the 25 TEM images that were randomly taken. The overall yield of the dimerization was ca 24%. This efficiency is quite high since our TEM samples contain all
of the particles, without any separation from the initial step. Even though Shumaker-Parry and co-workers produced 46% yield in case of 41 nm sized Au nanoparticles [13
], the washed particle number in the initial step was not counted and the yield is dependent on the nanoparticle size, and therefore it is difficult to compare it with our result. Our overall yield was similar when the concentration of aqueous tiopronin was over 2 mM. The dimerization through ligand-exchange under the same conditions, with the exception of the shorter reaction time (2 min) produced a ca 21% overall yield, but higher-ordered clusters were generated, in fact, twice as many as seen in the ligand-exchange assisted by sonication for 10 min due to the hydrophobic nanoparticle surface. Likewise, the ligand-exchange for a longer period of time (30 min) produced a ca 27% overall yield, but higher-ordered clusters were also generated, three times as many as observed in the ligand-exchange with sonication for 10 min. For a size effect on the dimerization, the dimerization of particles smaller than 60 nm under the same conditions produced a lot of clusters compared to those of 60 nm particles. In contrast to smaller particles (less than 20 nm in diameter), the dimerization of particles larger than 60 nm under the same conditions produced almost same overall yield and few clusters as those of 60 nm particles. According to a previous report, nanoparticles are mobile size-dependently at the toluene–water interface [25
], thus the ligand-exchange in our case must have occurred at multiple and random sites on the Au nanoparticle, and were time-dependently controlled by the two-phase system. The Au nanoparticle sample (60 nm) that is prepared using sonication for 10 min is more beneficial to separate dimers since the sample is mainly composed of monomers and dimers. Although our method can control the dimerization of larger particles better than that of smaller-sized nanoparticles, the size of particles fits well with the dimerization of the larger nanoparticles (typically 40–100 nm) which have higher local field enhancement at optimum surface plasmon resonance [26
TEM images of assembled Au nanoparticle dimers (marked on the left, separation of the dimers: less than 2 nm, and highlighted on the right) through ligand-exchange for 10 min and unreacted Au nanoparticle monomers.
Control experiments were conducted in order to demonstrate the advantage of the SAM layer and of two-phase mediated ligand-exchange. The first control experiment was that both citrate-stabilized Au nanoparticles ligand-exchanged with 20 mM tiopronin in single aqueous phase and the citrate-stabilized (unfunctionalized) Au nanoparticles proceeded to dimerization through DIC-mediated amide bond formation. Both of the Au nanoparticles produced a great deal of clusters due to the existence of carboxylic groups which stemmed from either citrate molecules or tiopronin molecules on the nanoparticle surface ((a) and (b) in ). This result indicates that the concentration of tiopronin cannot be dialed in the single aqueous phase, and that our alkyl- and hydroxyl-terminated SAM layer can be advantageous in controlling the number of carboxylic reaction sites in the ligand-exchange step.
Figure 2 SEM images of control experiments, (a) citrate-stabilized Au nanoparticles, (b) citrate-stabilized Au nanoparticles ligand-exchanged with tiopronin in single aqueous phase, and (c) SAM layered Au nanoparticles ligand-exchanged with tiopronin in single (more ...)
The second control experiment was performed in order to confirm the significance of the Au nanoparticle dimerization through a two-phase mediated ligand-exchange, the whole surfaces of the SAM layered Au nanoparticles as well as the citrate-stabilized Au nanoparticles (the same as above) were ligand-exchanged with tiopronin in water. The amide bond formation of both control samples resulted in heavy aggregations of nanoparticles ((b) and (c) in ). More importantly, this control experiment demonstrates that the ligand-exchange of the Au nanoparticles obtained by our toluene–water interface methodology is critical to produce Au nanoparticle dimers with a high yield, and that whole ligand-exchange leads to the aggregation of Au nanoparticles, whether or not a SAM layer on the Au nanoparticles exists.
The formation of the Au nanoparticle dimers can also be verified by EM field enhancement. The EM field enhancement has been supported by either resonance scattering for nanosensing [27
] or absorption for therapeutics [28
]. It was reported that interacting plasmons promoted the absorption mechanism of island structure composed of the ultrathin Au thin film (≤10 nm) [29
]. Additionally, it was expected that larger nanoparticles would be suitable for light scattering, while intermediate-sized nanoparticles would be good for photoabsorption [30
]. Therefore, in our study, the scattering of both a single monomer and a single dimer of the Au nanoparticle was measured in air at far field using dark field spectroscopy with reflected illumination. Even though particle to particle spectral variation due to the difference in the nanoparticle's morphology and gap size was observed, the general difference between the single monomer and single dimer was apparent, as follows. In the spectra taken from the single Au nanoparticle monomers, one plasmon resonance peak at 540 nm is observed (). In regard to the single Au nanoparticle dimers, two resonance peaks are found. The longer wavelength resonance peak at 680 nm originates from the excitation of the hybridized surface plasmon modes that are parallel to the long axis of the dimer with a separation of less than 2 nm, and these always have a characteristic red shift from the monomer plasmon resonance. The shorter wavelength resonance peak of the dimers is owing to the excitation of the hybridized surface plasmon modes that are perpendicular to the long axis of the dimer [6
]. All of the single monomers and single dimers measured in our experiment were found to have much stronger dark field scattering intensity from a single dimer than that seen from a single monomer, at both 540 and 680 nm, and this is approximately correspondent to that of our numerical simulation result. shows the wavelength dependence for the intensity of the scattering from a monomer and a dimer. At both 540 and 680 nm, a much stronger scattering intensity from a single dimer than that from a single monomer was found.
Figure 3 (a) The dark field scattering of a single Au nanoparticle monomer and a single dimer (the isolated monomers and dimers were placed on an indium tin oxide (ITO)-coated quartz plate prepared by e-beam lithography) and the dark field spectroscopy setup. (more ...)