The strategy of two-scale interface enrichment for constructing ultrasensitive SERS substrates based on the coffee ring effect of AgNP@β-CD

Abstract

Herein we introduced a Raman technique based on the two-scale interface enrichment for the detection of aromatic compounds. Silver nanoparticles were fabricated, β-cyclodextrins (β-CD) were physically absorbed on the surface of the silver nanoparticles as capturing layers of aromatic compounds for enriching the analytes on the interface of the silver nanoparticles in solution. After the analyte was captured, a mixture of silver nanoparticles and aromatic compounds were dropped on silicon wafer to form the coffee ring, which comprises densely packed silver nanoparticles loaded with analytes on the interface of the silicon wafer. In light of this, analytes were able to enrich twice in one process. Furthermore, densely-packed substrate produced a large number of “hot-spots” to improve the SERS effect. The combination of pre-concentration of β-CD and the SERS effect of the coffee-ring enhanced the detection ability of SERS to aromatic compounds. Due to the special structure of β-CD, this strategy can also be utilized in the SERS detection of other analytes.

---

Introduction

Raman techniques have attracted extensive attention since their discovery.^1,2 The finding and application of surface-enhanced Raman scattering (SERS) phenomenon transcends weak signal limits during the process of collecting Raman spectra, which proves to be a great advantage for the Raman technique, enabling it to be a rapid and sensitive analytical method.^3–5 Thus the approach of SERS has been widely applied in many realms, ranging from fundamental research to studies of biomedical and environmental issues.^6–8 Several strategies based on novel materials have been found to possess the capacity to improve SERS efficiency. Recently, fundamental studies have focused on hot spots or nanogaps among two or more nanoparticles.^9,10 The hot spots are important elements to Raman enhancement, which are characteristic of small nanoparticles aggregating on super-enhancing interstitial sites and large fractal aggregating where the hot spots arise.^11,12

The drop coating deposition Raman (DCDR) technique, based on the coffee ring effect, has gradually been recognized as a potential technique of fabricating 3-D hot spots substrates.^13–15 Ishan Barman et al. demonstrated that the application of this technique in the selective detection of HbA1c.¹⁶ Xu et al. developed a sensor-based Au coffee effect to detect polycyclic aromatic hydrocarbons.¹⁷ However, the nanoparticles do not specifically bind analytes. Therefore, analytes would randomly deposit on the substrate. Some of the analytes would not distribute on the edge,¹⁸ rendering low sensitivity and uncertainty in the results.

Cyclodextrin (CD), due to its special structure consisting of a hydrophilic exterior surface and hydrophobic interior core, is an ideal host–guest complex and can be applied in many fields such as drug release, molecule assembly.^19–21 Several studies have reported its usefulness in the detection of aromatic compounds. Chen et al. discovered that β-CD-functionalized nanoparticles could detect aromatic isomers via a UV-vis technique. However, its low sensitivity prevented its application to trace detection.²²

In this article, we fabricate AgNPs coated with β-CD (AgNP@β-CD) to form the coffee ring effect as a rapid and sensitive SERS platform for aromatic compound detection. Benefiting from the pre-concentration of β-CD, the SERS platform caught aromatic compounds on the nano interface of AgNP@β-CD and the 3D effective hotspots of the coffee ring of AgNP@β-CD on the silica interface endowed the SERS platform with an extremely low detection ability. In addition, this method together with its derivative method can also be applied to the detection of other analytes, such as polypeptides, heavy metal ions, and noxious gas.

Materials and methods

O-Phenylenediamine, 2,3-diaminonaphthalene, 2,4-dihydroxybenzoic acid, silver nitrate (AgNO₃), sodium hydroxide (NaOH), and chloroauric acid tetrahydrate (HAuCl₄), beta-cyclodextrin and glucose were purchased from Sinopharm Chemical Reagent Co. Nile blue A was obtained from Sigma Aldrich (USA). All of the reagents were analytically pure. Deionized water was purified with the Milli-Q system (resistivity > 18 MΩ).

Raman spectra were obtained on the edge of the substrate using a Renishaw Invia microRaman spectrometer at room temperature (∼25 °C) using a 785 nm excitation laser. The laser power was 12 mW and the acquisition time was 10 seconds in all measurements. The microstructures of the AgNP@β-CD were measured by a Zeiss ULTRA plus scanning electron microscope (SEM) at 15 kV and a JEM-2100EX (JEOL) transmission electron microscope (TEM) operated at 200 kV. UV-vis absorption spectra were collected with a Shimadzu UV3600 UV-vis-NIR spectrum-photometer. FTIR spectra were acquired using a Bruker Optik GmbH TENSOR 27 FTIR microscope.

Synthesis of Ag NP@β-CD

The synthesis of monodisperse silver nanoparticles stabilized with β-CD was performed according to the reported literature.²³ Briefly, 10 mL aqueous NaOH (0.01 mol L⁻¹), 20 mL aqueous β-CD and 10 mL aqueous glucose (0.013 mol L⁻¹) solutions were mixed at a constant speed and heated in a water bath. As soon as the solution reached 60 °C, 15 mL AgNO₃ (0.01 mol L⁻¹) aqueous solution was added. Finally, the solution was centrifuged three times at 8000 rpm for 10 min and then the supernatant was removed. The colloid solution was stored in a refrigerator at 4 °C for further analyses.

Substrate preparation

Silicon wafers were cut into 1 × 1 cm pieces. Hydrophobic and hydrophilic silicon wafer processing steps were as follows: in the hydrophilic treatment, the cleaned silicon wafer was heated in the mixed solution (H₂O₂ : NH₃H₂O : H₂O = 1 : 1 : 5 v/v/v) with a water bath at 60 °C for 30 min. In the hydrophobic treatment, the silicon wafer was soaked in 1% HF solution for 20 min to remove the surface oxide layer.

The SERS analyses

Diverse concentrations of analyte samples (1 μL) were mixed with AgNP@β-CD (9 μL) respectively. Then we extracted 1 μL of the mixture, dripped it onto the treated silicon wafer and dried it at normal temperature. The SERS signal was obtained on the edge of the ring deposition after evaporation of the nanoparticle suspension (Fig. S1†) (Scheme 1).

Scheme 1 Illustration of the self-assembly of the AgNP@β-CD via the coffee-ring effect and the detection of analyte.

Results and discussion

Characterization of the AgNP@β-CD

The absorption peak of AgNP@β-CD is at 404 nm in the UV-vis spectrum and there are no extra peaks in other wavelength regions (Fig. 1A). It is the typical absorption peak of the AgNP and the existence of β-CD has no impact on the peak position. The solution appearing as a transparent bright yellow indicated that the AgNP@β-CD is monodisperse in solution. FTIR spectra of the pure β-CD and AgNP@β-CD (Fig. 1B) explored the interactions between the β-CD and AgNP@β-CD. The band at 3605 cm⁻¹ of the β-CD molecule is assigned to the vibration of –OH, and the peak at 3578 cm⁻¹ in AgNP@β-CD is associated with –OH stretching. The band at 2920 cm⁻¹ appearing in two spectra is assigned to C–H stretching + C–H (CH₂). The band at 1140 cm⁻¹ (β-CD) and the band at 1145 cm⁻¹ (AgNP@β-CD) are associated with the bending of C–O–C. The FITR spectra confirmed that β-CD was successfully coated on to the surface of the silver particles.

Fig. 1 UV-vis spectra of AgNP@β-CD suspension (A) and FTIR spectra of β-CD and AgNP@β-CD nanoparticles (B).

The formation of the substrate

The SEM image (Fig. 2A) demonstrates that the average size of AgNP@β-CD was ca. 30 nm. As shown in Fig. 2B, the TEM image of AgNP@β-CD reveals that the surface of AgNP was successfully modified with a 2 nm thick β-CD. The morphology of the AgNPs was approximately a smooth sphere. The uniform and small sphere is able to move and reach the edge of the droplet easily. As shown in Fig. 2C, the ring width is ca. 15 μm and the ring diameter is 1.3 mm and most of the AgNPs@β-CD are deposited at the edge of the ring. The overall image clearly shows the pattern of the AgNP@β-CD depositing on the hydrophobic treated surface. A three phase contact line (TCL) pinning is one of the crucial conditions for formation of the coffee ring. The AgNP@β-CD was carried toward the TCL to form a ring deposition on the hydrophobic silicon wafer. After a drop of the mixture was dripped on to the hydrophilic surface (Fig. S2†), the TCL slid and led to the droplet spilling and forming an irregular deposition. To verify whether the AgNP@β-CD was totally deposited on the rim, an elemental mapping image (Fig. 2D) was performed. The yellow region represents the Ag element, which indicates that the AgNP@β-CD successfully reached the edge and formed a densely packed substrate.

Fig. 2 SEM images (A) and TEM images (B) of the AgNP@β-CD nanoparticles; SEM images of the substrate formed by the coffee-ring effect (C) and EDS analysis image (D).

The superiority of the coffee-ring pattern

In order to validate the superiority of the substrate, a series of patterns were deposited. We chose the OPD molecule as representative in this section. All substrates contained 5 μM OPD respectively. As shown in Fig. 3, AgNPs were dripped in to two patterns (Fig. 3A). Pattern A represented the material deposited via the coffee-ring effect; pattern B represented the material uniformly deposited. No matter what pattern was fabricated, no significant signals appeared. However, we found that the high peak is at 1270 cm⁻¹ after mixing with AgNP@β-CD. The result showed that the β-CD molecules played a key role in detecting the OPD. It specifically captured OPD molecules on the surface of the AgNPs. The OPD molecules were enriched on the interface of the materials and the solution. Analytes were found near the surface of the nanoparticles in large numbers, which produced a higher signal. In this protocol, β-CD molecule is a bi-functional reagent which plays an effective role in protecting agents, making the silver particles stable. It was utilized specifically to capture analytes on the surface of materials. Hence the analytes would not distribute on the substrate randomly. Instead, they would be carried to the edge of the ring and thus achieve highly sensitive detection results. Furthermore, under the coffee-ring effect, the signal was greater than the pattern which uniformly deposited on the silicon wafer (Fig. 3B). This result was attributed to the pre-concentration of the OPD molecules on the interface of the silicon wafer and the densely assembled substrate increased the number of SERS “hot spots”, leading to a signal enhancement thus rendering a high sensitivity to our platform. Twice enrichment processes occurred in one-step and the results showed a higher sensitivity compared to traditional methods.

Fig. 3 The SERS spectra of AgNPs (A) and AgNP@β-CD (B) substrate. All substrates contained 5 μM OPD. The characteristic peak at 1270 cm⁻¹. The spectra were recorded at 785 nm laser.

The stability of the platform

To study the stability of this platform in practical applications, the durability of the materials and the stability of the signal acquisition must be taken into careful consideration. To begin with, we evaluated the durability of the materials by UV-vis spectrometry (Fig. S3†). Spectra of AgNP@β-CD were acquired at different times, the absorption peak just slightly decreases with increasing time. The results depicted the durability of the materials. The β-CD layer was used to capture nanoparticles on the surface to prevent oxidation of the silver nanoparticles. At last, we dripped five droplets containing 10⁻⁵ NBA and AgNPs, respectively, and formed the substrate. The spectrum signal was acquired randomly on the edge of each ring. As shown in Fig. S4,† the signal strength was nearly consistent in each measurement. These results showed that the coffee ring pattern is reproducible.

The detection of aromatic compounds

Aromatic compounds are widely used in the dye, pesticide and medicine industry and recently in industrial synthesis. However, they are potentially hazardous to the environment. In our work, we select several typical aromatic compounds which contain OPD, 2,3-diaminonaphthalene and 2,4-dihydroxybenzoic acid. The droplet of mixture containing AgNP@β-CD and aromatic compounds rapidly evaporated at room temperature. The whole process of acquiring the SERS spectra is quite time-saving, taking less than 20 minutes. Two weak and broad peaks of the substrate would not affect the characteristic peak of aromatic compounds (Fig. S5†).

In order to certify the quantification of our platform, different concentration solutions were measured. The SERS spectra of aromatic compounds with different concentrations and calibration curves are shown in Fig. 4. We selected their characteristic peak, respectively (OPD: 1270 cm⁻¹; 2,3-diaminonaphthalene: 1743 cm⁻¹; 2,4-dihydroxybenzoic acid: 828 cm⁻¹). In the process of acquiring spectra, these characteristic peaks showed good stability in pre-experiment and the intensity of the peaks would vary (according to the concentration of the sample). It clearly showed that the intensity of the characteristic peak was influenced by the concentration of solution. The results demonstrated that this platform could be applied to detect aromatic compounds quantitatively.

Fig. 4 Representative SERS spectra with different concentrations ((A): OPD, (C): 2,3-diaminonaphthalene, (E): 2,4-dihydroxybenzoic acid) and the corresponding calibration curves (B, D, F).

To explore the detection ability of this platform when applied to different analytes, we acquired their detection limits. The detection limit of this platform is disparate for the different analytes (OPD: 100 pM; 2,3-diaminonaphthalene: 1 nM; 2,4-dihydroxybenzoic acid: 100 nM) under the optimized conditions. It relied on the combining capacity between the analytes and β-CD. All the analytes contained hydrophobic groups and hydrophilic groups. The results depicted that the combining capacity of OPD and β-CD is better than the others, due to the OPD structure which contains a hydrophobic benzene ring and two hydrophilic amino groups, which is conducive to its combination with β-CD. Moreover, we also selected other molecules as target analytes. As shown in Fig. S4,† when the object detected is nonylphenol, there are no significant Raman signals. We presumed that the hydrophobic long chain of nonylphenol hindered its combination with β-CD. Similarly, the SERS signal of bisphenol A is weak. The results show the AgNP@β-CD probes are specific for the analytes. Only when the structures of the analytes are suitable for the β-CD, the layer of β-CD on AgNP will capture them and their Raman spectrum will be acquired.

The SERS enhancement

In order to evaluate the superiority of the twice enrichment method, the SERS enhancement factor (EF) is an important parameter. To interrogate the SERS enhancement of the substrate, we calculated the Raman enhancement. The results are shown below:

EF = I_SERSC_Raman/I_RamanC_SERS

I and C represent the peak intensity and the concentration of the probe molecule, respectively. We calculate the EF value (EF = 2 × 10⁹), indicating a high SERS enhancement effect. The high EF value benefited from the special structure of the coffee-ring and the pre-concentration of the probe molecule. In addition, the results are not in a simple linear relationship with the level of enrichment. This phenomenon does not contribute solely to analyte enrichment. Nanoparticles aggregating in a narrow area can be a contributing factor, which leads to an increase in the large number of hot-spots. The hot-spots produced further improve the SERS enhancement. Higher packing densities contributed to higher densities of hot spot regions of the substrate displaying a greater overall SERS enhancement factor.

To further confirm the feasibility of this method in practical applications, as shown in Fig. 5, we tested the samples and detected the substrate background signal in DI water. No evident signal was detected in tap water. However, when aromatic compounds were added to the tap water, characteristic peaks were detected after interrogating these samples. It demonstrated the enormous potential of this method for the analysis of real water samples.

Fig. 5 The SERS spectra of aromatic compounds in practical application: deionized water (black), tap water (red), tap water spiked with 0.5 μM OPD (blue), 20 nM 2,3-diaminonaphthalene (pink) and 2 μM 2,4-dihydroxybenzoic acid (green).

Conclusion

In conclusion, we have presented stable AgNP modified with β-CD captured aromatic molecules and formed a robust SERS platform based on the coffee ring effect. This protocol promoted the separation and pre-concentration of the analytes in one step. Due to the special structure of the material and the enrichment of the analyte, trace amounts of aniline and other aromatic compounds could be detected via this SERS platform. This method can be effectively applied for the detection of trace pollutants. Moreover, the advantages of CD widen the application range of this technology, which can also be used for other analytes.

Acknowledgements

We gratefully acknowledge support from the Chinese 973 Project (Grant: 2012CB933302), the National Natural Science Foundation of China (Grant: 21175022), the Ministry of Science & Technology of China (Grant: 2012AA022703 and 2015AAO20502).

References

1. L. B. Yang, P. Li, H. L. Liu, X. H. Tang and J. H. Liu, Chem. Soc. Rev., 2015, 44, 2837–2848.
2. J. H. Ko, C. K. Lee and J. B. Choo, J. Hazard. Mater., 2015, 285, 11–17.
3. W. C. Soon, H. C. Chin, H. C. Chi, X. C. Siew, Z. Sarani, S. S. Mohd and M. H. Nay, RSC Adv., 2015, 5, 88915–88920.
4. S. Su, C. Zhang, H. Li, J. Chao, X. L. Zuo, X. F. Liu, C. Y. Song, C. H. Fan and L. H. Wang, ACS Appl. Mater. Interfaces, 2014, 6, 18735–18741.
5. J. Dong, M. D. Guo, W. Xie, Y. Li, M. Y. Zhang and W. P. Qian, Analyst, 2015, 140, 2741–2746.
6. K. Kim, K. L. Kim and K. S. Shin, Analyst, 2012, 137, 3836–3840.
7. S. M. Tabakman, Z. Chen, H. S. Casalongue, H. L. Wang and H. J. Dai, Small, 2011, 7, 499–505.
8. R. A. Halvorson and P. J. Vikesl, Environ. Sci. Technol., 2011, 45, 5644–5651.
9. J. Yan, S. Su, S. J. He, Y. He, B. Zhao, D. F. Wang, H. L. Zhang, Q. Huang, S. P. Song and C. H. Fan, Anal. Chem., 2012, 84, 9139–9145.
10. H.-Y. Chen, M.-H. Lin, C.-Y. Wang, Y.-M. Chang and S. Gwo, J. Am. Chem. Soc., 2015, 137, 13698–13705.
11. J. P. Camden, J. A. Dieringer, Y. M. Wang, D. J. Masiello, L. D. Marks, G. C. Schatz and R. P. Van Duyne, J. Am. Chem. Soc., 2008, 130, 12616–12617.
12. S. J. Lee, A. R. Morrill and M. Moskovits, J. Am. Chem. Soc., 2006, 128, 2200–2201.
13. R. D. Deegan, O. Bakajin, T. F. Dupont, G. Huber, S. R. Nagel and T. A. Witten, Nature, 1997, 389, 827.
14. P. J. Yunker, T. Still, M. A. Lohr and A. G. Yodh, Nature, 2011, 476, 308.
15. R. A. Halvorson and P. J. Vikesl, Environ. Sci. Technol., 2011, 45, 5644–5651.
16. I. Barman, N. C. Dingari, J. W. Kang, G. L. Horowitz, R. R. Dasari and M. S. Feld, Anal. Chem., 2012, 84, 2474–2482.
17. J. W. Xu, J. J. Du, C. Y. Jing, Y. L. Zhang and J. L. Cui, ACS Appl. Mater. Interfaces, 2014, 6, 6891–6897.
18. R. D. Venkateshwar and G. B. Madivala, RSC Adv., 2015, 5, 60079–60084.
19. K. Uekama, F. Hirayama and T. Irie, Chem. Rev., 1998, 98, 2045–2076.
20. M. N. Roy, M. C. Roy and K. Roy, RSC Adv., 2015, 5, 56717–56723.
21. Q. Li, Y. Zhang, Y. Jin, Q. B. Yang, J. S. Du and Y. X. Li, RSC Adv., 2015, 5, 68815–68821.
22. X. Chen, S. G. Parker, G. Zou, W. Su and Q. J. Zhang, ACS Nano, 2010, 4, 6387–6394.
23. P. F. Andrade, A. F. D. Faria, D. S. D. Silva, J. A. Bonacin and M. D. C. Gonc, Colloids Surf., B, 2014, 118, 289–297.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra01101b

---