Raman spectroscopy for hydrogen peroxide scavenging activity assay using gold nanoshell precursor nanocomposites as SERS probes

Abstract

In this work, we have developed a novel SERS-based approach to detect hydrogen peroxide (H₂O₂) scavenging activity by using gold nanoshell precursor nanocomposites (SiO₂/GNPs) as nanoprobes. H₂O₂ can reduce AuCl₄⁻ to Au⁰ and enlarge the gold nanoparticles (GNPs) that attached on the surface of SiO₂. As the concentration of H₂O₂ increases, the surface coverage of resultant gold on silica cores increases accordingly until continuous gold nanoshells (GNSs) are formed. During the growth process, there is a strong correlation between the SERS-activity of the GNSs and the amount of H₂O₂ that is used as reductant. When H₂O₂ reaches 250 μM, the resultant GNSs show the highest SERS-activity. H₂O₂ can be scavenged by antioxidants such as tannic acid and L-apple acid. Their H₂O₂ scavenging activities were determined by restraining the H₂O₂-mediated (250 μM) growth of SiO₂/GNPs. The decrease of the SERS-activity was proportional to the H₂O₂ scavenging activity of the antioxidant. The results showed that tannic acid had a much higher H₂O₂ scavenging activity than that of L-apple acid.

---

1. Introduction

The importance of evaluating the antioxidant activity has been widely reported in the past two decades not only for the extensive application of antioxidants in food industry and food nutriology, but also because of the discoveries of the strong correlations between reactive oxygen species (ROS) and many pathogenesis of human diseases,^1,2e.g. nervous system injuries,³ atherosclerosis and cancer.^4,5Hydrogen peroxide (H₂O₂), a key part of ROS, is generated in vivo under physiological conditions by peroxisomes (a kind of oxidative enzyme including glucose oxidase and D-amino acid oxidase) and superoxide dismutase which catalyze the dismutation of superoxide radicals. Along with superoxide anion radical (O₂˙⁻), H₂O₂ can damage many cellular components and further convert into more ROS such as hydroxyl radical. Thus, the evaluation of H₂O₂ scavenging activity is an important part for assessing the ROS scavenging activity. Generally, the evaluation of H₂O₂ scavenging activity is peroxidase based or enzyme-free but biomarker dependent.^1,6,7 Many methodologies based on different mechanisms have been developed for detecting H₂O₂ and/or H₂O₂ scavenging activity, e.g. chemiluminescence (CL),^2,7fluorometry,⁶colorimetry,^2,6cyclic voltammetry (CV),¹ electro spin resonance (ESR),¹etc. However, for the peroxidase-based approach, many compounds can be served as a substrate for the peroxidase enzyme, resulting in error assessment. On the other hand, the biomarker dependent methods require a complicated sample preparation process and it is time consuming.^1,6 Another common method for assessing the H₂O₂ scavenging activity is based on the intrinsic absorbance value of H₂O₂ at 230 nm,^8,9 which will decrease when H₂O₂ is scavenged by antioxidants. However, the antioxidants may also have an absorbance range at 230 nm which depress the precision and the accuracy of this method.¹ Therefore, it is still a challenge to develop a novel approach with high sensitivity, high speed, low cost, and easy operation to detect the H₂O₂ scavenging activity.

Many materials, such as quantum dots, metal nanoparticles and polymer spheres have been utilized for the detection of H₂O₂ scavenging activity.^10–13Gold nanoshells (GNSs), a layered dielectric core coated with a gold shell, have gained extensive attentions as they show charming tunable optical properties and good biocompatibility.^14–16 Many applications of GNSs have been reported, e.g. whole blood immunoassays, optical triggerable drug delivery, photothermal cancer therapy.^17–20 Besides their tunable optical properties, a great enhancement of Raman scattering cross-section of adsorbed molecules can be achieved by applying GNSs as surface-enhanced Raman scattering (SERS) substrates.^21,22 The SERS is a new powerful analytical technique with ultrahigh sensitivity.^23–27 In addition to the “whole-finger-prints” information and the low water background, SERS has high signal enhancement and fluorescence quenching effects, which are extremely desirable for the biological sample analysis. Because of its unique attributes, SERS has been used in a wide range of fields, including homeland security,²⁷food safety,²⁸ clinical diagnosis,^29,30 trace detection,^23,31 chemical and biochemical monitoring,^32–34etc., since it was discovered in the late 1970s.^35–37 However, the significance and potential applications of the SERS-activity changes during the SERS substrates forming process are often neglected.

In the present work, we first studied the changing trend of the SERS-activity of the gold nanoshell precursor nanocomposites (SiO₂/GNPs) upon treatment with different concentrations of H₂O₂. Then, we developed a new enzyme-free method for assessing H₂O₂ scavenging activity by applying SiO₂/GNPs as SERS nanoprobes. The SiO₂/GNPs were produced by adsorbing ultrasmall gold nanoparticles (GNPs) onto the silica core surface. H₂O₂ can reduce AuCl₄⁻ to Au⁰ and enlarge the GNPs, mediating the growth of SiO₂/GNPs. During the growth process, the SERS-activity of the SiO₂/GNPs with different enlargement conditions manifested an obvious change concomitant with the increase of gold coverage on the silica cores, which was quantitatively correlated with the concentration of H₂O₂ used. The presence of antioxidant with H₂O₂ scavenging activity would prevent the H₂O₂-mediated growth of SiO₂/GNPs and caused a drop in SERS-activity. The decrease of the SERS-activity was proportional to the H₂O₂ scavenging activity of the antioxidant. The method presented here is simple to design; furthermore, it embraces the advantages of SERS which are extremely desirable for the biological sample analysis.

2. Experimental section

2.1 Materials and reagents

3-(Aminopropyl)-triethoxysilane (APTES, 98%) and Silica colloidal spheres (∼110 nm) were obtained by Sigma and Nissan Chemical Ind., Ltd., Japan, respectively. Sodium borohydride (NaBH₄), hydrogen peroxide (H₂O₂, 30%), chloroauric acid terahydrate (HAuCl₄·4H₂O), potassium carbonate (K₂CO₃), and anhydrous ethanol were bought from Nanjing Sunshine Biotechnology Ltd., China. Niel blue A sulfate (NBA) was purchased from Sigma-Aldrich. Tannic acid and L-apple acid (L-malic acid) were purchased from Shanghai Aladdin Chemical Ltd., China. All of the chemical reagents were analytical grade and used without further purification. K₂CO₃/HAuCl₄ solution (K–H growth solution) was prepared by adding 3 mL of 1% (W V⁻¹) HAuCl₄·4H₂O stock solution into 98.5 mL of aqueous K₂CO₃ solution (0.5 g L⁻¹) under continuous stirring for 20 min and then aged in the dark at 4 °C for 1 day. H₂O used in the experiments was prepared by Milli-Q system (resistivity >18 MΩ).

2.2 Synthesis of GNPs

GNPs (∼5 nm) were prepared by the reduction of HAuCl₄ with NaBH₄ as described elsewhere.¹⁶ Briefly, 3 mL of 1% (W V⁻¹) HAuCl₄·4H₂O was mixed with 200 mL of H₂O under vigorous stirring for about 5 min, followed by the addition of 1 mL of K₂CO₃ (0.2 M). Then, 9 mL of freshly prepared NaBH₄ (0.5 mg mL⁻¹) was quickly added to the mixture. The mixture would turn into wine red rapidly, which indicated the generation of GNPs. The obtained solution was stored at 4 °C until use.

2.3 Preparation of SiO₂/GNPs

SiO₂/GNPs, the precursors of GNSs, were prepared as reported previously with some modifications.³⁸ Firstly, 39 mL of monodisperse silica colloidal spheres (0.013 g mL⁻¹ in ethanol) were amino-functionalized by mixing with 60 μL of APTES under vigorous stirring by magnetic bar at 70 °C for 3 h. In order to remove the excess reactants, the solution was centrifuged for 3 times and redispersed in ethanol. Then, 550 μL of pure APTES-functionalized silica spheres (0.043 g mL⁻¹) were redispersed in 15 mL of ethanol and added dropwise to 200 mL of as-prepared colloidal GNPs with vigorous stirring. The GNPs were attached onto the surface of silica nanoparticles to form SiO₂/GNPs nanocomposites. After centrifugation and removal of the supernatant for three times, pure SiO₂/GNPs were obtained and redispersed in H₂O. The peak absorbance (ca. 532 nm) of the obtained SiO₂/GNPs solution was adjusted to 0.5 for use. The prepared SiO₂/GNPs nanocomposites can be dispersed homogeneously in solvents such as water and ethanol for more than 1 month at 4 °C.

2.4 H₂O₂-mediated growth of SiO₂/GNPs

Firstly, 3.2 mL of H₂O was mixed with 0.5 mL of SiO₂/GNPs solution and 3 mL of as-prepared K–H growth solution for 2 min. Then, 100 μL of H₂O₂ was added under gently shaking. After 15 min, the UV/Visible absorption spectra of the resultant solution were performed on a Shimadzu UV3150 UV-vis-NIR spectrophotometer in absorbance mode. To study the growth process of GNSs, the concentration of H₂O₂ was varied from 0 μM to 400 μM. A Zeiss ULTRA-plus scanning electron microscope (SEM) was used to image the morphology of SiO₂/GNPs or GNSs at different stages of the shell forming operated at 15 kV.

2.5 H₂O₂ scavenging activity detection assay

Typically, 3 mL of H₂O was mixed with 200 μL of antioxidant solution and 100 μL 17 mM H₂O₂ (250 μM in the final reaction solution which was 6.8 mL). After incubating for 2 min, 0.5 mL of SiO₂/GNPs solution and 3 mL of as-prepared K–H growth solution were added and mixed under gently shaking at room temperature for 15 min. In order to remove the residual reactants, the solution was processed using centrifugation for 2 times (3000 rpm, 20 min) and the deposition was redispersed to 0.68 mL (The original solution was 6.8 mL before centrifuging.) A control experiment was tested by adding 200 μL of water instead of the antioxidant solution.

2.6 SERS measurements

In order to determine the SERS-activity of the GNSs, 0.5 mL 10 μM NBA was added into 0.5 mL GNSs solution and then incubated for half an hour before the Raman experiments. A Renishaw Invia Reflex system (equipped with a Peltier-cooled charge-coupled device (CCD) detectors and a Leica confocal microscope) was used to carry Raman experiments. Samples were excited with a 785 nm diode laser (∼0.6 mW) under linefocus mode and a grating of 1200 l mm⁻¹ was used. The exposure time was 10 s. The laser was focused onto the sample surface by using a 50× long working distance objective. All Raman experiments were carried out at room temperature (∼20 °C). Every Raman spectrum was averaged from 5 measurements. All experiments were performed in triplicate and values were averaged. Results were given as means ± the standard deviation (SD).

3. Result and discussion

3.1 H₂O₂-mediated growth of GNSs

It has been reported that many chemical reagents, e.g.formaldehyde, hydroxylamine-hydrochloride and sodium borohydride, could mediate the growth and generation of GNSs.³⁹ Our group has revealed that H₂O₂ could also reduce AuCl₄⁻ to Au⁰ and enlarge the GNPs on the SiO₂/GNPs surface.³⁸Scheme 1 outlines the main process of the GNSs formation. The ultrasmall GNPs are first attached onto the APTES-functionalized silica surface to form SiO₂/GNPs. The pre-adsorbed GNPs will act as nucleation sites for gold deposition. The gold is then selectively reduced onto these GNPs by H₂O₂, causing the GNPs seeds to grow larger and coalesce gradually to form gold island clusters or complete GNSs (Scheme 1SiO₂/GNPs to Complete GNSs).

Scheme 1 Schematic illustration of the H₂O₂-mediated GNSs formation process

Fig. 1 shows a set of SEM images of GNSs that is produced by different concentration of H₂O₂. It can be seen clearly that there were numerous GNPs (∼5 nm in diameter) on the APTES-functioned silica surface (Fig. 1a), making up the SiO₂/GNPs nanocomposites. The gold coverage of SiO₂/GNPs precursor was about 25% (obtained from SEM images). As the concentration of H₂O₂ increased, more AuCl₄⁻ was reduced onto the GNPs on the silica surface, the gaps/interstices between GNPs became narrower and finally coalesced (Fig. 1b–f). Once complete GNSs forms, the additional deposition of gold only caused the increase of the outer shell thickness and the decline in the core-radius/shell-thickness ratio of GNSs. The H₂O₂-mediated growth process was confirmed by the localized surface plasmon study, whose plasmon peak shifted from 532 nm to 760 nm as the concentration of H₂O₂ increased from 0 μM to 400 μM, as shown in Fig. 2.

Fig. 1 SEM images of GNSs produced by different concentrations of H₂O₂: a) 0 μM, b) 25 μM, c) 50 μM, d) 150 μM, e) 250 μM, f) 400 μM. The concentration of the H₂O₂ here means the concentration in the final reaction solutions (6.8 mL). The scale bar for the SEM images was 100 nm.

Fig. 2 Absorption spectra of the SiO₂/GNPs after incubating with different concentration of H₂O₂ for 15 min (from a to k: 0, 25, 50, 75, 100, 125, 150, 200, 250, 300 and 400 μM).

3.2 SERS-activity of GNSs with different enlargements

Concomitant with the increase of gold coverage on the silica cores, the SERS-activity of GNSs exhibited a pronounced change, which quantitatively correlated with the concentration of H₂O₂ (Fig. 3). In order to demonstrate the SERS-activity of the SiO₂/GNPs or GNSs, 5 μM Nile blue A sulfate (NBA) was chosen as Raman probe molecule. The higher the SERS intensity, the stronger the SERS-activity of the GNSs. Fig. 3a shows the SERS spectra of gold nanoshell solutions that were produced by different concentrations of H₂O₂. The spectra present the characteristic Raman shifts of NBA at 592 cm⁻¹ and 1638 cm⁻¹, whose modes were formed by positively charged nitrogen.⁴⁰ Because of the high specificity and the relative strong intensity in the spectrum, we took the 592 cm⁻¹ mode of NBA as reference mode for quantitative analysis and the SERS intensity referred in this work meant the intensity of the 592 cm⁻¹ mode unless specified. The plot of SERS intensity against the H₂O₂ concentration that was used to form GNSs is presented in Fig. 3b–c. As can be seen, the SERS intensity went up monotonically until it reached its peak at ∼42 000 counts when 250 μM H₂O₂ was used. However, if the concentration of H₂O₂ increased further (400 μM), the SERS intensity would decrease contrarily. The results indicated that the SERS intensity of Raman probe molecule correlated well with the concentration of H₂O₂ used to form gold nanoshell substrates in the range of 0–250 μM. Remarkably, even a slight variation of H₂O₂ (less than 5 μM) was enough to lead to a distinguishable change of the SERS intensity. Compared with the conventional method for measuring H₂O₂, which is based on the intrinsic absorbance value of H₂O₂ at 230 nm,^1,8,9 our SERS-based method was more sensitive. The effective test range for our method was 0–250 μM.

Fig. 3 a) SERS spectra of GNSs produced by different concentrations of H₂O₂: 25 μM; 50 μM; 150 μM and 250 μM from bottom upwards. b) and c): Plot of the SERS intensity of Nile blue A sulfate (NBA) against the concentration of H₂O₂. Each spectrum was averaged from 5 measurements. The samples were excited with a 785 nm laser and with 10 s exposure time. The concentration of the Raman probe molecule, NBA, was 5 μM.

3.3 Raman enhance effect of the gaps/interstices on the shell surface

Since the SERS phenomenon was discovered, many theories have been proposed for its explanation and two of them are currently used. One is called “electromagnetic enhancement” (EM) which was proposed by Albrecht and Creighton;³⁶ the other is named “charge transfer enhancement” or “chemical enhancement” which was proposed by Jeanmaire and Van Duyne.³⁷ From the point of view of EM theory, the SERS phenomenon depends intimately on the localized surface plasmons of metal nanostructures which are excited by the incident laser. The enhancement factor (EF) of GNSs with a core radius of r and an outer shell thickness of R can be approximatively estimated from the expression reported by Kerker et al.,^21,41

G = | (1 + 2g₀) (1 + 2g) |², (1)

where g₀ and g are polarizability factors of the nanoshell at the excitation frequency and the Raman scattering frequency, respectively; and g is given as:where ε₁, ε₂ and ε₃ are dielectric constants of the core, the outer layer metallic shell and the surrounding medium, respectively; q = r/(r + R). The g₀ has a similar expression. The calculated EFs of the GNSs were ca. 203, 185 and 145 (with 18 nm, 20 nm and 30 nm shell thickness on a 110 nm core, respectively). An experimental dielectric function of gold was used in the calculations.⁴² The excitation laser was 785 nm. The calculated EFs were for the 592 cm⁻¹ scattering mode. If the enhancement was entirely due to the local electromagnetic field, the EFs should decrease by 8.9% and 28.6% when the metal shell grows from 18 nm to 20 nm and 30 nm, respectively. However, the EF we obtained in the experiment increased by 113.3% and 40% from 18 nm to 20 nm and 30 nm shell thickness. This was not a big surprise as the Kerker's expression was based on a model with an ideal smooth surface. We suggested that the “hot spots” in the gaps/interstices on the shell surface contributed a lot to the total SERS enhancement. During the growth process of GNSs, the effect of the gaps/interstices on the enhancement should first increase as they become narrower which was preferred for the amplification of the local electromagnetic fields, then decreased because these “hot spots” area was sealed off by the further deposition of Au. This may be the main reason for the tendency of the SERS-activity changes (first increase and then decrease) in the growth process of GNSs.

3.4 Detection of H₂O₂ scavenging activity

After studying the correlation between the SERS-activity of GNSs and the concentration of H₂O₂ used, we determined the H₂O₂ scavenging activity of two antioxidants, tannic acid and L-apple acid, by restraining the H₂O₂-mediated formation of GNSs from SiO₂/GNPs. Briefly, the H₂O₂ used to form GNSs, with a final concentration of 250 μM, was first mixed with a certain amount of antioxidant before the forming process (Scheme 2a). Since the H₂O₂ was scavenged by the antioxidant, the growth of the SiO₂/GNPs would be inhibited, leading to a drop of the SERS-activity (Scheme 2b). By studying the change of the SERS-activity, we can estimate the H₂O₂ scavenging activity (SA_HP) of antioxidants. Briefly, in order to get the SA_HP value, the obtained SERS intensity was first presented as the percentage of SERS inhibition I_S–A by applying the following equation: I_S–A = ΔI/I₀ (%), I₀ was the difference of SERS intensities of the Raman probe molecule by applying the SiO₂/GNPs and the GNSs produced with 250 μM H₂O₂ as SERS substrates; ΔI was the difference of SERS intensities of Raman probe molecule by applying the GNSs produced with and without antioxidants as SERS substrates (Scheme 2b). The amount (concentration) of antioxidant to provoke 50% reduction in SERS-activity (SERS intensity here) was named IC₅₀ (when I_S–A = 50%). By plotting I_S–A against the concentration of antioxidant and then fitting it with an appropriate equation, we could get the IC₅₀ of the antioxidant approximately. Finally, the H₂O₂ scavenging activity, defined as SA_HP = 1/IC₅₀ and expressed as μM⁻¹, could be obtained.

Scheme 2 a) Schematic diagram of the methodology for detecting H₂O₂ scavenging activity: GNSs produced with [GNSs (I)] or without antioxidant [GNSs (II)]. b) SERS spectra of the Raman probe molecule by applying different substrates: SiO₂/GNPs, GNSs (I) and GNSs (II) from bottom upwards.

Fig. 4a shows the SERS spectra of different gold nanoshell solutions produced by 250 μM H₂O₂ together with different concentrations of L-apple acid. As can be seen, the SERS intensity decreased gradually as the concentration of L-apple acid increased, which indicated the inhibition of H₂O₂-mediated growth of GNSs. This was confirmed by Fig. 5, which shows the SEM images of GNSs produced by 250 μM H₂O₂ together with 50 μM (Fig. 5a) and 200 μM (Fig. 5b) L-apple acids. The relationship between the SERS intensity and the concentrations of L-apple acid is presented in Fig. 4b as triangular markers. A correlation between I_S–A and L-apple acid concentration is plotted as dots in Fig. 4b. The solid curve was a fitting I_S–A curve which can be done by many scientific softwares such as ORIGIN and MATHEMATICA (we found that the relationship between I_S–A and L-apple acid concentration fitted well with the exponential associate function). From the fitting curve, one can obtain the IC₅₀ value of the antioxidant and then the SA_HP value. Based on the detection protocol described above, the effective test concentration range for L-apple acid was 5–500 μM, the IC₅₀ value was 150 μM and the SA_HP value was 6.67 × 10⁻³ μM⁻¹. By the same token, we detected the H₂O₂ scavenging activity of tannic acid and we got the effective test concentration range for tannic acid as 5–200 μM, the IC₅₀ value as 8.2 μM and the SA_HP value as 122 × 10⁻³ μM⁻¹ (Fig. 3c and 3d; interestingly, the exponential associate function was also the best fitting function for I_S–A).

Fig. 4 a) and c): SERS spectra of GNSs produced by using 250 μM H₂O₂ together with a different concentration of antioxidants: a) L-apple acid: 0 μM, 50 μM, 300 μM, 500 μM from a to d, and c) tannic acid: 0 μM, 5 μM, 50 μM, 200 μM from a to d. b) and d): The SERS intensity of NBA (triangular marker) and I_S–A (dots) against the concentration of b) L-apple acid and d) tannic acid. The solid curves in b) and d) were the fitting curves. The concentration of the Raman probe molecule was 5 μM.

Fig. 5 SEM images of the GNSs restrained by a) 50 μM and b) 200 μM L-apple acid.

The results shown here indicated that tannic acid expressed superior H₂O₂ scavenging activity than L-apple acid. That was not a big surprise, because it has been reported that the H₂O₂ scavenging activity may have a relationship with the chemical structures and in particular the number of hydroxyl group.^7,43,44Fig. 6 shows the chemical structure of tannic acid and L-apple acid. Tannic acid has much more hydroxyl groups than that of L-apple acid. Both of them are naturally occurring antioxidants and have been universally used in daily foodstuff. Tannic acid is found in seeds and/or barks of many plants and L-apple acid is found in lots of vegetables and fruits. Ingestion of these antioxidants may keep the balance between antioxidants and oxidants in living organisms, and lower the probability of cardiovascular disease and cancer.⁴⁵ Frankel and Meyer have pointed out that a single assay is insufficient for the accurate assessment of multifunctional antioxidants and more than one method is needed to provide adequate information on evaluation, as different antioxidants respond with variable behavior in tests performed under different conditions.⁴⁶ Thus, the method reported here may serve as an additional means for assessing the activities of multifunctional antioxidants.

Fig. 6 Chemical structures of tannic acid and L-apple acid.

4. Conclusion

In conclusion, we have introduced the Raman spectroscopy technique to detect the H₂O₂ and H₂O₂ scavenging activity by applying SiO₂/GNPs as SERS nanoprobes. The H₂O₂ scavenging activities of tannic acid and L-apple acid were tested in our study. The results showed that tannic acid expressed superior H₂O₂ scavenging activity than L-apple acid. The sensitivity of the SERS-based method was less than 5 μM; furthermore, our method embraced the advantages of SERS which were extremely desirable for biological sample analysis. Additionally, we emphasized that the gaps/interstices on the GNSs surface played an important role in the SERS enhancement. Another purpose is to attract attention to the importance and the potential applications of SERS-activity change in the SERS probes/sensors design.

5. Acknowledgement

We gratefully acknowledge support from the National Natural Science Foundation of China (Grant 90923010), Ministry of Science & Technology of China (Grants 2007AA022007 and 2008DFA51180) and Specialized Research Fund for the Doctoral Program of Higher Education of China (Grant 20090092110026).

Notes and references

1. L. M. Magalhaes, M. A. Segundo, S. Reis and J. L. Lima, Anal. Chim. Acta, 2008, 613, 1–19.
2. L. G. Wood, P. G. Gibson and M. L. Garg, J. Sci. Food Agric., 2006, 86, 2057–2066.
3. N. Delanty and M. A. Dichter, Acta Neurol Scand., 1998, 98, 145–153.
4. H. Esterbauer, R. Schmidt and M. Hayn, Adv. Pharmacol., 1997, 38, 425–455.
5. K. H. Cheeseman, Lipid peroxidation and cancer. In: DNA and Free Radicals, (ed.: B. Halliwell and O. I. Aruoma), New York and London: Ellis Horwood, 1993, pp 109–144.
6. M. Pazdzioch-Czochra and A. Widenska, Anal. Chim. Acta, 2002, 452, 177–184.
7. A. Arnous, C. Petrakis, D. P. Makris and P. Kefalas, J. Pharmacol. Toxicol. Methods, 2002, 48, 171–177.
8. R. F. Beers and I. W. Sizer, J. Biol. Chem., 1952, 195, 133–140.
9. R. J. Ruch, S. J. Cheng and J. E. Klaunig, Carcinogenesis, 1989, 10, 1003–1008.
10. H. Jiang and H. X. Ju, Anal. Chem., 2007, 79, 6690–6696.
11. H. K. Lee, K. Lee, I. K. Kim and T. G. Park, Adv. Funct. Mater., 2009, 19, 1–7.
12. J. Wang, N. D. Zhou, Z. Q. Zhu, J. Y. Huang and G. X. Li, Anal. Bioanal. Chem., 2007, 388, 1199–1205.
13. S. H. Kim, B. Kim, V. K. Yadavalli and M. V. Pishko, Anal. Chem., 2005, 77, 6828–6833.
14. S. J. Oldenburg, R. D. Averitt, S. L. Westcott and N. J. Halas, Chem. Phys. Lett., 1998, 288, 243–247.
15. Y. Wang, W. P. Qian, Y. Tan and S. H. Ding, Biosens. Bioelectron., 2008, 23, 1166–1170.
16. S. H. Ding, W. P. Qian, Y. Tan and Y. Wang, Langmuir, 2006, 22, 7105–7108.
17. L. R. Hirsch, J. B. Jackson, A. Lee, N. J. Halas and J. L. West, Anal. Chem., 2003, 75, 2377–2381.
18. S. R. Sershen, S. L. Westcott, N. J. Halas and J. L. West, J. Biomed. Mater. Res., 2000, 51, 293–298.
19. D. P. O'Neal, L. R. Hirsch, N. J. Halas, J. D. Payne and J. L. West, Cancer Lett., 2004, 209, 171–176.
20. L. R. Hirsch, R. J. Stafford, J. A. Bankson, S. R. Sershen, B. Rivera, R. E. Price, J. D. Hazle, N. J. Halas and J. L. West, Proc. Natl. Acad. Sci. USA, 2003, 100, 13549–13554.
21. M. Kerker and C. G. Blatchford, Phys. Rev. B, 1982, 26, 4052–4063.
22. C. E. Talley, J. B. Jackson, C. Oubre, N. Grady, T. Huser, C. Hollars, S. Lane, P. Nordlander and N. J. Halas, Nano Lett., 2005, 5, 1569–1574.
23. S. Nie and S. R. Emory, Science, 1997, 275, 1102–1106.
24. K. Kneipp, H. Kneipp, I. Itzkan, R. R. Dasari and M. S. Feld, Chem. Rev., 1999, 99, 2957–2976.
25. C. McDonagh, C. S. Burke and D. D. MacCraith, Chem. Rev., 2008, 108, 400–422.
26. N. P. W. Pieczonka, G. Moula and R. F. Aroca, Langmuir, 2009, 19, 11261–11264.
27. R. S. Golightly, W. E. Doering and M. J. Natan, ACS Nano, 2009, 3, 2859–2869.
28. M. Lin, L. He, J. Awika, L. Yang, D. R. Ledoux, H. Li and A. Mustapha, J. Food Sci., 2008, 73, 129–134.
29. X. Qian, X. H. Peng, D. O. Ansari, Q. Yin-Geon, G. Z. Chen, D. M. Shin, L. Yang, A. N. Young, M. D. Wang and S. Nie, Nat. Biotechnol., 2008, 26, 83–90.
30. M. Culha, D. Stokes, L. R. Allain and T. Vo-Dinh, Anal. Chem., 2003, 75, 6196–6201.
31. K. E. Shafer-Peltier, C. L. Haynes, M. R. Gluksberg and R. P. Van Duyne, J. Am. Chem. Soc., 2003, 125, 588–593.
32. R. J. Dijkstra, W. J. J. M. Scheenen, N. Dam, E. W. Roubos and J. J. Ter Meulen, J. Neurosci. Methods, 2007, 159, 43–50.
33. C. S. Levin, B. G. Janesko, R. Bardhan, G. E. Scuseria, J. D. Hartgerink and N. J. Halas, Nano Lett., 2006, 6, 2617–2621.
34. K. Hyunhyub, S. Singamaneni and V. V. Tsukruk, Small, 2008, 4, 1576–1599.
35. M. Fleischmann, P. J. Hendra and A. J. McQuillan, Chem. Phys. Lett., 1974, 26, 163–166.
36. M. G. Albrecht and J. A. Creighton, J. Am. Chem. Soc., 1977, 99, 5215–5217.
37. D. L. Jeanmaire and R. P. Van Duyne, J. Electroanal. Chem., 1977, 84, 1–20.
38. F. Kong, H. Liu, J. Dong and W. Qian, Biosens. Bioelectron., 2010 DOI:10.1016/j.bios.2010.03.029.
39. B. E. Brinson, J. B. Lassiter, C. S. Levin, R. Bardhan, N. Mirin and N. J. Halas, Langmuir, 2008, 24, 14166–14171.
40. G. M. do Nascimento and M. L. A. Temperini, J. Raman Spectrosc., 2008, 39, 772–778.
41. M. Kerker, Acc. Chem. Res., 1984, 17, 271–277.
42. P. B. Johnson and R. W. Christy, Phys. Rev. B, 1972, 6, 4370–4379.
43. R. Pulido, L. Bravo and F. Saura-Calixto, J. Agric. Food Chem., 2000, 48, 3396–3402.
44. R. Scherer and H. T. Godoy, Food Chem., 2009, 112, 654–658.
45. R. L. Prior, X. L. Wu and K. Schaich, J. Agric. Food Chem., 2005, 53, 4290–4302.
46. E. N. Frankel and A. S. Meyer, J. Sci. Food Agric., 2000, 80, 1925–1941.

---