The influence of complexing agent concentration on particle size in the process of SERS active silver colloid synthesis

Abstract

A one-step chemical reduction route towards silver colloid particles with controllable sizes ranging from 45 to 380 nm is reported in this article. Silver particles, prepared by the reduction of [Ag(NH₃)₂]⁺ complex with various reducing sugars, were characterised by means of transmission electron microscopy (TEM), dynamic light scattering (DLS) measurement of particle size distribution, and UV-VIS spectroscopy. The concentration of ammonia in the reaction mixture and the choice of the reducing sugar are the key parameters in the control of particle size. Synthesised silver colloid particles were successfully tested for use in surface-enhanced Raman spectroscopy (SERS) with 1-methyladenine as a testing substance.

---

Introduction

Nowadays, the study of nanometer-sized particles is an emerging branch of colloid chemistry. These particles possess a specific quality—their properties are strongly dependent on their dimensions. This size quantization effect is mainly due to the arrangement of energetic levels (differing from that of the bulk phase) and also the high fraction of surfacial atoms.¹ Colloid particles have become an important building element in the intensively developing area of nanotechnology, where, for example, they are used as substrate specific sensors for biological and medicinal applications.² Silver, thanks to its physicochemical properties and the easy preparation of its colloid particles, is one of the most widely used metals in modern nanotechnology. The wide application of silver colloid particles began with the discovery of surface-enhanced Raman scattering in the early 1970s.³ Today surface-enhanced Raman spectroscopy is not only a research method but is also widely used in practical analytical chemistry,^4,5 because of its extremely high sensitivity, which makes possible the detection of even a single molecule adsorbed on the silver surface.^6–8

The applicability of silver colloid particles or their hydrosols to certain applications is dependent on their dimensions and morphology. For example, the maximum signal amplification in SER spectroscopy (10¹⁵) was achieved with silver particles between 110 to 120 nm for the 514.5 nm laser wavelength.⁹ For laser wavelengths between 647 and 488 nm, optimum particle sizes range from 70 up to 200 nm.¹⁰ Other properties are size-dependent, for instance the catalytic efficiency for the reduction of certain organic dyes,¹¹ or the non-linear optical properties of these particles.¹²

The controlled preparation of silver colloid particles ranging in size from a few nanometers to hundreds of nanometers, according to the requirements of a particular application, is therefore a major research task in this area. Methods based on the chemical reduction of silver salt solution currently prevail. The primary approach is based on a two-step reduction synthesis. In this process, tiny silver particles formed by the primary reduction of the silver salt using a strong reducing agent are, in the second step, enlarged by further reduction using a weaker reduction agent. In this case, the 20 nm-sized silver “seeding” particles, prepared with the use of NaBH₄, are enlarged up to 170 nm by subsequent reduction with ascorbic acid.^13,14 If a weaker reducing agent is used in the first step instead of a strong one, then the resulting particles after second enlargement step reveal a relatively broad size distribution. Thus, the silver particles with an average size of about 45 nm, prepared by the reduction of silver salt with citrate according to the procedure of Lee and Meisel,¹⁵ are enlarged up to an average size of 120 nm; however, the resulting particles are considerably polydisperse.¹⁶ A novel two-step route towards silver colloid particles of the controlled size was suggested by Emory et al., who separated highly polydisperse silver hydrosol from citrate into several size fractions by filtration on membranes.⁷

Methods allowing the one-step preparation of silver particles with controlled size have recently been described. By the reduction of silver complexes with ammonia by glucose, known as the “silver mirror” reaction or Tollens process, it is possible to prepare silver films consisting of particles between 60 and 180 nm,¹⁷ or stable silver hydrosols with particle size ranging from 20 to 50 nm,¹⁸ depending on the initial system component concentration. With the use of a stabiliser (e.g., gum arabic), a size range from 100 to 1200 nm can be achieved by changing the stabiliser concentration.¹⁹

Another general parameter allowing the control of the size of the prepared particles is a difference in the redox potentials of the reducing agent and the silver ions undergoing reduction. This difference is easily controllable in a defined way by changing in the silver ion complex forming agent concentration in the reaction mixture.²⁰ Thus the possibilities for particle size control in the previously-mentioned “silver-mirror” reaction can be extended because, besides the change in the ammonia concentration, a broad range of reducing sugars can be used in this preparation method.

Thus, the aim of this work is, from the basis of the “silver mirror” reaction, to develop and evaluate a simple one-step method that makes possible the synthesis of silver colloid particles with a controlled size in the broad ranges that are applicable in different fields, including SER spectroscopy, antimicrobial applications, or catalysis. For the purposes of this study, prepared silver colloid particles were tested in SER spectroscopy.

Experimental

Silver nitrate (99.9%) was purchased from Safina, Czech Republic. Ammonia (25% w/w aqueous solution), D-glucose (p.a.), and D-fructose (p.a.) were supplied by Lachema, Czech Republic; D(+)-xylose (≥98%), D(+)-maltose (99+%), and 1-methyl adenine by Sigma-Aldrich. All the reagents employed in this work were of analytical grade. Aqueous solutions were prepared using deionised water. All solutions were filtered using a 200 nm filter prior to their use in the preparation of silver colloids.

The silver colloid particles were prepared by the chemical reduction of silver nitrate in the presence of ammonia, forming a complex with silver ions of [Ag(NH₃)₂]⁺ composition. Four structurally dissimilar reducing sugars were used as reductants—xylose representing five-carbon aldoses, glucose as a representative of six-carbon aldoses, fructose as a six-carbon ketose, and maltose representing disaccharides. The pH of the reaction system was adjusted to about 12.5 using an NaOH solution, so that the reaction took 3 to 10 min to complete. The reduction was initiated by the addition of reducing sugar into the reaction mixture. Initial concentrations of the reaction mixture components were 10⁻³ mol dm⁻³ for AgNO₃ and 0.01 mol dm⁻³ for the reducing sugar, and the concentration of ammonia varied from 0.2 to 0.005 mol dm⁻³. The course of reduction was monitored continually by reaction system transparence measurements set at a wavelength of 635 nm. The saturation tendency in the system transparence was considered to be the point of completion of the reduction process of the silver ions. All experiments were carried out at laboratory temperature (ca. 20 °C) in a 30 ml cuvette protected from light, and the reaction mixture was stirred with the aid of a magnetic stirrer. The cuvette was cleaned after every experiment using diluted (1 : 1) nitric acid.

The particle size distribution of the prepared colloid particles was measured by the dynamic light scattering method (DLS) on a Zeta Plus instrument (Brookhaven Instr. Co.). The average values of the particle size and half-widths of a lognormal approximation of real particle size distribution curves were determined from the DLS measurements. Microscopic measurements of particle sizes were performed with Zeiss Opton 109 (Zeiss Jena) and CM12 TEM/STEM (Philips) transmission electron microscopes. UV-VIS absorption spectra were obtained using a Helios α (Unicam) spectrophotometer. All the reported values of particle sizes in this article were obtained by DLS, if not stated otherwise.

For the purpose of SER spectroscopy with 1-methyladenine, freshly prepared silver colloid was diluted 5 times with deionised water and activated by an NaCl solution. The addition of NaCl corresponded to 2 × 10⁻³, 10⁻² and 0.1 mol dm⁻³ final concentration of chlorides in the reaction mixture. One microliter of 10⁻² mol dm⁻³ 1-methyladenine solution was added into 1 ml of the colloid solution. Raman spectra were measured after 15 min incubation in both non-activated and chloride-activated solutions. To ascertain the surface enhancement magnitude, the Raman spectrum of a 10⁻² mol dm⁻³ solution of 1-methyladenine in deionised water was also measured.

Raman spectra were recorded at laboratory temperature on a Jobin-Yvon T 64 000 spectrometer equipped with a liquid nitrogen-cooled CCD detector. The argon laser (Coherent Innova 90C FreD, Coherent Inc., λ = 514.5 nm) was used for excitation. Spectra were registered in the range from 600 to 2000 cm⁻¹ with 1 cm⁻¹ resolution and 10 s scan time; 16 accumulations were made. The laser light power incident onto a sample was 100 mW, with a 0.1 × 1 mm slit used.

Results and discussion

Preparation of silver colloid particles

The reduction of the [Ag(NH₃)₂]⁺ complex cation, which was formed by the addition of ammonia into the reaction mixture containing silver nitrate, was carried out by reducing sugars and proceeded via an autocatalytic mechanism, as manifested by the characteristic sigmoid shape of the corresponding kinetic curves (see Fig. 1). The length of the induction period (the initial stage of slow product formation) was unambiguously dependent on the ammonia content in the reaction system—it increased from seconds to minutes with ammonia concentrations ranging from 0.005 mol dm⁻³ to 0.2 mol dm⁻³. The maximum reaction rate changed from 5.6 × 10⁻⁵ mol dm⁻³ s⁻¹ (for the lowest ammonia concentration) to 1.3 × 10⁻⁵ mol dm⁻³ s⁻¹ (for the highest ammonia concentration).

Fig. 1 Kinetic curves of silver colloid formation recorded as change of transparency T during the [Ag(NH₃)₂]⁺ reduction by glucose for the ammonia concentrations in the reaction system: 0.005, 0.01, 0.02, 0.05, 0.1 and 0.2 mol dm⁻³ (from the left).

Both average particle sizes and the particle size distributions of the synthesised silver colloids were analysed by DLS and TEM. The average particle size values and half-widths of lognormal distributions of silver colloids depend on the reaction conditions, as determined by DLS. They are summarised in Table 1. Apart from a systematic shift in particle sizes obtained towards the lower values from TEM measurements, no significant differences were found and in this way the observed dependence of silver particle sizes on ammonia concentration was verified. The systematic shift is caused by the difference in the measuring principles and is discussed elsewhere.²¹ The resulting silver particles were of spherical shape, their size distribution dependent on both the ammonia concentration and the type of reducing sugar used. The differences in the particle sizes and morphology depend on reaction conditions, as illustrated by TEM micrographs of samples prepared by reduction of [Ag(NH₃)₂]⁺ using glucose at variable ammonia concentration (Fig. 2). Fig. 3 shows additional TEM images obtained for different reducing sugars and the same concentration of ammonia (0.005 mol dm⁻³) in the reaction mixture.

Fig. 2 TEM images of colloid silver particles prepared by reduction of [Ag(NH₃)₂]⁺ using glucose for (a) 0.02, (b) 0.05, (c) 0.1, and (d) 0.2 mol dm⁻³ ammonia concentration in the reaction mixture.

Fig. 3 TEM images of colloid silver particles prepared by reduction of [Ag(NH₃)₂]⁺ using (a) glucose, (b) xylose, (c) fructose, and (d) maltose for 0.005 mol dm⁻³ ammonia concentration in the reaction mixture.

Table 1 Average particle diameters, d, and half widths of lognormal size distributions, hw, obtained from DLS for the silver colloids prepared by the reduction of [Ag(NH₃)₂]⁺ at different ammonia concentration using various reducing sugars

Reducing sugar	Ammonia concentration/mol dm^−3
	0.005		0.01		0.02		0.05		0.1		0.2
	d/nm	hw/nm	d/nm	hw/nm	d/nm	hw/nm	d/nm	hw/nm	d/nm	hw/nm	d/nm	hw/nm
Xylose	54	24	57	24	67	28	292	135	355	164	372	161
Glucose	57	24	63	28	87	49	270	133	302	156	336	150
Fructose	161	81	168	86	204	75	243	120	295	154	380	202
Maltose	47	13	68	36	116	57	239	112	328	167	352	158

---

For glucose, xylose and maltose, the lowering of the ammonia concentration in the reaction system led to an abrupt decrease in the average size of the formed silver particles between 320 and 380 nm at 0.2 mol dm⁻³ down to 45–60 nm at 0.005 mol dm⁻³ ammonia concentration (Fig. 4a, b; curves A, B and C). The shape of the curves of the average particle size vs. ammonia concentration curves and the size ranges obtained were similar for all three of the above-mentioned reducing sugars. In the case of fructose, the decrease in the average particle size was not so significant (Fig. 4b, curve D)—only from 380 nm to 170 nm.

Fig. 4 The dependence of the average size of colloid silver particles prepared by reduction of [Ag(NH₃)₂]⁺ by glucose (A), xylose (B), maltose (C), and fructose (D) on ammonia concentration.

The absorption spectra of colloid silver particles prepared by the reduction using glucose are shown in Fig. 5. UV-VIS spectra of samples prepared at higher ammonia concentrations (0.05–0.2 mol dm⁻³) with the aid of glucose, xylose and maltose possess flat maxima at 400–500 nm with a high absorption in the whole visible spectral region.

Fig. 5 UV-VIS absorption spectra of colloid silver particles prepared by reduction of [Ag(NH₃)₂]⁺ by glucose at different ammonia concentrations: 0.005 (A), 0.01 (B), 0.02 (C), 0.05 (D), 0.1 (E) and 0.2 mol dm⁻³ (F). Colloid solutions were diluted three times before the measurement of absorption spectra.

This is characteristic for metallic particles with a size on the order of hundreds of nanometers. On the other hand, the spectra of colloids prepared at lower concentrations of ammonia (0.005–0.02 mol dm⁻³) reveal a sharp maximum of surface plasmon absorption at about 420 nm, typical for silver particles whose dimensions are several tenths of nanometers.²² If fructose is used as a reducing sugar, no particles of sizes lower than 100 nm are produced at any ammonia concentration, as is evident from the absorption spectra (where only low flat maxima are observed), as for colloids prepared at high ammonia concentration with the three above-mentioned sugars.

Concerning the colloid stability of the prepared silver hydrosols, it was found to be strongly dependent on particle size. The colloids with particle sizes below 100 nm, prepared at lower ammonia concentrations, are stable for several months, while larger particles reveal a high level of instability as a result of the relatively quick sedimentation process (hours). The colloid stability of silver particles prepared at lower ammonia concentrations is well documented by the time dependencies of absorption spectra (Fig. 6), exhibiting a sharp plasmon absorption maximum even three months after preparation.

Fig. 6 The time dependence of absorption spectra of silver colloid particles prepared by the reduction of [Ag(NH₃)₂]⁺ by maltose at the 0.005 mol dm⁻³ ammonia concentration. The storage times of colloids were: fresh (A), one-week (B), one-month (C), two months (D) and three months (E). Colloid solutions were diluted five times before the measurement of absorption spectra.

The classical theory of new phase formation and growth is applicable in order to explain the observed dependencies of average silver particle sizes on ammonia concentration. This kind of model is appropriate if a weak reducing agent is used, rather than an aggregation mechanism which is more relevant for the reduction of silver ions by a strong reductant, e.g. NaBH₄,²³ or in a solvent reduction process at elevated temperature.²⁴ Following the theory of formation and growth of a new phase, a nucleus can be formed only under suitable thermodynamic conditions characterised by an over-saturation quantity on which the frequency of new phase nuclei formation is dependent.²⁵ Those nuclei which reach the critical dimensions can grow in further reaction stages to give final stable particles. In this subsequent stage, the silver formed by the reduction is preferentially deposited on already-existing nuclei, as a heterogeneous mechanism of nucleation is energetically more favourable than a homogeneous one. Regarding the easier deposition of silver onto the already formed silver particles, the abrupt decrease occurs in over-saturation. The formation of further nuclei by the homogeneous mechanism is then suppressed. At a higher reduction rate, given mainly by the high difference in the redox potentials of Ag⁺/Ag and the reducing agent, a higher degree of over-saturation and a higher frequency of nuclei formation is achieved at the beginning of the reduction process. Therefore, more critical nuclei are formed in the reaction system. The more nuclei that are formed in the initial stage, the smaller the resulting particles are. This is because a given amount of silver (the same for all experiments) is consumed by a larger number of nuclei.

The formation of an [Ag(NH₃)₂]⁺ complex cation in the presence of a sufficiently high ammonia concentration leads to a decrease in the Ag⁺/Ag standard redox potential from +0.799 for uncomplexed Ag⁺/Ag down to +0.38 V for [Ag(NH₃)₂]⁺.²⁰ The difference between the reducing sugar redox potential (it is slightly negative, about −0.15 V for pH 7 and sugars used) and the redox potential of silver is therefore decreased, which results in a decrease in the overall reaction rate. Because of the decreased reaction rate, the number of nuclei formed is lowered, which causes the observed increase in the final average particle size.

Therefore, the influence of the complexing agent concentration on the solution's redox potential is reflected in the resulting silver colloid particle sizes. The potential of Ag⁺/Ag redox couple (silver electrode) is given by Nernst's eqn. (1):

E_Ag/Ag+ = E^f_Ag/Ag+ + (RT / F) × ln [Ag⁺] (1)

where E^f_Ag/Ag+ is the formal potential of the silver electrode and [Ag⁺] is the concentration of free Ag⁺ ions which is influenced by silver ammonia complex formation:

Ag⁺ + 2 NH₃ ↔ [Ag(NH₃)₂]⁺ (2)

This equilibrium is characterised by the practical stability constant of the resulting complex cation, depicted by eqn. (3):

K′ = [Ag(NH₃)₂]⁺ / [Ag⁺] × [NH₃]² (3)

The concentration of free Ag⁺ ions in the reaction mixture can be evaluated from eqn. (3) and next to this the dependence of the silver electrode potential on the ammonia concentration can be evaluated:

E_Ag/Ag+ = E^f_Ag/Ag+ + (RT / F) × ln {[Ag(NH₃)₂⁺] / K′ × [NH₃]²} (4)

The result of the computations discussed above, along with the dependence of silver colloid particle size on ammonia concentration, is given in Fig. 7 (the value of K_β = 6.3 × 10⁻⁸ was used²⁰ in this calculation instead of the unknown value of K′).

Fig. 7 The correlation between colloid silver particles average size d_Ag [curve (A)] and Nernstian term E_C = −(RT / F) × ln [Ag⁺] for redox potential of silver cation [curve (B)] in the reaction mixture containing ammonia. Silver particles were prepared by reduction of [Ag(NH₃)₂]⁺ by maltose at different ammonia concentrations.

In addition to the complex agent, particle formation is also affected by the strength of the reducing agent and its molecular structure, which influences its ability to adsorb on the growing particle, which in turn catalyses electron transfer between the reductant and the Ag⁺ ion undergoing reduction. Thus, differences in the electrochemical and structural properties of the reducing agent affecting the overall rate of silver complex reduction are reflected in the size of the resulting silver particles prepared with different reducing sugars at the lowest ammonia concentration in the reaction system (0.005 mol dm⁻³). The well-known fact that aldehydes are generally stronger reductants than ketones is in accordance with the finding that glucose (six-carbon aldose) produces silver particles with an average size of 60 nm, while the use of fructose (six-carbon ketose) results in 170 nm-sized silver particles. The structural effects related to the different number of carbons in the molecule do not manifest themselves—xylose (five-carbon aldose) yields particles of about 60 nanometers, as in the case of glucose. However, deeper structural changes, such as a changeover to disaccharides, affect the resulting silver particle size. Contrary to glucose, maltose, a disaccharide consisting of two glucose units, gives significantly smaller particles—only 45 nm in diameter. Such a significant reduction in particle size is attributable to a higher (practically twofold) increase in the concentration of the actual reducing agent—glucose—released during the alkaline maltose hydrolysis.

SERS measurements

The application potential of the prepared silver colloid particles was tested in surface-enhanced Raman spectroscopy of 1-methyladenine. For these experiments, the 60 nm-sized particles were prepared by the reduction of ammonia complex cation with xylose at 0.01 mol dm⁻³ concentration of ammonia. Chloride ions were used as an activation agent, which, at a concentration of 0.1 mol dm⁻³, effectively suppressed the background Raman spectrum of the freshly prepared silver hydrosol (Fig. 8). The effect is most probably connected to the silver particle aggregation induced by the high chloride concentration. As manifested by the DLS measurement, 0.1 mol dm⁻³ NaCl causes an instantaneous aggregation, which is reflected in the increase of the average particle size up to 230 nm, but lower chloride concentrations do not produce this effect.

Fig. 8 Raman spectra of silver hydrosol prepared by the reduction of [Ag(NH₃)₂]⁺ with xylose, non-activated (A) and activated by NaCl with concentration of 0.002 (B), 0.01 (C) and 0.1 mol dm⁻³ (D).

The influence of chloride ions on the Raman spectrum of added 1-methyladenine is exactly opposite to that on the spectrum of pure colloid. For the 0.1 mol dm⁻³ sodium chloride concentration, the highest surface enhancement was observed for 1-methyladenine Raman spectra (Fig. 9). Comparing the intensities of the characteristic peaks in surface-enhanced spectrum and in the Raman spectrum of homogeneous 1-methyladenine solution, the enhancement factor is ca. 10⁵. It can be concluded that the silver particles synthesised by the method used are effective in the enhancement of Raman scattering.

Fig. 9 Difference surface enhanced Raman spectra of 1-methyl adenine (10⁻⁵ mol dm⁻³) in the presence of silver hydrosol prepared by the reduction of [Ag(NH₃)₂]⁺ with xylose. The spectra were obtained by subtraction of the spectrum of the pure silver hydrosol from the spectrum recorded for 1-methyl adenine adsorbed on silver hydrosol. Silver hydrosols were activated by 0.002 (A), 0.01 (B), and 0,1 mol dm⁻³ (C) of NaCl. Raman spectrum of 1-methyl adenine (10⁻² mol dm⁻³) in deionised water (D) is included for comparison.

Conclusion

The reduction of the [Ag(NH₃)₂]⁺ complex by various reducing sugars was found to be a simple one-step synthesis of spherical silver colloid particles with controlled size, applicable in surface-enhanced Raman spectroscopy. It has been proven that a decrease in ammonia content in the reaction mixture from 0.2 mol dm⁻³ to 0.005 mol dm⁻³ leads to a decrease in the average particle sizes from 380 down to 45 nm. For the same ammonia concentration, the particle size decrease also depends on the reducing sugar used and is significantly different for fructose compared to glucose, xylose, and maltose. In the UV-VIS spectra of the colloids prepared at low ammonia concentrations, a strong maximum of the surface plasmon absorbance at 420 nm was observed, as evidence of the formation of nanometer-sized silver particles. Silver hydrosol prepared by the reduction of the silver ammonia complex cation by xylose and activated by chloride ions was successfully tested for use in the SERS with 1-methyladenine as a testing substance. An enhancement factor of about 10⁵ was achieved with the prepared sol.

Acknowledgements

The financial support of the MSM 6198959218 and 1M0512 grants from the Ministry of Education of the Czech Republic is gratefully acknowledged. Sincere gratitude to N. Pizúrová (IPM ASCR Brno, Czech Republic) for TEM measurements and V. Mašek (Institute of Biophysics ASCR Brno, Czech Republic) for SERS measurements.

References

1. A. Henglein, Ber. Bunsen-Ges. Phys. Chem., 1997, 101, 1562.
2. P. Alivisatos, Nat. Biotechnol., 2004, 22, 47.
3. M. Fleischmann, P. J. Hendra and A. J. McQuillan, Chem. Phys. Lett., 1974, 26, 163.
4. Y. Wang, Y. S. Li, Z. X. Zhang and D. Q. An, Spectrochim. Acta, Part A, 2003, 59, 589.
5. K. Faulds, W. E. Smith and D. Graham, Anal. Chem., 2004, 76, 412.
6. K. Kneipp, Y. Wang, R. R. Dasari and M. S. Feld, Appl. Spectrosc., 1995, 49, 780.
7. S. R. Emory and S. Nie, J. Phys. Chem. B, 1998, 102, 493.
8. M. Ishikawa, Y. Maruyama, J. Y. Ye and M. Futamata, J. Biol. Phys., 2002, 28, 573.
9. S. M. Nie and S. R. Emory, Science, 1997, 275, 1102.
10. S. R. Emory, W. E. Haskins and S. M. Nie, J. Am. Chem. Soc., 1998, 120, 8009.
11. N. R. Jana, T. K. Sau and T. Pal, J. Phys. Chem. B, 1999, 103, 115.
12. K. L. Kelly, E. Coronado, L. L. Zhao and G. C. Schatz, J. Phys. Chem. B, 2003, 107, 668.
13. S. Schneider, P. Halbig, H. Grau and U. Nickel, Photochem. Photobiol., 1994, 60, 605.
14. N. Shirtcliffe, U. Nickel and S. Schneider, J. Colloid. Interface Sci., 1999, 211, 122.
15. P. C. Lee and D. Meisel, J. Phys. Chem., 1982, 86, 3391.
16. L. Rivas, S. Sanchez-Cortes, J. V. Garcia-Ramos and G. Morcillo, Langmuir, 2001, 17, 574.
17. Y. Saito, J. J. Wang, D. N. Batchelder and D. A. Smith, Langmuir, 2003, 19, 6857.
18. Y. D. Yin, Z. Y. Li, Z. Y. Zhong, B. Gates, Y. N. Xia and S. Venkateswaran, J. Mater. Chem., 2002, 12, 522.
19. K. P. Velikov, G. E. Zegers and A. van Blaaderen, Langmuir, 2003, 19, 1384.
20. D. V. Goia and E. Matijevic, New J. Chem., 1998, 1203.
21. M. Takeo, Disperse Systems, Wiley-VCh, Weinheim, 1999, pp. 330.
22. T. Jensen, L. Kelly, A. Lazarides and G. C. Schatz, J. Cluster Sci., 1999, 10, 295.
23. D. L. Van Hyning, W. G. Klemperer and C. F. Zukoski, Langmuir, 2001, 17, 3128.
24. M. Giersig, I. Pastoriza-Santos and L. M. Liz-Marzan, J. Mater. Chem., 2004, 14, 607.
25. P. G. Debenedetti and H. Reiss, J. Chem. Phys., 1998, 108, 5498.

---