Plasmonics

, Volume 1, Issue 1, pp 45–51

Photochemical Synthesis and Multiphoton Luminescence of Monodisperse Silver Nanocrystals

Authors

  • Thomas Kempa
    • Eugene F. Merkert Chemistry CenterBoston College
  • Richard A. Farrer
    • Eugene F. Merkert Chemistry CenterBoston College
  • Michael Giersig
    • Department of Nanoparticle TechnologyCenter of Advanced European Studies and Research
    • Eugene F. Merkert Chemistry CenterBoston College
    • Department of Chemistry and BiochemistryUniversity of Maryland
Original Paper

DOI: 10.1007/s11468-006-9008-5

Cite this article as:
Kempa, T., Farrer, R.A., Giersig, M. et al. Plasmonics (2006) 1: 45. doi:10.1007/s11468-006-9008-5

Abstract

A rapid, photochemical solution-phase synthesis has been developed for the production of monodisperse, nanometer-sized silver particles. The stabilizer used in the synthesis can be used to control the average diameter of the particles over a range from 1 to 7 nm. The same reaction mixture can also be employed to deposit patterns of nanoparticles with a laser via multiphoton absorption. The particles exhibit strong multiphoton absorption-induced luminescence when irradiated with 800-nm light, allowing emission from single nanoparticles to be observed readily.

Key words

Silver nanoparticlePhotochemical synthesisMultiphoton absorptionLuminescence

Applications of noble-metal nanoparticles have grown exponentially over the past decade. Silver is one of the least expensive of the noble metals and is only weakly reactive in the bulk. In nanoparticle form, silver has properties that can be exploited for applications such as surface-enhanced Raman spectroscopy (SERS) [1], even down to the single-molecule level [2,3]. In addition, subnanometer silver clusters show considerable promise as luminescent media for single-molecule studies [4,5]. Although the preparation of monodisperse nanoclusters of metals such as gold and platinum with average diameters that can be less than 1 nm is well documented [6,7], the synthesis of monodisperse silver nanoclusters is generally a considerably more difficult prospect, and it has been difficult to attain diameters of less than 3 nm in solution [8,9]. Here we report a facile solution-phase photochemical synthesis that, in a matter of minutes, produces monodisperse silver nanoclusters with controllable average diameters that can range from 1 to 7 nm. We further show that two-photon absorption (TPA) can be used to pattern the nanoparticles on a substrate. In addition, the particles exhibit strong multiphoton-absorption-induced luminescence (MAIL) upon irradiation with ultrafast pulses of 800-nm light, allowing the emission of individual particles to be monitored.

The role of kinetics in determining the average size and dispersity of semiconductor nanoparticles has long been recognized and used to advantage [10]. Many such schemes require the rapid mixing of precursors to initiate the synthesis of nanoparticles. However, this type of strategy has received considerably less attention in the synthesis of noble-metal nanoparticles. A key element in such an approach is the ability to reduce silver cations quickly, such that nanoparticles with small diameters can be trapped kinetically. Photochemical reduction is an attractive means of accomplishing this end, as it allows for complete mixing of the reagents before the synthesis is initiated. Whereas a number of groups have reported photochemical syntheses of silver nanoparticles [1126], in all of this previous work, the resultant nanoparticles have been polydisperse and often relatively large as well. The polydispersity of the particles produced is probably related to the fact that these syntheses involve ultraviolet exposure times of many minutes.

We reasoned that rapid photoreduction would facilitate the kinetic trapping of small, monodisperse nanoparticles. To this end, we chose the efficient photobase N-methylnifedipine [27] as a reducing agent. The structure of N-methylnifedipine and its absorption spectrum in tetrahydrofuran (THF) are shown in Figure 1. Photoexcitation of N-methylnifedipine produces a strong reducing agent that allows for the rapid conversion of silver ions to silver metal.
https://static-content.springer.com/image/art%3A10.1007%2Fs11468-006-9008-5/MediaObjects/11468_2006_9008_Fig1_HTML.gif
Figure 1

Absorption spectrum and activation mechanism of N-methylnifedipine.

To create nanoparticles, 10.0 mL of a 1.00-mM solution of AgClO4 in THF was mixed with 10.0 mL of a 0.75-mM solution of N-methylnifedipine in THF and placed in a test tube that was highly transmissive in the near ultraviolet. The mixture was then exposed to an ultraviolet lamp for 1 s. The particles were extracted into water to remove any remaining photobase. For transmission electron microscopy (TEM) studies, it was desirable also to remove any remaining silver salt and unbound stabilizer molecules, so the stabilizers were exchanged with oleic acid and the particles were reextracted into toluene. A drop of solution was then placed on a TEM grid and allowed to evaporate. TEM of the resulting sample (Figure 2) revealed polydisperse silver nanoparticles. Whereas there was a significant population of particles with diameters on the order of 1 nm, some particles had diameters greater than 20 nm.
https://static-content.springer.com/image/art%3A10.1007%2Fs11468-006-9008-5/MediaObjects/11468_2006_9008_Fig2_HTML.jpg
Figure 2

TEM of silver nanoparticles prepared without a stabilizer. The scale bar is 30 m long.

To prevent the formation of the larger nanoparticles, the reaction mixture was modified to include 10 mL of a 1.0 mM sodium citrate solution in THF as a stabilizer for the particles. A TEM of the nanoparticles that resulted after the mixture was exposed to an ultraviolet lamp for 1 s is shown in Figure 3a. The addition of citrate resulted in the production of monodisperse silver nanoparticles with an average diameter of 1.3 nm, a diameter polydispersity (Pd) of 1.10, and a mass polydispersity (Pm) of 1.51. Exposure times of up to 1000 s were tested, and in all cases, the results were similar to those for an exposure time of 1 s, indicating that the reaction has reached completion by this time, and that the nanoparticles remain kinetically stable once they have been formed. The rapid formation of nanoclusters implies that N-methylnifedipine is a more efficient photoreducing agent than those employed in previous studies. We note that the UV-induced fragmentation of silver nanoparticles into nanoclusters in the size range here has been reported previously [28]. This fragmentation process requires exposure times that are orders of magnitude longer than the ones used here and is accompanied by the formation of larger nanoprisms. We can therefore rule out fragmentation as the source of the nanoclusters that we observe.
https://static-content.springer.com/image/art%3A10.1007%2Fs11468-006-9008-5/MediaObjects/11468_2006_9008_Fig3_HTML.gif
Figure 3

TEM images of silver nanoparticles created with the sodium salts of (a) citric, (b) tartaric, (c) aspartic, (d) glutamic, (e) malonic, and (f) oxalic acids. The insets show size-distribution histograms for each image. All scale bars are 30 nm long.

Having determined that the basic reaction scheme presented above could produce monodisperse nanoparticles, different polyfunctional molecules containing carboxylic-acid and primary-amine moieties were tested as surface stabilizers to control the average diameter of the particles. In each case, a 1.0-mM solution of the stabilizer in THF was used in place of the sodium citrate solution in the above synthesis, and in each case, monodisperse silver particles were obtained, as seen from the TEMs in Figure 3. The compounds that were investigated as potential stabilizers include the sodium salts of tartaric (Figure 3b), aspartic (Figure 3c), glutamic (Figure 3d), malonic (Figure 3e), and oxalic (Figure 3f) acids. The data in Table I illustrate that the average nanoparticle diameter can indeed be controlled via the choice of stabilizer. The general trend is that smaller stabilizers produce larger nanoparticles, presumably because the diameters are determined predominantly by the minimum radius of curvature of a particle that the stabilizer can cover effectively. Thus, employing larger stabilizer molecules may allow for the production of even smaller nanoclusters. Because aspartate and glutamate contain amine groups, which have a higher affinity for metal clusters than do carboxylic acids, these anions are chemically distinct from the other stabilizers employed here. This enhanced affinity may be responsible for the particularly favorable dispersity noted in particles prepared with these stabilizers.
Table I

Average diameter, diameter polydispersity, and mass polydispersity of nanoparticles with different stabilizers.

Stabilizer

d〉 (nm)

Pd

Pm

Citrate

1.3

1.10

1.51

Tartrate

1.4

1.08

1.54

Aspartate

2.4

1.02

1.23

Glutamate

3.2

1.02

1.23

Malonate

3.3

1.03

1.21

Oxalate

6.4

1.01

1.06

Shown in Figure 4 are high-resolution TEMs of two representative nanoparticles stabilized with aspartate. It is clear from these images that the nanoparticles are predominantly crystalline, although some defects can be seen. The particles are also not perfectly spherical. Both of these observations are in line with what might be expected from a kinetically controlled growth process. Indeed, citrate-capped silver nanocrystals in this size range tend to assume a cuboctahedral geometry that makes it difficult to observe clean crystallographic lines [29], as is the case here. The lattice spacing measured from the electron diffraction images in Figure 4 is 2.3 Å, which is consistent with the spacing of the (111) planes of cubic bulk silver [30].
https://static-content.springer.com/image/art%3A10.1007%2Fs11468-006-9008-5/MediaObjects/11468_2006_9008_Fig4_HTML.jpg
Figure 4

Images (a) and (b) show high-resolution TEMs of representative nanoparticles stabilized with aspartate; the micrographs are approximately 5 nm on a side. Images (c) and (d) show the corresponding electron-diffraction patterns.

The photochemical nature of the synthesis can be used to advantage to pattern silver particles on a microscopic scale. Rather than excite the photobase resonantly with single-photon excitation, we employed two-photon excitation using the 800-nm output of a Ti:sapphire laser. Because there is no transition in the photobase that is resonant with this light, two photons must be absorbed simultaneously to cause electronic excitation, and the absorption probability therefore scales as the square of the laser intensity. As a result, TPA can be localized to within the tight focal volume that results from sending the laser beam through a microscope objective [31], which, in turn, should localize the production of nanoparticles.

Our TPA apparatus has been described in detail elsewhere [32]. Briefly, a commercial Ti:sapphire laser (Coherent Mira Basic) was used to produce pulses with a center wavelength of 800 nm and a repetition rate of 76 MHz. After dispersion compensation, the pulse length was ∼100 fs. The laser output was introduced into the reflected-light port of an upright microscope (Zeiss AxioPlan 2) and focused using a 20×, 0.5-NA, nonimmersion objective. The laser spot size was large enough to overfill the back aperture of the objective. The position of the sample was controlled in three dimensions with a motorized stage (Ludl BioPrecision) and the focusing drive of the microscope. Transmitted light was used to view the sample on a charge-coupled device camera during the deposition process.

A solution was prepared consisting of 10-mL portions of 5.00 mM AgClO4, 3.75 mM N-methylnifedipine, and 5.00 mM sodium malonate dissolved in THF. The sample cell was constructed from two clean glass microscope slides with a 1-mm rubber O-ring serving as a spacer and seal. A 300-mesh copper TEM grid with a thin coating (10 nm) of amorphous carbon was secured to the surface of one of the microscope slides such that it was in the center of the O-ring. A few drops of the solution were placed inside the O-ring, and the second microscope slide was brought in contact with it to seal the liquid sample above the TEM grid. Teflon tape was used to secure the sample tightly. The microscope was focused on the carbon-film surface of the TEM grid, and a shutter was opened to allow the laser output to reach the sample. The stage was then scanned to produce two-dimensional patterns. An average power of 3 mW at the sample was sufficient for the observation of material deposition.

Shown in Figure 5a is a low-resolution TEM image of a typical two-dimensional pattern deposited on a TEM grid. The deposited lines are on the order of 3 μm wide. Zooming in on one of the dark regions reveals the presence of monodisperse silver nanoparticles, as can be seen in the high-resolution TEM image in Figure 5b. The diameters of the particles are comparable to the 3.3 nm observed for UV exposure in the bulk solution with malonate as the stabilizer. Although the use of TPA to pattern silver has been reported, the goal of these previous studies was to form continuous silver patterns [33,34]. This is the first demonstration to our knowledge of the use of TPA to create patterns of isolated nanoparticles.
https://static-content.springer.com/image/art%3A10.1007%2Fs11468-006-9008-5/MediaObjects/11468_2006_9008_Fig5_HTML.jpg
Figure 5

Low-resolution (a) and high-resolution (b) images of silver nanoparticles deposited on a TEM grid using TPA. The scale bar in (a) is 10 μm long, and the scale bar in (b) is 10 nm long.

Noble metals are known to luminesce upon multiphoton excitation in the infrared [3539]. Although this MAIL is generally weak, we have recently demonstrated that gold nanoparticles can exhibit highly efficient visible MAIL when excited with ultrafast pulses in the near infrared [40]. The high efficiency of MAIL is believed to be associated with large multiphoton absorption cross sections arising from the strong electric field enhancements generated at nanoparticles that have asperities [40].

The somewhat irregular shapes of our nanocrystals suggested that they might also be able to generate MAIL efficiently. To determine whether this was the case, nanoparticles that were synthesized with citrate as the stabilizer were transferred to toluene as described above, and a dilute solution was evaporated on a coverslip. A two-photon microscope with the sensitivity to detect fluorescence from single molecules [41] was used to observe the samples using an excitation wavelength of 800 nm. Two typical MAIL images obtained at orthogonal emission polarizations are shown in Figure 6. The laser power used in these images was approximately 10 mW at the sample, and the brightest particles gave more than 200 counts/ms.
https://static-content.springer.com/image/art%3A10.1007%2Fs11468-006-9008-5/MediaObjects/11468_2006_9008_Fig6_HTML.jpg
Figure 6

False-color MAIL images of silver nanoparticles on a glass coverslip at an excitation power of 10 mW. White corresponds to >100 counts in 1 ms. The images are 15 μm on a side, and panels (a) and (b) were obtained at orthogonal polarizations.

The same particles exhibit different emission intensities in Figure 6a and b, indicating that the emission is polarized, as was observed for gold nanoparticles. The silver nanocrystals are also highly photostable; they neither blink nor photobleach. The MAIL intensity of a particle does fluctuate to some extent over time, which may result from orientational dynamics or may be akin to the spectral fluctuations observed previously in silver nanoclusters [4]. As was the case for gold nanoparticles, the MAIL spectrum spans much of the visible spectrum, although the emission from the two metals is distinguishable.

As shown in Figure 7, a logarithmic plot of the MAIL intensity as a function of excitation power reveals that excitation occurs through TPA. In contrast, we found previously that high-efficiency MAIL is a three-photon process in gold nanoparticles at the same wavelength [40]. It is somewhat surprising that MAIL should require fewer photons for silver than for gold. We believe that in the case of gold nanoparticles, the most efficient emission arose from particles with especially sharp features. Such features make it possible to generate large-enough field enhancements to promote three-photon absorption, which occurs at an energy at which the density of states is considerably greater than that at the energy for TPA. In contrast, for larger gold structures, MAIL is a two-photon process at 800 nm [38,39]. In the case of silver, MAIL is considerably less efficient for the silver particles studied here than it was for the most efficient gold particles studied [40], suggesting that the field enhancement in the silver particles is not as great. This is consistent with the TEM images of the silver particles, which are somewhat irregular but do not exhibit large asperities.
https://static-content.springer.com/image/art%3A10.1007%2Fs11468-006-9008-5/MediaObjects/11468_2006_9008_Fig7_HTML.gif
Figure 7

Power dependence of MAIL for two different silver nanoparticles.

In summary, we have demonstrated a simple photochemical synthesis of monodisperse silver nanoparticles with controllable average diameters in the range of a few nanometers. Nanoparticles not only can be created in the bulk using single-photon absorption, but also can be deposited in a controlled fashion on surfaces using TPA of 800-nm light. The ability to pattern these particles with high resolution may have a number of applications, including the fabrication of SERS devices. The nanoparticles also exhibit efficient MAIL upon absorption of two photons of 800-nm light and are highly photostable. As a result, these particles should prove to be excellent photolabels for observation of single-molecule dynamics.

Acknowledgements

This work was supported by the National Science Foundation, Grant ECS-0088438. J.T.F. is a Research Corporation Cottrell Scholar and a Camille Dreyfus Teacher-Scholar. T.K. is a Beckman Scholar and thanks the Hahn-Meitner Institute for a Summer Student Fellowship. We thank Michael Hilgendorff for assistance with some of the TEM measurements reported here.

Copyright information

© Springer Science+Business Media, Inc. 2006