This contribution summarizes some of the salient aspects of lectures presented by the authors at the conference of the Inter-American Photochemical Society held in Mendoza, Argentina in May 2011.
Colloidal solutions of noble metal nanoparticles have colors that are totally different from their characteristic bulk appearance. Thus, gold is usually red, whereas solutions containing silver nanoparticles (AgNP) are frequently yellow. Their colorful appearance is due to a characteristic transition due to a surface plasmon band (SPB), which in the case of spherical AgNP occurs at around 400 nm and is responsible for their yellow color (1,2). The SPB is due to the collective motion of electrons upon interaction of the metal nanoparticle with the incident light beam; the position and shape of this band (and thus the color of the material) depends on many parameters, including chemical composition, size, morphology and surface coverage (3–5). Long before “nano” gained its current popularity, or before the origin and nature of SPB transitions was recognized, metal nanoparticles were employed as pigments in materials, such as decorative stained glasses, abundant in many churches around the world (2). Perhaps a message contained in this particular application is that under the right conditions (e.g. in stained glasses) metal nanoparticles are robust materials with remarkable longevity and stable to ambient conditions of light and thermal exposure. As a way of introduction, Fig. 1 shows representative spectra of spherical gold, silver and copper nanoparticles in aqueous solutions.
Silver atoms in solution: on the way to AgNP
The first step in the synthesis of metal nanoparticles in solution is the reduction of metal ions by a suitable reducing agent. Among the nonphotochemical ways of reducing metal ions, agents such as BH4−, citrate, ascorbic acid and hydroxylamine have been common choices (6–8). Photochemical approaches to noble metal ion reduction frequently take advantage of the electron donating properties of some free radicals that can in turn be generated photochemically. Among these, ketyl radicals and α-aminoalkyl radicals have proven very versatile; both can be produced by the Norrish Type I photocleavage reactions (9–11). In this contribution, we illustrate the reaction with ketyl radicals that have proven convenient for silver and that can be readily generated from common benzoins. Scheme 1 illustrates the reduction of Ag+ by the radicals formed by photodecomposition of Irgacure-2959™ (I-2959), a benzoin available commercially (12,13).
While numerous benzoins are available, I-2959 has proven extremely valuable. In particular, its convenient solubility in solvents as diverse as water, toluene and THF (at millimolar concentrations) make I-2959 remarkably versatile. Its photodecomposition occurs with a quantum yield of 0.29 and triplet lifetime of only 11 ns (13). The latter makes triplet quenching by metal ions an inefficient process, thus contributing to the overall high efficiency of metal ion reduction. That is, triplet quenching is undesirable because triplets that are quenched do not generate the free radicals needed as reducing agents (12).
Note that Scheme 1 only follows the fate of the ketyl radicals and is not concerned with the fate of the substituted benzoyl radical formed in the first reaction. For the purposes of this contribution, suffice to say that its ultimate fate is the formation of the corresponding carboxylic acid (HEBA) that, through mild binding, contributes to the stability of metal nanoparticles.
The first stage on the way to forming AgNP is the formation of Ag2. We have studied the dimerization of Ag atoms in nonpolar solutions and shown that the rate constant for this process is near diffusion controlled. Our studies also suggest that initial growth occurs by “even numbers,” that is Ag4 is reached by 2Ag2 → Ag4, as opposed to processes that could be mediated by Ag3. This is inferred from the absolute selectivity to rapidly form Ag2 and not Ag3, or larger clusters at very short time scales (14). The results were complemented by two-laser experiments that generated Ag0 in the presence of Ag2.
Beyond these initial stages, the growth of these initial clusters to form AgNP with thousands of atoms is an entirely spontaneous process; this is not to say that the process cannot be controlled, in fact, later in this article, we show that size and morphology can be readily tuned to produce nanostructures with desirable properties (3,15).
One particular aspect of AgNP growth in nonpolar media is worth discussing. A convenient precursor for Ag0 in toluene or THF is AgCF3COO, given its solubility in nonpolar solvents. Reduction by the mechanism of Scheme 1 is sluggish, but can be readily assisted by addition of equimolar amounts of amines, such as cyclohexylamine (10). Its role is probably to act as a receiver for the proton formed in Scheme 1. Much to our initial surprise, the reaction leads to strongly fluorescent solutions. Detailed studies of this system have shown that the reduction of AgCF3COO in nonpolar media leads to AgNP decorated by Ag2, or small Ag clusters and that these species are responsible for the strong fluorescence (10).
In aqueous media, reduction of Ag+ (usually from AgNO3) leads to nonfluorescent spherical AgNP with spectra comparable to that included in Fig. 1. If small nanoparticles are preferred, their growth can be arrested by adding suitable stabilizing (i.e. surface covering) agents. We find that citrate is a convenient one for AgNP, and in its presence the particles obtained are around 3 nm in diameter and stable indefinitely (at least 6 months) in solution (3).
Controlling the size and shape of AgNP
When we engaged in this research, a number of strategies to control size and morphology of AgNP where already known; however, each one of these approaches led to particles that frequently differed in other ways, such as the presence of different stabilizing agents, or the fact that their diverse history could lead to “chemical debris” with different properties reflecting variations in the synthetic procedure (16–25).
In our laboratory, we were able to devise a strategy that relies on a single synthetic methodology and a single stabilizing agent, always at the same concentration, yet achieve a remarkable diversity of morphologies and of spectroscopic properties. In our approach, AgNP seeds of ca 3.3 nm in diameter are exposed to light from narrow band LEDs (3). This is illustrated in Fig. 2 for the conversion of small spherical AgNP into decahedra upon irradiation with light centered at 455 nm.
The broad diversity of morphologies of AgNP is achieved by simply changing the LED source to excite with different wavelength maxima. For example, triangular AgNP plates can be made with LED excitation with 590 or 627 nm light, where higher aspect ratio plates are made with longer wavelengths. Overall we were able to make larger spheres, decahedra, hexagonal and triangular plates, and nanorods with 405, 455, 505, 590/627 and 720 nm emitting LED, respectively. One of the most interesting features of this morphology control is the associated change in visible spectra of each of these species, and thus the color, as shown in Fig. 3.
It is often thought that capping ligands are the controlling factor for realizing different shapes of AgNP even in light driven shape conversions. However, we have shown that with one capping ligand, citrate, we are able to obtain a multitude of shapes. There is an induction period to the shape control, which has been attributed to the formation of anisotropic particles in a reversible surface oxidation/citrate reduction process, whereby reduced ions do not deposit homogeneously on the surface due to long wavelength excitation. The particles then aggregate and coalesce into various shapes as controlled by the wavelength of irradiation. For short wavelength shape conversion, as is the case with decahedral particles with 455 nm LED, the conversion is fast (<2 h for completion), likely due to a significant absorption of LED irradiation by the original seeds. However, for obtaining shapes such as rods using 720 nm, irradiation times are often around 1 week using comparable light intensity; we attribute the need for very long exposures to the minimal excitation of AgNP seeds by such long wavelength irradiation. However, to make shapes like triangular plates using LED with intense and relatively monochromatic light has drastically reduced the time required as compared with Xe lamp irradiation.
Photochemical applications of AgNP
Our ability to make nanostructures with a relatively unprotected surface (i.e. no covalent bonds such as Ag-S or Ag-P) and with controlled morphology would be of little interest if we could not take advantage of these materials in novel areas of application. In this contribution, we concentrate on uses that are triggered by the absorption of light by the nanostructure, frequently by excitation of the SPB. Many other applications can be conceived in the areas of catalysis (26) and nanomedicine, uses that in many cases do not involve, or require photoexcitation. Others in our group at the University of Ottawa are exploring these avenues.
Applications of silver-based nanostructures that require photoexcitation can be classified according to the phenomenon involved, or by the actual application. Some possibilities—in no way exhaustive—are listed in Table 1.
|Plasmon-mediated phenomena||Applications of SPB excitation|
|Efficient light to heat conversion||Surface heating and thermal chemistry|
|Field enhancement effects||Enhanced vibrational spectroscopy|
|Hole or electron transfer||Enhanced electronic transitions|
Light-to-heat conversion can be viewed as a rather trivial effect that simply reflects the laws of energy conservation; however, the ultrafast (subpicosecond) relaxation that follows SPB excitation causes major time dependent temperature gradients that can trigger reactions that normally require high temperature at formally near-ambient temperatures. These processes have been mainly examined for AuNP, where a study from our group revealed that molecules near the surface experience temperatures as high as 500°C for submicrosecond times following SPB excitation, which in the case of AuNP occurs at around 530 nm (27).
The nature of field enhancement effects have been elegantly demonstrated by Novotny in a classic experiment in which the fluorescence of a single molecule was recorded as a single AuNP was brought in its close proximity (28). As molecules come in close proximity to metal nanoparticles the electromagnetic (EM) field around the particles increases the efficiency of electronic and vibrational transitions reaching a maximum at approximately 5 nm from the surface of the particles. At least in the case of electronic transitions, when molecules are extremely close to the surface of metal nanoparticles, enhancement factors become negligible, and electronic transitions are actually reduced in an efficient quenching process (Fig. 4).
One of the best-established applications of plasmon field enhancements is in surface enhanced Raman spectroscopy (SERS), where nanoparticles have proven valuable tools for signal enhancement (15,29–32).
In our laboratory, we have studied the effect of the size of spherical AgNP on the intensity of SERS spectra of molecules bound to the AgNP surface, specifically rhodamine-6G (R6G). When metal particles are excited, there is a resultant EM field induced around the particles and the size/strength of this field is proportional to the size of the particles (larger particles have stronger EM fields). However, smaller particles have a higher surface area to volume ratio and for a given amount of total silver, larger particles have less total surface area available for binding molecules. Therefore, for a constant amount of silver there should be an optimal size of AgNP that allows maximum surface coverage, whereas still providing significantly enhanced Raman spectra. We have determined that, for R6G, the optimal size of AgNP for SERS is approximately 50 nm under surface saturated conditions. When the enhancement factor (G-value) per R6G molecule in a monolayer on the surface of AgNP is calculated (larger particles have less surface area for binding) and is plotted against AgNP the relationship in Fig. 5c is observed where enhancement factor increases for larger AgNP. It is expected that these results can be extended to other surface bound molecules as well, as the EM field is not significantly influenced by the medium (15).
Using SPB excitation to achieve subwavelength resolution may find applications in very diverse areas, from new functional materials to information technologies and in the field of nanomedicine. In a recent contribution, we concentrate on an example where SPB excitation is used to create acrylic features approximately 1/20 of the wavelength of the excitation light. We note that simple microscopy techniques can normally resolve features that are about one-third of the wavelength of light, whereas state of the art lithography is approaching 1/9 of the excitation wavelength (33). Thus, we note that 1/20 would be a major resolution improvement.
For resolution of subwavelength features, we have taken advantage of the field enhancement phenomenon of SPB excitation to selectively excite electronic transitions in azo-bis-isobutyronitrile (AIBN) that is in close proximity (<10 nm) of the AgNP surface, and in doing so induce radical generation followed by polymerization. Selectively decomposed AIBN in the vicinity of AgNP (due to SPB excitation) releases N2 and produces two radicals, which are capable of cross-linking trimethylpropane triacrylate. One of the keys to this study is to choose a wavelength where the AgNP absorb, and therefore have a significant induced EM field, but where AIBN has a minimal absorption in the bulk matrix (34). Figure 6a shows that 405 nm light (delivered by an intense LED) fulfills these requirements. The scheme in Fig. 6b describes the effect of selective polymerization due to the antenna effect near the particle surface.
The polymerization process was monitored by both scanning electron microscopy (SEM) and atomic force microscopy (AFM), and in both cases approximately 8–10 nm features were confirmed. Figure 7 gives a representative AFM image of particles after polymerization. An interesting feature of the process is that nanoparticle dimers and larger aggregates were preferentially retained on the films, a feature that is consistent with an EM enhancement as the mechanism for spatially favored photoinitiation.
Further experiments are currently underway in exploiting the phenomena surrounding plasmon excitation of noble metal nanoparticles, and in particular, AgNP. We are currently investigating the novel methods to selectively generate positive lithographic features to provide subwavelength resolution positive tone images. We are also investigating the safe levels of light intensity and concentration of nanoparticles for the use of AgNP in medical applications, above which denaturing of proteins and cell death would be imminent. Yet, in other studies, the Scaiano group is looking at novel approaches to making surface plasmon lasers that have potential applications in optoelectronics, where coherent plasmon fields are desirable. A general goal of this research is to strive toward interesting and novel ways of exploiting SPB phenomena to ultimately benefit many different areas of research.