Influence of various chloride ion concentrations on silver nanoparticle transformations and effectiveness in surface enhanced Raman scattering for different excitation wavelengths

Abstract

The effect of six various chloride ion concentrations (25, 50, 100, 200, 400, and 800 mM) on time-dependence and surface enhanced Raman scattering (SERS) signal intensity was investigated for silver nanoparticles (∼28 nm) with a high monodispersity and long time stability. The experiments were performed using three lasers with excitation wavelengths in the visible region (532 nm, 633 nm, 780 nm). Adenine was used as a model analyte. The treatment procedure, when the various sodium chloride solutions added to silver nanoparticles, led to an enhancement in the Raman signal at all studied concentration levels of sodium chloride. Nevertheless, low-concentration chloride ions differently influenced the time course of enhancement efficiency contrary to high-concentration chloride ions. The final concentration of chloride ions equal to 25 mM did not have any pronounced influence on the silver particle sizes and morphologies. The final concentration of chloride ions varying from 50 to 200 mM led to the etching and coalescence of silver nanoparticles. Higher concentrations of chlorides (400 mM) caused re-crystallization of primary silver nanoparticles to one order larger crystallites (400 nm). From the point of view of SERS, the time dependent profiles of Raman signal enhancement differ only slightly for all the used final concentrations of chloride ions when using excitation at 532 nm. On the contrary, for excitation wavelengths 633 nm and 780 nm, the time dependent profiles of Raman signal enhancement were very different when using above mentioned six various final concentrations of chloride ions.

---

Introduction

Due to Fleischmann's observation of Raman spectra of pyridine adsorbed at a silver electrode,¹ and mainly Creighton's observations of intense Raman scattering of pyridine molecules adsorbed on silver colloid particles,² surface-enhanced Raman scattering (SERS) spectroscopy has become a powerful analytical technique where metal particles or layers are exploited for detection of important molecules.^3,4 SERS is a very sensitive technique which allows the acquisition of spectra at very low concentrations of analytes.^5–7 This ability makes SERS a versatile tool for easy and rapid detection of the important compounds which may be a clue to, for example, criminal investigations or diagnostic purposes. Silver or gold nanoparticles which are the most frequently used SERS substrates^8–12 have been successfully applied for measurements of addictive substances,^13–15 unsafe food,^16–18 or, especially in the last few years, bacteria.^19–27

Among the metals, silver based substrates are widely used for SERS purposes, because of the cost, relatively easy preparation, and very good enhancement effects of such materials. Studies revealed that the shape, size and distance between silver nanoparticles play an important role providing an enhancement of the Raman signal.^28,29 Recently, it has been shown that the highest value of Raman signal enhancement was achieved on the so-called ‘hot silver particles’. The optimal sizes of such particles are between 110 to 120 nm for the commonly used argon laser with a wavelength of 514.5 nm.^7,30 On the other hand, silver nanoparticles with sizes of a few tens of nanometers can be utilized as effective SERS substrates too. Nevertheless, these small particles themselves usually do not provide significant enhancement of the Raman signal. For these purposes, they must be treated, usually by the addition of some inorganic ions, mainly chlorides. The presence of ions leads to the aggregation of particles where the junction sites among nanoparticles provide the very strong enhancement of the SERS signal. The most used silver colloids for SERS purposes (treated by the addition of ions) are prepared using of sodium citrate, sodium borohydride and hydroxylamine hydrochloride as reducing agents. Nevertheless, there are a few studies comparing influence of different concentrations of ions on characteristics of silver colloid and its activity in SERS.^31–35 Liu et al. provided comprehensive studies of 66 mM halides (NaCl, NaBr, and NaI) activation of borohydride-reduced silver colloid for SERS of riboflavin.³³ Dong et al. reported a comparative study of the effects of various sodium salts (NaCl, NaNO₃, Na₂SO₄, and NaI) on the Raman signal of methylene blue (MB) adsorbed on hydroxylamine-reduced silver colloid with particle size of about 34 nm.³⁶ Among the four mentioned salts used for SERS effect, the addition of NaCl solution provides the greatest enhancement of Raman signal due to the synergy of the modification of silver particle surface and the particle aggregation caused by chlorides, while NaI gives the least enhancement of Raman signal. On the other side, Aubard et al. observed quenching of SERS-signal from benzoic acid after the addition of various anions (Br⁻, I⁻, and Cl⁻). This work brought interesting results about mechanism of ions exchange which proceed on silver particle surface. They have observed the decreasing SERS intensities with the increasing concentration of sodium salts up to 10 mM.³⁷

There are no studies that have investigated the influence of ions in a wide range of concentrations on changes of one type silver nanoparticles with respect to its efficiency to enhance of Raman signal. For such study, we choose maltose-reduced silver nanoparticles (NPs) with high monodispersity and great time stability.³⁸ For treating of the silver NPs with an average particle size equal to 28 nm, we applied six different final concentrations of chloride ions (25, 50, 100, 200, 400, and 800 mM). The changes in silver nanoparticle characteristics were monitored by measuring the average particle sizes and by recording the UV/Vis spectra. The kinetics of Raman signal enhancement after addition of chloride ions was investigated for three excitation wavelengths (532, 633, and 780 nm).

Experimental

Materials and chemicals

Silver nitrate (AgNO₃, p.a., Fagron), ammonia (NH₄, aqueous solution 28% w/w, p.a., Sigma), sodium hydroxide (NaOH, p.a., Lachema), D(+)-maltose monohydrate (C₁₂H₂₂O₁₁·H₂O, p.a., Sigma) were used for preparation of silver nanoparticles without any further purification. Sodium chloride (NaCl, p.a., Sigma-Aldrich) was used as a treatment agent of primarily prepared silver nanoparticles. The adenine with certified purity of 99% (Sigma-Aldrich) was used without any further purification too. Deionized water was used for all experiments with conductivity 0.05 μS cm⁻¹ obtained from instrument Aqual 29 (Merci).

Instrumentation

The size of synthesized silver nanoparticles as well as the course of changes of particle sizes formed after the addition of sodium chloride to the primarily prepared silver nanoparticles was characterized by a dynamic light scattering (DLS) using a Zeta Plus analyzer (Brookhaven, USA). UV/Vis absorption spectra of the silver nanoparticles were obtained by using a Specord S600 (Analytic Jena AG, Germany). Transmission electron microscopy (TEM) images of silver nanoparticles were recorded on a Jeol JEM 2010 (Japan) electron microscope at 160 kV of the acceleration voltage. SERS and normal Raman spectra were recorded using a DXR Raman Microscope (Thermo Scientific) equipped with a thermoelectrically cooled (−50 °C) charge-coupled device (CCD) camera and a 4× objective. All spectra were measured at room temperature in a quartz cell. SERS spectra have been measured by three laser sources with wavelengths of 532 nm (diode-pumped, solid state laser), 633 nm (He–Ne gas laser) and 780 nm (frequency-stabilized single mode diode laser). In all three cases, the SERS spectra were acquired in the range from 40 to 1880 cm⁻¹ where the spectral acquisition was repeated 16 times with 5 seconds accumulation time. For all three different excitation wavelengths, the laser light power incident onto a sample was adjusted at to 8.0 mW and slit to 50 μm.

Preparation of silver nanoparticles

The maltose reduced silver nanoparticles, used as a SERS substrate, were prepared by a procedure described by Panacek et al.³⁹ This procedure lies in a reduction of the silver ammonia complex cation [Ag(NH₃)₂]⁺ with D-maltose. The concentrations of the reaction components were 10⁻³ mol L⁻¹ and 10⁻² mol L⁻¹ for AgNO₃ and the reducing sugar, respectively. The concentration of the used ammonia was 5 × 10⁻³ mol L⁻¹. The basic environment of the reaction system was adjusted to the value of pH at 11.5 ± 0.1 by adding of sodium hydroxide solution. A reaction time was achieved during 4 minutes. These maltose-reduced silver nanoparticles are nearly monodispersed with an average particle size of about 28 nm. The as-prepared silver nanoparticles were characterized by measuring the UV/Vis absorption spectra with narrow surface plasmon absorption peak at 410 nm wavelength. All the measurements were performed at the laboratory temperature (20 °C). The resulting brownish aqueous dispersion of the silver nanoparticles exhibits a long stability in time during a period of ∼2 years.³⁸

Sample preparation for characterization and SERS measurements

For the purpose both of DLS and UV/Vis measurements, 0.4 mL of the stock aqueous solution of silver nanoparticles was diluted by deionized water ranged from 1.20 to 1.55 mL and then 0.050–0.400 mL of 1 M or 4 M NaCl solution was added. After adding solution of chloride ions, the mixture was quickly shaken and immediately the measurements began. For SERS measurements, six working mixtures with different final concentration of sodium chloride were prepared as follows: the amount of 0.4 mL of the stock aqueous solution of silver nanoparticles was added to deionized water ranged from 1.18 to 1.53 mL. Then, 0.050–0.400 mL of 1 M or 4 M NaCl solution was added in order to characterize the influence of chlorides. After adding chloride ions to the diluted dispersion of silver nanoparticles, the mixture was quickly vortexed and finally 20 μL of 10⁻³ M solution of adenine was added (the final concentration of adenine was 10⁻⁵ M) thus, the total volume was 2 mL. The working mixture was shaken and the SERS measurements began immediately. This treatment procedure was the same for each laser source. The total volume of working solution was held constant for each measurement when the various chloride concentrations and different lasers were used. The experiments were repeated three times in order to verify the reproducibility. The SERS enhancement factors were calculated as a ratio of SERS signal of 10⁻⁵ M adenine to Raman signal of 0.1 M adenine for the strongest band in the spectrum at ∼734 cm⁻¹.

Results and discussion

In our previous work, it has been shown that the high concentrated aqueous solution of NaCl (a final concentration of 400 mM) is very efficient for the SERS activation of maltose-reduced silver nanoparticles (∼28 nm). Such very high concentration of chloride ions leads to rapid re-crystallization of the primarily prepared silver nanoparticles to particles with the size around 400 nm which very efficiently enhanced Raman signal of adenine both for visible (488 nm) and near infrared laser excitation (1064 nm).³⁸

In this study, we focus on effects of various concentrations of chloride ions on the characteristics of silver nanoparticles (particle size, UV/Vis absorption, SERS activity) because such complex study is missing and is important for the utilization and understanding of silver nanoparticles as enhancers of Raman signal after treatment by chloride ions. The experiments were performed for six different final concentrations of chlorides (25, 50, 100, 200, 400, and 800 mM) in aqueous dispersion of silver nanoparticles and for three excitation lines (532 nm, 633 nm, 780 nm). Silver nanoparticles obtained by the above mentioned procedure are nearly monodispersed with a longtime stability. After addition of NaCl solution to the diluted dispersion of silver nanoparticles, the changes in its properties were monitored by measuring of average particle size and by recording of UV/Vis spectra. Change of the average particle size measured by DLS as well as the UV/Vis absorption spectra were recorded each minute during 30 minutes in order to have information regarding the ongoing transformation processes of primary silver nanoparticles. In order to confirmation of particle size changes including their morphology, samples at the time of 15 minutes after the addition of NaCl solution were taken for TEM analysis. In the case of SERS measurements, the SERS spectra were collected each third minute during a period of 30 minutes after the addition of the NaCl solution. For the evaluation and comparison of Raman intensities, the observation of the whole spectrum is important. Nevertheless, for a simplification of calculation we have chosen only one peak at ∼734 cm⁻¹. The intensities of the other adenine peaks varied in proportion to intensity of the main peak at ∼734 cm⁻¹. The SERS spectrum in Fig. 1 shows that this peak (marked with star) is the strongest among all and comes from the ring-breathing mode in the adenine structure (inset picture in Fig. 1).

Fig. 1 SERS spectrum and chemical structure of adenine.

SERS measurements were performed with excitation wavelengths of 532 nm, 633 nm and 780 nm. Raman intensity dependences on time are shown in the Fig. 2. For the excitation wavelength 532 nm, there were only slight differences among the varying concentration of chloride ions (see Fig. 2a).

Fig. 2 Time-dependence profiles of SERS signal intensities of the 734 cm⁻¹ band of adenine after the addition of six different concentrated solution of NaCl for the excitation wavelengths of 532 nm (A), 633 nm (B) and 780 nm (C).

For all used final NaCl concentrations, we have observed approximately twenty percent difference in intensities of the adenine band at ∼734 cm⁻¹. In the case of excitation at 633 nm, there we have observed considerable differences among the SERS intensities of adenine for each concentration of NaCl (Fig. 2b). The highest enhancement of the Raman signal was obtained with the use of 400 mM final concentration of chloride ions. The SERS signal values achieved for the excitation wavelength of 780 nm were lower than in the case with excitation wavelengths of 532 nm and 633 nm, as seen in the Fig. 2c. There the highest enhancement of the Raman signal was observed when 400 mM final concentration of chlorides was used, although the absolute enhancement was considerably lower compared to 532 nm and 632 excitation lines. The SERS enhancement factors for the excitation lasers of 532 nm, 633 nm, and 780 nm were estimated to be 7.2 × 10⁴, 2.0 × 10⁵, and 2.4 × 10⁵, respectively.

As can be seen (Fig. 2a), the enhancement and time-dependent enhancement profiles for excitation at 532 nm are very similar for different final concentrations of chlorides on the contrary of using of 633 nm and mainly 780 nm excitation line. The possible explanations of considerably high Raman signal enhancement for 532 nm and 633 nm excitation wavelengths when using of 25 mM chlorides lies in the so-called hot spot regions.⁴⁰ Such concentration of chloride ions can cause temporarily approaching of certain fraction of particles in the solution, which are close to each other enough to form active sites for Raman signal enhancement. Nevertheless, for excitation wavelength 780 nm the using of 25 mM chlorides was ineffective. The increase in SERS intensities at 780 nm excitation wavelength using such low concentration of chloride ions was not observed due to the fact that the addition of chloride ions did not induce changes in silver nanoparticle characteristics sufficiently and the surface plasmon resonance of such treated Ag NPs is off the resonance with the use of the laser 780 nm. The original size of the primary silver nanoparticles has been changed from 28 nm to 40 nm. Immediately, during the first minute after the adding of 25 mM solution of NaCl to the primary Ag NPs, there was change in the particle size from 28 nm to 40 nm, and after that the average particle size gradually decreased during a period of 30 minutes as is shown in the time-dependence of particle size measured by the DLS method (Fig. 3a). Because of the every minute averaging of particle size, temporary approaches of two or more Ag NPs have influence on slight growth of the average particle size to 40 nm. Nevertheless, on the basis of the observation of the UV/Vis absorption spectra and TEM image, there weren't observed considerable changes in particle size. Fig. 3d shows the UV/Vis absorption spectra of Ag NPs mixed with the final concentration of chloride ions equal to 25 mM. It has been observed obviously that the addition of 25 mM chlorides doesn't change the position of the absorption peak during the whole measurement, because the change of the average particle size was very negligible. Nevertheless, the weak increase of absorption at 532 nm and 633 nm has been observed. This correlates with TEM image which was taken at the time of 15 minutes after the addition of chloride ions solution (Fig. 5a). TEM image of primarily prepared Ag NPs is presented in ESI (Fig. S1†).

Fig. 3 Time dependences of the average size of Ag NPs and its UV/Vis absorption spectra recorded during 30 minutes after the addition of the NaCl solution. The time dependences represented in plots (A)–(F) correspond to final concentration 25 mM, 50 mM and 100 mM of NaCl in mixture, respectively.

Fig. 4 Time dependences of the average size of Ag NPs and its UV/Vis absorption spectra recorded during 30 minutes after the addition of the NaCl solution. The time dependences represented in plots (A)–(F) correspond to final concentration 200 mM, 400 mM and 800 mM of NaCl in mixture, respectively.

Fig. 5 Representative TEM images of the treated Ag NPs after 15 minutes from the addition of the NaCl solution with final concentration (A) 25 mM, (B) 50 mM, (C) 100 mM, (D) 200 mM, (E) 400 mM and (F) 800 mM.

For the final concentration of chlorides equal to 50 mM, there were observed higher SERS intensities for the laser 532 nm, which were stable during the whole experiment (Fig. 2a). In the case of excitation at 633 nm, a considerable growth of enhancement was achieved within few minutes after the addition of Cl⁻ (Fig. 2b). The enhancement slowly grew further up with the increasing particle size. As it can be seen in the Fig. 2c, there is nearly negligible growth of SERS intensities for laser excitation at 780 nm. This fact can be explained by a considerable change in the size of Ag NPs after the addition of Cl⁻ to Ag NPs (Fig. 3b). After the third minute following the addition of sodium chloride, there was an abrupt increase in the particle size. The size increased until the 25^th minute, where it was about 300 nm. In addition, particle size changes were accompanied by a change in the color of the working mixture of Ag NPs (from brownish to grey). Fig. 3e exhibits the UV/Vis absorption spectra recorded immediately after the addition of 50 mM chlorides. After three minutes, the surface plasmon resonance peak of the original nanoparticles (410 nm) was suppressed and at the same time the secondary peak in the region 570 nm appeared. This indicates that the Ag NPs are still suitable for lasers 532 nm, 633 nm but not for 780 nm. These changes correlate with a respective TEM image (Fig. 5b), which revealed the presence of larger silver particles. On the basis of performed measurements, there it can be deduced that the addition of 50 mM chlorides causes transformations of original nanoparticles suitable for SERS measurements at 532 nm and 633 nm, but this low concentration level of chlorides has still poor effect for enhancement of the Raman signal using laser 780 nm.

The laser with excitation wavelength at 780 nm is generally not commonly used, because the most common SERS-active substrates exhibiting the largest enhancement effects are based on colloidal silver, which has the strongest plasmon field lower than mentioned excitation wavelength at 780 nm. It is known that the surface plasmon resonance wavelength is shifted with a size of the particles or with distance among them and it has to be in resonance with Raman laser to achieve the best signal enhancement. The mostly used lower final concentrations of sodium chloride (in orders of units of mM) are not sufficient to increase the size of the particles enough to enhance the Raman signal equipped with laser of longer wavelengths, but as we have observed in our previous study, the addition of highly concentrated sodium chloride solution (hundreds of mM) to maltose-reduced silver colloid increases the particle size enough to achieve an effectiveness of enhancement of the Raman signal even for near infrared laser excitation (1064 nm).³⁸

In terms of activation by 100 mM and 200 mM chloride ions, the very similar effect of these concentrations was observed, albeit the Raman intensities were higher with the use of 200 mM NaCl. Regarding the laser operating at 532 nm, there the Raman intensities for both concentration levels are stable during a whole measuring time (Fig. 2a). Addition of final concentration 100 or 200 mM NaCl didn't cause any significant changes in Raman intensities during the measurements when laser 633 nm was utilized (Fig. 2b and c); the enhancement was stable during the whole monitored time. For laser 780 nm, the signal intensities increased for 50% regarding 100 mM and 500% regarding the use of 200 mM (Fig. 2b and c). Correlation among the DLS method, UV/Vis absorption spectra and the morphologies observed in the TEM image indicates that these Cl⁻ concentrations induced changes in the particle size with the formation of both larger and smaller particles due to the re-growth and diffusion of the origin nanoparticles in the presence of the 100 mM Cl⁻ and 200 mM Cl⁻. Considerable growth of the average particle size from 28 nm up to 400 nm was observed (Fig. 3c and 4a). Most of the changes occur within the first few minutes after the addition of chloride ions. During the thirty minute record, the surface plasmon band was shifted to ∼650 nm and the plasmon band at 410 nm for origin nanoparticles was suppressed (Fig. 3f and 4d). The suppression of the original peak at 410 nm and at the same time the emerging of the secondary peak clearly indicates the onset gradual changes in the average size³⁸ or morphologies⁴¹ of silver particles. As the size of the silver particles increases, the value of UV/Vis absorption maximum becomes closer to excitation wavelengths, and higher SERS-signal enhancement is achieved. All these characteristic changes were accompanied by a change in the color of the working mixture of Ag NPs (from brownish to grey).

Particles treated with 400 mM chlorides caused the most effective Raman enhancement for all the three mentioned excitation sources, compared to the other concentrations of chlorides. In the time region 1–16 minute after the addition of 400 mM chlorides to Ag NPs, there we observed a gradual rise in Raman intensities (Fig. 2a–c). It is attributed to an increase in the particle size caused by the presence of highly concentrated chloride ions. After the mentioned time, the SERS signal decreases because the colloid becomes over-aggregated. Due to the aggregation effect, large radii of new particles (up to 400 nm), reached after 3 minutes after the addition of NaCl solution, leads to a decrease of the available particle surface, and thus its total saturation can be achieved more easily,⁴² as is shown in the Fig. 4b. The particle growth was accompanied by both a change in the color of the working mixture of Ag NPs (from brownish to grey) and a distinct red shift of the maximum in the UV/Vis absorption (Fig. 4e).

For the highest concentration of chlorides, 800 mM, and laser excitations 532 nm and 633 nm, there were observed decreasing Raman intensities with increasing incubation time (Fig. 2a and b). One of the possible explanations for the observed SERS-signal intensities decrease can be attributed to a surface-coverage limitation caused by a competition of the Cl⁻ ions and molecules of adenine for a surface on the present silver particles.³⁷ The measurement of SERS on silver surface with excitation source 780 nm, there it has been observed the maximum intensity sooner, compared to 400 mM chlorides, already in the 7^th minute after the addition of chloride ions solution. However the signal immediately decreased (Fig. 2c). Following DLS measurements uncovered a significant growth of the average particle size to 450 nm (Fig. 4c), which is in a good accordance with the respective UV/Vis absorption spectra where, in the first minute of the experiment, the peak at 410 nm decreased and broadened and in parallel, peak at 700 nm raised (Fig. 4f). Further aggregation of the particles leads to their destabilization which causes a total loss of both absorption maxima. As it has been noticed, the simultaneous decrease of absorption and the long-wavelength shift is more dramatic than in the case of the previous lower chlorides concentrations, which is in a correlation with a TEM image of particles treated using 800 mM chlorides (Fig. 5f). Higher concentrations of NaCl added to original silver colloid led to a formation of large silver aggregates which was accompanied by a change in the color of the working mixture of Ag NPs (from brownish to grey). These aggregates sediment, and therefore the amount of SERS-active particles decrease.

For comparison, we also performed additional experiments regarding impact of three final chlorides concentrations (25 mM, 100 mM, and 400 mM) on the changes of Ag NPs with average particle size equal to approx. 50 nm and 120 nm (see ESI†).

Conclusions

The presented study was focused on the influence of six different final concentrations of chloride ions (25, 50, 100, 200, 400, and 800 mM) used as an agent for silver nanoparticles activation for purposes of surface enhanced Raman spectroscopy. The study correlates selected concentration levels of chloride ions with the intensity of the SERS signal using of adenine as a model compound. It was found out that the concentration level of chloride ions plays a crucial role and a careful tuning of experimental design has to be performed for each laser wavelength. Lower final concentrations of chloride ions (tens of mM) led to a transformation of silver nanoparticles in a different manner (mainly aggregation) compared to higher concentrations (hundreds of mM). Higher concentration of NaCl (in this case 400 mM) let to the best results for all three tested laser wavelengths (532, 633 and 780 nm). Lower concentrations of chlorides are also usable for 532 nm laser, but its action has less or no effect on the Raman signal enhancement for lasers 633 and 780 nm. On the other hand, treatment with 800 mM chlorides led to a rapid loss of the Raman signal due to the fast destabilization of silver particles for all three tested laser wavelengths.

Acknowledgements

The authors are grateful for the support by the project LO1305 of the Ministry of Education, Youth and Sports of the Czech Republic. This work has been also supported by the Operational Program Education for Competitiveness – European Social Fund (CZ.1.07/2.3.00/20.0056), internal grant of Palacky University in Olomouc (PrF_2014_032) and by the Technology Agency of Czech Republic (Project no. TA03011368).

Notes and references

1. M. Fleischmann, P. J. Hendra and A. J. McQuillan, Chem. Phys. Lett., 1974, 26, 163–166.
2. J. A. Creighton, C. G. Blatchford and M. G. Albrecht, J. Chem. Soc., Faraday Trans. 2, 1979, 75, 790–798.
3. H. Huang, Ch. Shende, A. Sengupta, F. Inscore, C. Brouillette, W. Smith and S. Farquharson, J. Raman Spectrosc., 2012, 43, 701–705.
4. R. L. Garrell, Anal. Chem., 1989, 61, 401A.
5. V. Ranc, J. Hruzikova, K. Maitner, R. Prucek, D. Milde and L. Kvitek, J. Raman Spectrosc., 2011, 43, 971–976.
6. M. Moskovits, Rev. Mod. Phys., 1985, 57, 783–826.
7. K. Kneipp, Y. Wang, R. R. Dasari and M. S. Feld, Appl. Spectrosc., 1995, 49, 780–784.
8. L. Kvitek, R. Prucek, A. Panacek, R. Novotny, J. Hrbac and R. Zboril, J. Mater. Chem., 2005, 15, 1099–1105.
9. P. C. Lee and D. Meisel, J. Phys. Chem., 1982, 86, 3391–3395.
10. N. Leopold and B. Lendl, J. Phys. Chem. B, 2003, 107, 5723–5727.
11. R. Prucek, A. Panacek, J. Soukupova, R. Novotny and L. Kvitek, J. Mater. Chem., 2011, 21, 6416–6420.
12. R. Prucek, V. Ranc, O. Balzerova, A. Panacek, R. Zboril and L. Kvitek, Mater. Res. Bull., 2014, 50, 63–67.
13. F. Inscore, Ch. Shende, A. Sengupta, H. Huang and S. Farquharson, Appl. Spectrosc., 2011, 65, 1004–1008.
14. R. Pérez, A. Rupérez and J. J. Laserna, Anal. Chim. Acta, 1998, 376, 255–263.
15. P. C. White, Sci. Justice, 2000, 40, 113–119.
16. L. He, Z. Liu, M. Lin, J. Awika, D. R. Lidoux, H. Li and A. Mustapha, Sensonry and Instrumentation for Food Quality, 2008, 2, 66–71.
17. N. N. Yazgan, I. H. Boyaci, A. Topcu and U. Tamer, Anal. Bioanal. Chem., 2012, 403, 2009–2017.
18. B. Liu, P. Zhou, X. Liu, X. Sun, H. Li and M. Lin, Food Bioprocess Technol., 2012, 6, 710–718.
19. R. Prucek, V. Ranc, L. Kvitek, A. Panacek, R. Zboril and M. Kolar, Analyst, 2012, 137, 2866–2870.
20. M. Kazanci, J. P. Schulte, C. Douglas, P. Fratzl, D. Pink and T. Smith-Palmer, Appl. Spectrosc., 2009, 63, 214–223.
21. Y. L. Liu, Y. R. Chen, X. W. Nou and K. L. Chao, Appl. Spectrosc., 2007, 61, 824–831.
22. S. Ramya, R. P. George, R. V. Subba Rao and R. K. Dayal, Appl. Surf. Sci., 2010, 256, 5108–5115.
23. N. P. Ivleva, M. Wagner, A. Szkola, H. Horn, R. Niessner and C. Haisch, J. Phys. Chem. B, 2010, 114, 10184–10194.
24. N. P. Ivleva, M. Wagner, H. Horn, R. Niessner and C. Haisch, J. Biophotonics, 2010, 3, 548–556.
25. J. Guicheteau, S. Christesen, D. Emge and A. Tripathi, J. Raman Spectrosc., 2010, 41, 1632–1637.
26. C. Fan, Z. Q. Hu, A. Mustapha and M. S. Lin, Appl. Microbiol. Biotechnol., 2011, 92, 1053–1061.
27. S. L. Hennigan, J. D. Driskell, N. Ferguson-Noel, R. A. Dluhy, Y. Zhao, R. A. Tripp and D. C. Krause, Appl. Environ. Microbiol., 2012, 78, 1930–1935.
28. L. Kvitek, R. Prucek, A. Panacek, J. Soukupova and M. Vanickova, Nanocon 2009, Conference Proceedings, 2009, pp. 74–82.
29. K. G. Stamplecoskie and J. C. Scaiano, J. Phys. Chem. C, 2011, 115, 1403–1409.
30. S. M. Nie and S. R. Emory, Science, 1997, 275, 1102–1106.
31. N. R. Yaffe and E. W. Blanch, Vib. Spectrosc., 2008, 48, 196–201.
32. M. G. Espinoza, M. L. Hinks, A. M. Mendoza and D. P. Pullman, J. Phys. Chem. C, 2012, 116, 8305–8313.
33. F. Liu, H. Gu, Y. Lin, Y. Qi, X. Dong, J. Gao and T. Cai, Spectrochim. Acta, Part A, 2012, 85, 111–119.
34. X. Dong, H. Gu, J. Kang, X. Yuan and J. Wu, Colloids Surf., A, 2010, 368, 142–147.
35. M.-S. Hsu, Y.-W. Cao, H.-W. Wang, Y.-S. Pan, B.-H. Lee and Ch.-L. Huang, ChemPhysChem, 2010, 11, 1742–1748.
36. X. Dong, H. Gu and F. Liu, Spectrochim. Acta, Part A, 2012, 88, 97–101.
37. J. Aubard, E. Bagnasco, J. Pantigny, M. F. Ruasse and G. Lévi, J. Phys. Chem., 1995, 99, 7075–7081.
38. R. Prucek, A. Panacek, A. Fargasova, V. Ranc, V. Masek, L. Kvitek and R. Zboril, CrystEngComm, 2011, 13, 2242–2248.
39. A. Panacek, L. Kvitek, R. Prucek, M. Kolar, R. Vecerova, N. Pizurova, V. K. Sharma, T. Nevecna and R. Zboril, J. Phys. Chem. B, 2006, 110, 16248–16253.
40. L. Weiyang, H. C. C. Pedro, L. Xianmao and X. Younan, Nano Lett., 2009, 9, 485–490.
41. Y. Yang, S. Matsubara, L. Xiong, t. Hayakawa and M. Nogami, J. Phys. Chem. C, 2007, 111, 9095–9104.
42. S. Sanchez-Cortes, J. V. Garcia-Ramos, G. Morcillo and A. Tinti, J. Colloid Interface Sci., 1995, 175, 358–368.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra13881c

---