, 1:79

Preparation and Optical Characterization of Core–Shell Bimetal Nanoparticles

  • A. Steinbrück
  • A. Csáki
  • G. Festag
  • W. Fritzsche
Original Paper

DOI: 10.1007/s11468-005-9000-5

Cite this article as:
Steinbrück, A., Csáki, A., Festag, G. et al. Plasmonics (2006) 1: 79. doi:10.1007/s11468-005-9000-5


Chemical approaches allow for the synthesis of highly defined metal heteronanostructures, such as core–shell nanospheres. Because the material in the metal nanoparticles determines the plasmon resonance-induced absorption band, control of particle composition results in control of the position of the absorption band. Metal deposition on gold or silver nanoparticles yielded core–shell particles with modified optical properties. UV–vis spectroscopy on solution-grown, as well as surface-grown, particles was conducted and provided ensemble measurements in solution. Increasing the layers of a second metal leads to a shift in the absorption band. A shell diameter comparable to the original particle diameter leads to a predominant influence by the shell material. Extent of shell growth could be controlled by reaction time or the concentration of metal salt or reducing agent. Besides optical characterization, the utilization of atomic force microscopy, scanning electron microscopy, and transmission electron microscopy yielded important information about the ultrastructure of nanoparticle complexes. Surface-grown core–shell particles were superior in terms of achievable shell thickness, because of difficulties encountered with solution-grown particles due to salt-induced aggregation.


Nanoparticle Gold Silver AFM TEM UVvis spectroscopy 


For hundreds of years man has used nanoparticles to color glass, examples of which are still visible in colorful church windows. Michael Faraday [1] and Gustav Mie [2] are the most famous of past scientists who dealt with the synthesis and theoretical calculations of colloidal solutions. Today, many calculations are based on Mie's theory (1908).

The color of colloidal solutions is explained as the result of a collective oscillation of electrons of the conductive electron band. This effect generates a so-called surface plasmon (SP) band in the spectra where absorption reaches a maximum level at a certain wavelength of light. The location of the SP band can be influenced by the material, size, shape, and surrounding medium of a nanoparticle (solvent, ligands) as well as interparticle spacing. In bimetallic particles, the SP band depends further on the composition and the distribution of the two metals [7].

The chemical synthesis of nanoparticles has improved since the days of Faraday. Today, the principles of the synthesis process of nanoparticles are well established. Nanoparticles can be produced from several metals, such as Au, Ag, Cu, Pt, Pd, Ru, and others, with control over the size and the shape of particles. A variety of shapes such as spheres [6,8, 9, 10], rods with different aspect ratios [11, 12, 13], prisms [14,15], and cubes with varying edge lengths [16] and even more complex-shaped nanoparticles [17, 18, 19] have been created. Also, physical methods such as electron beam lithography can be used to fabricate highly defined metal nanostructures for optical studies [3, 4, 5].

In recent years, bimetallic nanoparticles have gained increasing attention. Two types of bimetallic nanoparticles can be classified: (1) particles with a homogeneous distribution of two metals called alloys; and (2) particles with heterogeneous arrangement of two metals leading to the so-called core–shell nanoparticles. Core–shell nanoparticles can be synthesized by successive reduction of two metals. Nanoparticles created during the first reduction process are used as seeds for the second reduction [20, 21, 22]. Alloy nanoparticles are synthesized in solution by the simultaneous reduction of the two metals of interest [23,24]. Laser treatment of core–shell nanoparticles can also result in the formation of alloy nanoparticles [25]. For the characterization of nanoparticles, several methods are used. Transmission electron microscopy (TEM) and scanning electron microscopy (SEM), complemented by atomic force microscopy (AFM), can be used to visualize the size, shape, and ultrastructure of the particles; UV–vis spectroscopy is applied to analyze the optical properties.

In this article, the synthesis and characterization regarding the structural and optical properties of core–shell nanoparticles will be discussed. The metals of interest are gold and silver.

Materials and methods

Synthesis of nanoparticles

Solution-grown core–shell nanoparticles

Solutions of gold nanoparticles (15, 30, and 60 nm in size) and silver nanoparticles (40 and 60 nm in size) were purchased from British Biocell International (BBI, Cardiff, UK). Additionally, gold nanoparticles with mean diameters of 12 and 47 nm were prepared according to a modified protocol described by Turkevitch et al. [8]. Silver particles of ca. 40 nm mean diameter size were synthesized according to a protocol described by Hutter et al. [26].

All nanoparticle solutions were characterized by SEM and/or TEM before use.

Silver enhancement was performed using the silver enhancement kit of BBI in 1:20 dilution. Furthermore, the commercial silver enhancement kit R-Gent from Aurion (Seligenstadt, D) and a homemade silver enhancement solution according to a protocol described by Hacker [27] were used in parallel. For gold enhancement, aqueous solutions of H[AuCl4] and NH2OH (0,5 mM both) were used according to the protocol described of Weizmann et al. [28]. Negative controls without silver or gold salt, or without reducing agent, were also prepared.

Surface-grown core–shell nanoparticles

Asymmetric core–shell nanoparticles were synthesized on microscopic glass slides (Roth, Karlsruhe, Germany) or thermally oxidized silicon chips. Therefore, nanoparticles (solutions see above) were bound unspecifically to the surface. Silver enhancement was performed using the silver enhancement kit of BBI according to the customers' manual. Solutions of H[AuCl4] and NH2OH (see above) were used for gold enhancement. For removal of core–shell nanoparticles from the surface, ultrasonic treatment was applied.

A centrifuge (Herolab UniCen 15DR, Wiesloch, Germany) was used for adjustment of the particle concentration and for washing procedures. Particles smaller than 15 nm were centrifuged at 12,900 g and larger particles at 5700 g.

Characterization methods

Core–shell nanoparticles were characterized by UV–vis spectroscopy (Nanodrop, Wilmington, DE, USA). To document the color of the solutions, color images of droplets were taken. For determination of shape and ultrastructure of specimens, scanning electron microscopy (SEM) and transmission electron microscopy (TEM) measurements were performed using a Zeiss DSM 960 (Jena, Germany) and a Zeiss CEM 902A (Jena, Germany), respectively.

For surface-grown core–shell nanoparticles, atomic force microscopy (AFM, NanoScope III from DI, Santa Barbara, CA, USA), together with the software ImageJ (freeware), was used to determine the size of the generated particles based on brightness-encoded height measurements.

Results and discussion

Solution-grown core–shell nanoparticles: Au–Ag

For silver enhancement, three different kinds of enhancement solutions were tested. The first was a homemade enhancement solution based on silver acetate and hydroquinone [27]. Furthermore, we used two commercial enhancement kits. In principle, all enhancement solutions worked for the tested system. However, the homemade solution was light-sensitive, which resulted in autonucleation of pure silver particles over time as indicated by UV–vis spectroscopy and TEM measurements (data not shown). The commercial enhancement kits were light-insensitive for a reaction time in the lower minute range, but the BBI kit showed stronger reaction with the gold solution than the Aurion kit (data not shown). For these reasons, the BBI kit was used in all succeeding experiments.

Gold nanoparticles of 30 nm in size were enhanced using different concentrations of a silver salt solution. During the enhancement reaction, silver salt is reduced by a reducing agent and is deposited onto the gold nanoparticles in the solution. The progress of the reaction can be followed by the naked eye as the color of the solution changes from red to orange to yellow (see Figure 1b, bottom).
Figure 1

(a) UV–vis absorption spectra of 30-nm-sized gold nanoparticles enhanced with increasing concentrations of silver salt in solution (B–D). (b) Scheme of enhancement and photographic pictures of droplets containing Au core/Ag shell particles. TEM (c, d) and SEM pictures (e, f) of 30 nm gold nanoparticles enhanced with silver in solution.

Immediately after the reaction, UV–vis absorption spectra were measured for all nanoparticle solutions (Figure 1a). The spectra of pure gold and silver particles serve as a reference (A and E, respectively, in Figure 1a). Pure gold particles of 30 nm in size (core in this experiment) show a characteristic SP band around 520 nm. Pure 40-nm-sized silver particles yield a peak around 415 nm. For core–shell structures, absorption spectra with two absorption maxima are expected [7] and measured as shown in Figure 1a (B–D). Due to the two interacting metals at the interface, peak positions for silver and gold are slightly shifted. As expected, a blue shift for the original pure gold peak is observed (Figure 1a). The silver peak seamed to be blue-shifted for low silver fractions compared to the 40-nm silver reference. However, in the early reaction state, the shell is much thinner than 40 nm (E in Figure 1a). However, with increasing amount of silver, the shell thickness grows resulting in a red-shift of the SP maximum for the silver fraction (B–D in Figure 1a). For increasing amounts of silver salt in the solution, the silver/gold ratio in the spectra increases, indicating a growing silver shell.

We also studied the influence of particle size on enhancement rate via investigation of the silver enhancement in solution for gold nanoparticles of mean diameter sizes of 15 and 60 nm (BBI), and 12 and 47 nm (self-made), respectively (Figure 2). The concentration of all four solutions is adjusted to the stock solution of 30 nm gold nanoparticles (BBI; 2×1011 particles/ml) by centrifugation or dilution, respectively. Thus, the results can be directly compared. For larger nanoparticles (47 and 60 nm), only little changes in the absorption spectra were detected, probably because of the larger surface that has to be enhanced with a constant amount of silver and resulting in a thinner layer (Figure 2b). For smaller nanoparticles, a thicker shell of silver can be deposited. This is reflected in the spectra by an increasing silver/gold ratio (Figure 2a). In Figure 3, the shift of the corresponding gold peak of the absorption spectrum is plotted versus the increasing amount of silver added for all three commercial gold solutions, based on the shift of the SP band as reference for the shell thickness. As predicted, 60-nm particles had spectra indicating the thinnest shells, 15-nm particles had the thickest shells, and 30-nm particles had intermediate shells.
Figure 2

UV–vis absorption spectra of (a) 12-nm (D, E) or 15-nm-sized gold nanoparticles (B, C), and (b) 47-nm (C) or 60-nm-sized (B) gold nanoparticles enhanced with silver salt in solution.

Figure 3

Dependence of absorption maximum of gold nanoparticles on the concentration of silver salt (%) added to differently sized commercial gold colloid solutions.

To obtain an impression of the achievable shell thickness, TEM and SEM imaging were applied (Figure 1c–f). Aside from a few particles that did not indicate any sign of enhancement (not shown), particles enhanced into various shapes (Figure 1c and f), and occasionally aggregated particles (Figure 1e) were visualized.

In parallel, we performed control samples either lacking the reducing agent or the silver salt. As expected in both cases, only one plasmon band attributed to gold nanoparticles was detected and the maxima were not significantly shifted by 1–2 nm (data not shown). In summary, silver deposition on gold is only possible by adding both silver salt and reducing agent, based on the electrochemical series.

Solution-grown core–shell nanoparticles: Ag–Au

As a general observation, we noted that silver nanoparticles were more sensitive to changes in salt conditions than gold nanoparticles.

For gold enhancement, a protocol according to Weizmann et al. [28] was used to deposit gold onto silver nanoparticles. The principle of the reaction is the same as that for silver enhancement and is schematically described in Figure 4b (top). The color change is also visible to the naked eye and is documented for droplets of 40-nm self-made silver particles in Figure 4b (bottom).
Figure 4

(a) UV–vis absorption spectra of 40 nm silver nanoparticles enhanced with increasing concentration of gold salt in solution (B, D), controls without reducing agent (C, E), and negative control (F). (b) Scheme of enhancement and photographic pictures of droplets containing Ag core/Au shell particles. SEM (c) and TEM pictures (d) of 40 nm silver nanoparticles enhanced with gold in solution.

In Figure 4a, spectra of solution-grown Ag–Au core–shell nanoparticles with 40-nm-sized silver nanoparticles (BBI) are presented. No clear two-peak spectrum was detected for this system. We expect a very thin shell under the conditions with only weak influence on the spectrum. We speculate that a small shoulder for gold is hidden under the broadened spectrum. SEM and TEM images (Figure 4c and d) show that no continuous shell has been formed. The created shell consists of small particles of gold. Control experiments were performed without the addition of reducing agent in parallel. The UV–vis spectra showed the same characteristics as the normal samples with gold salt solution and reducing agent added (curves C and E in Figure 4a). This means that the enhancement reaction is possible without any additional reducing agents. This is understandable because gold is more noble than silver (electrochemical series). When only reducing agent (but no gold salt solution) is added to the silver nanoparticle solution, there is no significant change in the absorption spectrum compared to the spectrum of untreated silver (F and G in Figure 4a).

Surface-grown core–shell nanoparticles

In solution, the thickness of the coating of core–shell nanoparticles is limited because the nanoparticles react sensitively toward higher salt concentrations thereby resulting in aggregations. This problem is overcome for surface-grown core–shell nanoparticles. One can enhance surface bound nanoparticles 10 times the original size or even larger. Another reason to synthesize core–shell nanoparticles on surfaces is that one can create asymmetrically shaped particles as the core is overgrown (see scheme, Figure 5b). We call these particles “toffifee”-like particles because their geometry of a spherical core surrounded by a semispherical shell resembles the toffifee candy. The SP bands of such “toffifee”-like particles are expected to show other bands than spherically shaped core–shell particles because the optical properties depend on (beside the material, the size, and the surrounding medium) the shape.
Figure 5

(a) Average height of Au core/Ag shell particles on SiO2 after different enhancement times (stepwise from 2 to 20 min). (b) Scheme of enhancement. (c–e) AFM pictures of 30 nm gold particles (c), followed by silver enhancement for (d) 6 and (e) 15 min, respectively [all pictures 10×10 μm; height scale: (c) 30 nm and (d, e) 100 nm].

Gold or silver nanoparticles, respectively, were unspecifically bound onto planar glass or SiO2 substrates and imaged with AFM (Figure 5c). Coating with silver or gold enhancement solution was carried out stepwise to ensure that the reactants are not consumed during the reaction. After every step, a washing procedure was performed. Then AFM images were again recorded to measure the height of the enhanced particles at each step (Figure 5d and e). As shown in Figure 5a for 30-nm gold nanoparticles enhanced with silver, the growth follows a linear relationship with respect to enhancement time. A good signal-to-noise ratio is obtained within the chosen conditions, which means that enhancement is mainly directed to the nanoparticles, but not to the whole surface. Figure 6 shows the summary of surface-grown bimetallic nanoparticles consisting of 30- and 60-nm-sized silver-enhanced gold particles and of 40- and 60-nm-sized gold-enhanced silver particles, respectively. A linear height growth during enhancement is observed for all types of particles, but the enhancement rate varies. This is probably attributable to the different surface densities caused by the unspecific binding of the particles on the substrate. This was confirmed by AFM images (data not shown). Also, a certain size distribution for the bimetallic particles is observed after the enhancement (see error bars in Figure 5a). Depending on the used enhancement solution, there are even occasional particles that will not be enhanced in all images (not shown).
Figure 6

Measured height of surface-grown core–shell particles: (left) 30- and 60-nm-sized silver-enhanced gold particles; (right) 40- and 60-nm-sized gold-enhanced silver particles.

Following the enhancement steps and AFM imaging, the core–shell particles were removed from the surface by ultrasonication in order to study the optical properties by UV–vis spectroscopy in solution. Because of the resulting random orientation, any effects expected from the asymmetrical shape are averaged in solution. In Figure 7a, absorption spectra of bimetallic particles with 60 nm gold core are shown as an example. Theoretically, it is expected that two-peak spectra characteristic for core–shell particles will be detected again, but the signals for silver and gold peaks are not well defined. Compared to solution-grown core–shell particles, signal from the core should be damped because the thick shell (especially when it is silver) should give the primary spectral response. There are some possible explanations for the broadened spectra and the plateau over the entire region of interest between 400 and 500 nm. We do not know the fraction of unenhanced gold particles that might broaden the “gold” fraction of the spectrum. From the SEM and TEM pictures (Figure 7b–h), we observed various sizes and shapes of particles and also some aggregates formed during the enhancement process. In general, a broad size distribution results in broadened SP bands. Furthermore, some silver aggregates are built on the surface even though there is low background. In summary, all these deviations from ideal conditions can influence the sum spectrum. Therefore, it is necessary to reduce the background and the formation of aggregates. Moreover, higher-efficiency methods for the removal of the surface-grown core–shell nanoparticles are required.
Figure 7

(a) UV–vis spectra of surface-grown bimetallic particles with 60 nm gold core after removal from surface. (b) SEM image of 60-nm-sized gold nanoparticles enhanced with silver for 6 min. (c–h) TEM images of 30-nm-sized gold nanoparticles enhanced with silver for 3 min (c–e) or 6 min (f–h), respectively.


We have shown the synthesis of heterogenous bimetallic nanoparticles, the so-called core–shell nanoparticles, in solution and on surfaces. The metals of interest were gold and silver. We created gold core–silver shell particles and vice versa. The absorption spectra showed two-peak spectra as reported in the literature for core–shell systems. However, due to the sensitivity to changes in salt concentration, the shell thickness for solution-grown core–shell nanoparticles is limited. In contrast, it is possible to increase shell thickness for surface-grown core–shell particles 10 times the former diameter in solution.


We will try to stabilize our nanoparticles with ligands against changes in salt concentration in order to obtain thicker shells in solution, too. It is also interesting to find an easy method for size separation because a sharp size distribution will result in sharpened absorption peaks. Another goal is to modify bimetallic nanoparticles with DNA and to build constructs based on DNA hybridization. Approaches with fluorescence tags are also possible.

We will also continue to characterize the optical properties of single particles.


We would like to thank Andrea Assmann for SEM imaging, Katrin Buder (IMB Jena) for help with TEM measurements, Uwe Klenz for 3D rendering, and James Vesenka for correcting the manuscript.

Copyright information

© Springer Science+Business Media, Inc. 2006

Authors and Affiliations

  • A. Steinbrück
    • 1
  • A. Csáki
    • 1
  • G. Festag
    • 1
  • W. Fritzsche
    • 1
  1. 1.Photonic Chip Systems DepartmentInstitute for Physical High Technology JenaJenaGermany

Personalised recommendations