Design and preparation of a recyclable microfluidic SERS chip with integrated Au@Ag/TiO₂ NTs

Abstract

In the present paper, we report on the design and fabrication of a recyclable microfluidic SERS chip with an integrated Au@Ag/TiO₂ NTs (gold core silver shell nanoparticles (NPs) with decorated TiO₂ nanotube arrays (NTs)) composite SERS substrate. The stripped TiO₂ NTs with regularly arranged pore structures were prepared on a Ti foil by easy tape lithography followed by anodic oxidation. The Au@Ag NPs were fabricated on the mouth surface and the inner walls of the TiO₂ NTs within a microchannel without clogging via a facile self-assembly chemical plating composite method. Under optimized conditions, the microfluidic SERS chip provided excellent sensitivity to R6G molecules with a detection limit of 10⁻¹⁰ M and a SERS enhancement factor (EF) of 1.15 × 10⁸. Versus single Au NPs, Au@Ag NPs and the Au@Ag/TiO₂ plate, the optimized Au@Ag/TiO₂ NTs showed improved SERS peak intensities for the C–C–C bending (613 cm⁻¹) of 20, 2.3 and 1.6 times enhancement, respectively. The enhancement effect was attributed to both electromagnetic and chemical enhancement. The photocatalytic performance of the Au@Ag/TiO₂ NTs was exploited to recycle the microfluidic SERS chip through UV light photocatalytic purification. The experimental result shows that the microfluidic SERS chip featured high reproducibility and could be used as an efficient SERS detector for many biochemical molecules.

---

1 Introduction

Surface enhanced Raman spectroscopy (SERS) has attracted the attention of many researchers because it offers fast detection and high sensitivity.^1–4 However, in the open space, sample molecules are often inhomogeneously distributed on the surface of the SERS enhancement substrate.^5,6 Accordingly, the SERS detection reproducibility is often poor. In addition, sample molecules on the surface of the SERS substrate can be easily contaminated, and this can interfere with spectral analysis.⁷ Recently, embedding SERS substrates into a microfluidic chip can make SERS detection more reproducible, efficient, environmentally friendly and safer.^8–13

The choice of SERS substrate is an important factor that affects the SERS performance. Typical SERS-active substrates are composed of noble metals (Au, Ag, Cu, Pt) in the form of zero- (0D), one- (1D) and two-dimensional (2D) nanostructures, taking advantage of the electromagnetic (EM) enhancement mechanism.¹⁴ Among the various materials, Ag NPs often have the best SERS activity,¹⁵ albeit, Au NPs have many advantages over Ag NPs, including long-term stability, homogeneity and biocompatibility.¹⁶ Consequently, bimetallic gold silver core–shell nanoparticles (Au@Ag NPs) can be selected as a SERS substrate and can offer stability and high SERS activity.^17–20 However, metal SERS substrates generally cannot be reused because of the difficulty of removing residual molecules from the substrate surface, which is consequently inefficient. In addition to noble metals, semiconductors and transition metal oxides have emerged as potential SERS-active substrates.²¹ ZnO nanocrystals,²² NiO,²³ Cu₂O²⁴ and TiO₂ ²⁵ flat surfaces and nanostructures exhibit SERS enhancement on the metal oxide side in the absence of plasmonic enhancement via a charge transfer mechanism (CT).^26–28 In particular, TiO₂ has been extensively used in degrading organic contaminants because of its high photocatalytic activity, nontoxicity and stability.^29–33 By combining TiO₂ three-dimensional (3D) nanostructures (such as TiO₂ nanotubes, NTs) and Ag/Au NPs, it is possible to obtain a synergic enhancement effect of EM and CT while at the same time exploiting the relatively high specific surface area to increase the loading of metal NPs and to yield many SERS-active “hot spots”, which can provide a huge Raman enhancement.^34,35 After testing, the substrates can be purified via a subsequent UV irradiation process and can be reused for several SERS measurements.

Fabricating 3D tubular nanostructure SERS substrates is challenging by metal deposition. However, recently, some techniques have been reported for their fabrication, such as photochemical reduction,³⁶ electrochemical deposition,³⁷ ultrasonic assistance reduction,³⁸ microwave-assisted method³⁹ and the sputter deposition method.⁴⁰ However, most of these methods lead to clogging of the nanoscale cylindrical pores, which can completely cover the porous layer with a large amount of metal NPs. This ultimately decreases their SERS efficiency because of their low specific density of “hot spots”. Recently, Yajie Chen et al. reported a simple but cost-effective method for the in situ seed growth of Au NPs on the mouth surface and the inside of the TiO₂ NTs without clogging.⁴¹ This in situ growth of Au NPs used electroless-deposited Au NPs as seeds for the catalytic reduction of additional gold to grow Au NPs inside the TiO₂ NTs. However, the challenge of precisely controlling the morphologies of Au NPs inside the TiO₂ NTs remains due to the uncontrollability of the seed size, shape and distribution density. To solve these problems, a self-assembly chemical plating composite method could be applied. This approach uses a chemical self-assembly method to controllably prepare orderly metal NP seeds inside TiO₂ NTs within a microchannel. This is followed by controllable growth of the metal seeds, so it represents a chemical plating method. Currently, this method has been used to controllably prepare Au NPs on a glass surface and silver film surface.^42,43 However, there are no reports on preparing Au@Ag NPs on the mouth surface and the inner walls of 3D TiO₂ NTs via this self-assembly chemical plating composite method.

Here, a novel recyclable microfluidic SERS chip with integrated Au@Ag/TiO₂ NTs composite SERS substrate was designed, and Au@Ag NPs were immobilized on the mouth surface and the inner walls of the TiO₂ NTs within a microchannel by using a facile, controlled self-assembly chemical plating composite method. Monodisperse Au NPs were prepared and the thickness of the Au core was controlled via the PDDA concentration. The thickness of the silver shell was controlled by the plating time. Using R6G as a sample analyte, the factors impacting the SERS signal were studied and optimized. The synergistic SERS enhancement effects of Au, Ag TiO₂ NTs were studied as well as the SERS performance of the microfluidic SERS chip. Finally, under UV light irradiation, the change process of the SERS signal intensity of the adsorbed R6G molecules with irradiation time was observed. The recyclability of the microfluidic SERS chip was also evaluated. We aimed to obtain a recyclable microfluidic SERS chip with high sensitivity and good reproducibility.

2 Experimental

2.1 Chemical and materials

Deionized water (DI) was used for all the aqueous solutions. Ti foil (99.9%) was purchased from Beijing Xinruige Material Co. Ltd, hydrogen tetrachloroauric acid (HAuCl₄·4H₂O, AR) was purchased from Sinopharm Chemical Reagent Co. Ltd, silver nitrate (AgNO₃, AR) from Aladdin Industrial Corporation, hydroxylamine hydrochloride (NH₂OH·HCl, AR) from Chengdu KeLong Chemical Reagent Co. Ltd, poly dimethyl diallyl ammonium chloride (PDDA, wt 35%, AR) from Heowns Biochemical Technology Co. Ltd, Rhodamine 6G (R6G, AR) from Sigma-Aldrich Co. LLC, the SU-8 photoresist and developer from Microchem Crop and poly(dimethylsiloxane) (PDMS) from Dow Corning Corp (Sylgard 184). And all the other regents were of analytical regent (AR) grade and purchased from Chongqing Chuandong Chemical Group.

Gold colloid was prepared by using the method developed by Frens.⁴⁴ Briefly, 100 ml of 0.01% HAuCl₄·4H₂O was brought to a vigorous boil while stirring, then 8 ml of 1% trisodium citrate was rapidly added to the solution. After 15 min, the prepared gold colloid was cooled to room temperature and kept at 4 °C until use.

2.2 Preparation of a microfluidic SERS chip with integrated Au@Ag/TiO₂ NTs

2.2.1 Preparation of a microfluidic chip with integrated TiO₂ NTs. The “sandwich” microfluidic chip consisted of three layers, as shown in Fig. 1(a), namely a Ti foil with stripped TiO₂ NTs as the bottom layer, a channel layer made of PDMS in the middle and a cover layer made of PDMS at the top. Stripped TiO₂ NTs were prepared on a Ti foil via easy tape lithography followed by anodic oxidation. Polyimide adhesive tape was manually aligned and bonded to the clean Ti foil to protect the Ti surface from oxidation (NTs growing area = 10 × 0.3 mm²). TiO₂ NTs were grown by anodic oxidation of the Ti foils with polyimide adhesive tape in an electrolyte solution made of ammonium fluoride (NH₄F, 98%, 10 g) and deionized water (500 ml) and glycerin (99.5%, 500 ml).⁴⁵ Briefly, the anodization was performed in a potentiostatic configuration by applying 30 V for 1 h in a two-electrode electrochemical cell. After anodization, the polyimide adhesive tape was removed from the Ti foils, and the samples were rinsed with DI water and dried in air. The as-grown porous anodic layers had poor adhesion to the Ti substrate but were easily removed, even by a gentle touch, and so were annealed at 450 °C for 2 h in air to crystallize them into the anatase phase. Concerning the fabrication of the microfluidic device, the PDMS microchannel layer was fabricated through soft-lithography technology. Briefly, a 300 μm thickness SU-8 photoresist was spin-coated and pre-baked for 30 min to evaporate the organic solvent. By using UV-lithography, a 300 μm thickness, 300 μm wide and 1 cm long template was successfully prepared. Then, the polydimethylsiloxane (PDMS) prepolymer and curing agent were thoroughly mixed in a 10 : 1 weight ratio. Subsequently, the mixture was poured onto the surface of the pre-prepared template and cured for 1 h at 90 °C. When the PDMS was dried, it was peeled off from the template and the PDMS channel was derived. The PDMS cover plate with a thickness of 0.4 mm was fabricated in the same way. Holes were punched through the cover plate at both ends of the channel, and were utilized as the inlet and outlet. The PDMS channel layer was manually aligned and bonded onto the Ti foil with stripped TiO₂ NTs by means of “stamp and stick” bonding.⁴⁶ In detail, a thin layer of the PDMS mixture was spun on a glass slice and then selectively transferred to the PDMS channel layer using a stamping process. The PDMS channel layer was then bonded to the Ti foil with stripped TiO₂ NTs by thermal treatment at 90 °C for 30 min. Finally, the PDMS cover plate was manually bonded on the PDMS channel layer to form a “sandwich” microfluidic chip with TiO₂ NTs.

Fig. 1 (a) Components of the “sandwich” microfluidic chip with integrated TiO₂ NTs; (b) construction of the microfluidic SERS chip; (c) cross-section of the microfluidic channel with an integrated Au@Ag/TiO₂ NTs SERS substrate.

2.2.2 In situ fabrication of Au@Ag NPs in the microchannels with TiO₂ NTs. Au@Ag NPs were in situ immobilized on the mouth surface and the pore walls of the TiO₂ NTs within microchannels by the self-assembly chemical plating composite method. First, PDDA solutions with concentrations of 0.5%, 1%, 1.5%, 2%, 4% (wt%) were injected into the microchannels. After keeping for 1 h and then thoroughly rinsing with DI water, the prepared gold colloid was injected into the microchannels and kept for 6 h. During the process, the gold colloid in the microchannels was substituted by fresh gold colloid every 30 min. After that, the chemical silver plating solution (content of 0.1 ml 0.099 M of hydroxylamine hydrochloride and 39.9 ml 0.2 mM of AgNO₃) was passed through the microchannels at a rate of 20 μl min and the reaction time was 0 min, 5 min, 10 min, 20 min, 30 min, 60 min, respectively. Finally, microfluidic SERS chips with integrated Au@Ag/TiO₂ NTs were obtained.

2.3 Characterization

The surface morphology and microstructure of the substrates were characterized with field emission scanning electron microscopy (FESEM; FEI Nova Nano SEM 400 operated at 10.0 kV) and transmission electron microscopy (TEM, Hitachi-7800). The elemental composition of the Au@Ag/TiO₂ NTs were confirmed by energy-dispersive X-ray spectroscopy (EDS, Oxford Instruments X-MaxN). The Raman spectra were collected using a Raman spectrometer (LabRAM HR Evolution, HORIBA Jobin Yvon, France) with the 633 nm line of a He–Ne laser as the excitation source. The laser power at the sample was set to 4.5 mW.

2.4 SERS measurements for R6G in the microfluidic SERS chips

To evaluate the SERS performance of the microfluidic SERS chip with integrated Au@Ag/TiO₂ NTs, R6G was selected as the test sample. Here, 10⁻¹⁰ M, 10⁻⁹ M, 10⁻⁸ M, 10⁻⁷ M, 10⁻⁶ M and 10⁻⁵ M of R6G solutions were injected into the SERS microchannels for 30 min, and then, the microchannels were thoroughly washed and filled with deionized water. The Raman scattering spectra were collected through a 50× microscopy objective lens (NA = 0.5) with a focal spot about 1 μm in diameter. The Raman mapping image was acquired from the surface of the Au@Ag/TiO₂ NTs substrate. The incident laser power was kept constant at 0.425 mW with total accumulation times of 1 s.

2.5 In situ photodegradation and SERS measurements

The photodegradation experiments of R6G were performed with the prepared SERS chip as follows: 10⁻⁵ M R6G solution was injected into the SERS microchannel for 30 min, and then, the microchannel was thoroughly washed and filled with water. A Philips UV lamp (25 W, maximum at 365 nm) was used as the UV light source. The vertical distance between the lamp and the microfluidic SERS chip was 5 cm. The SERS spectra of the adsorbed R6G were recorded every 10 min during UV irradiation. When no Raman peak could be observed, the microchannel was thoroughly rinsed in deionized water, and the above operations were repeated for 6 cycles. In order to avoid sample damage being caused by the laser, the laser power at the sample was reduced to 0.425 mW and the acquisition time was reduced to 1 s.

3 Result and discussion

3.1 Design and fabrication of the microfluidic SERS chip with integrated Au@Ag/TiO₂ NTs

In order to realize the reuse and to improve the preparation flexibility of the microfluidic SERS chip, TiO₂ was chosen as the substrate material due to its high photocatalytic activity and Raman enhancement effect.^44–48 Meanwhile, a TiO₂–PDMS–PDMS sandwich microfluidic chip was designed. By utilizing the excellent plasma characteristic of gold and silver – and the good stability, biocompatibility and oxidation resistance of nano-gold as well as the large specific surface area of the 3D NTs structure – an integrated Au@Ag/TiO₂ NTs composite nanostructure was designed. The TiO₂ NTs were prepared on a Ti foil via easy tape lithography followed by anodic oxidation. The surface of the TiO₂ NTs was modified with PDDA. The Au NPs were then assembled on the surface of the PDDA-modified TiO₂ NTs via electrostatic adsorption between PDDA and Au NPs. The Fermi level of Au was higher than the conduction band of TiO₂. Once the Au NPs were formed, electrons in the conduction band shift to the Au NPs and can store charge overwhelmingly since the Helmholtz capacitance of the metal–solution interface is much higher than that of the semiconductor–solution interface.⁴⁷

e_(CB)⁻ → e_Au⁻ (1)

As a result, when Au/TiO₂ NTs are used as precursors, silver ions in the solution obtain electrons only from Au NPs and thus locate on the Au surface to form Au@Ag NPs instead of producing new Ag NPs on TiO₂ NTs. The thickness of the Ag shell would then continue growing until the silver ions are depleted in solution.

e_(Au)⁻ + Ag⁻ → Au@Ag (2)

It is well known that the electromagnetic enhancement effect of metal NPs depends on their size and distribution density.^48,49 For systems with the same particle size, the electromagnetic coupling effect of metal NPs increase with the increasing distribution density of metal NPs. Therefore, the density of Au@Ag NPs distributed on the surface of TiO₂ NTs must be controlled by controlling the density of the Au core. In this experiment, the distribution density of the Au core on the surface of TiO₂ NTs depended on the PDDA concentration. Furthermore, under the same test conditions, the SERS enhancement effect of the different Au/TiO₂ NTs prepared with different concentrations of PDDA were compared using 10⁻⁵ M R6G as the test sample (Fig. 2). We found that the SERS signal intensity of R6G increased with the increasing concentrations of PDDA and reached a maximum at 2%. There was then a gradual decreased when the PDDA concentration was greater than 2%. This could be explained as follows: when the PDDA concentration was 0.5–2%, the density of Au NPs on TiO₂ NTs increased, and the interparticle electromagnetic coupling improved with the increasing concentrations of PDDA. When the PDDA concentration was 2%, the density of Au NPs on the surface of TiO₂ NTs reached the maximum, and the synergistic Raman enhancement effect of the Au NPs and TiO₂ NTs was optimized. When the PDDA concentration was greater than 2%, the density of the Au NPs remained constant due to the strong electrostatic repulsion between adjacent NPs. With further increasing the concentration of PDDA, the distance between the sample molecules and the surface of the TiO₂ NTs increased due to also increasing the thickness of the adsorbed PDDA layer on the surface of TiO₂ NTs. Accordingly, the Raman enhancement effect of TiO₂ gradually decays. Therefore, the optimal PDDA concentration was 2%.

Fig. 2 SERS spectra of 10⁻⁵ M R6G adsorbed on Au/TiO₂ NTs prepared by using different concentrations of PDDA: (a) 0.5%; (b) 1%; (c) 1.5%; (d) 2%; (e) 4%.

With the same size of Au core, the SERS enhancement effect of Au@Ag NPs was not only associated with the distribution density of the NPs, but also with the thickness of the silver shell.^50,51 The thickness of the silver shell was dependent on the chemical silver plating time. The SEM characterization results of Au@Ag/TiO₂ NTs with different chemical silver plating times are shown in Fig. 3(a)–(e). It was found that TiO₂ NTs had regularly arranged pore structures with 100 nm pore diameter. Before chemical plating, 18 nm diameter spherical gold NPs were evenly distributed on the mouth surface and on the inside of the TiO₂ NTs (Fig. 3(a)). The diameters of the NPs then increased from 21 nm to 25 nm, 28 nm and 36 nm as the plating time increased from 10 min to 20 min, 30 min and 60 min (Fig. 3(b)–(e)). The results involved the study of over 200 NPs. To study the core–shell structure of the NPs on the surface of the TiO₂ NTs, the morphology of the NPs with a plating time of 20 min could be clearly obtained via a TEM image (Fig. 3(f)). The NPs had a core–shell structure, and the average thickness of the shells was about 3.5 nm. The shell grew as an almost uniform layer around the core. Meanwhile, the EDS spectrum of the composite nanostructure with a plating time of 20 min showed that Au and Ag were present on the surface of the TiO₂ (Fig. 3(g)). Based on these characterization results and growth mechanism analysis, we confirmed that the NPs were Au–Ag bimetallic core–shell nanostructures, and that the thickness of the silver shell varied as a function of plating time. Versus the average diameters of the Au seeds, the average silver shell thicknesses of Au@Ag NPs were 1.5 nm, 3.5 nm, 5 nm and 9 nm at the plating times of 10 min, 20 min, 30 min and 60 min, respectively.

Fig. 3 SEM images of the Au@Ag/TiO₂ NTs substrate with different chemical silver plating times: (a) 0 min, (b) 10 min, (c) 20 min, (d) 30 min, (e) 60 min; (f) TEM image of the Au@Ag NPs with a chemical plating time of 20 min; (g) EDS images of the Au@Ag/TiO₂ NTs SERS substrate with a chemical plating time of 20 min.

To further evaluate the influence of the plating time on the SERS enhancement effect of Au@Ag/TiO₂ NTs, 10⁻⁵ M R6G was selected as the test sample, and the SERS spectra were collected from different Au@Ag/TiO₂ NTs with different plating times under the same test condition (Fig. 4(A)). The SERS signals were observed with six typical Raman bands appearing from 500 to 1700 cm⁻¹. The peaks of these Raman bands were located at 613, 1180, 1305, 1357, 1508 and 1642 cm⁻¹ and were assigned to the xanthene ring C–C–C bending, C–C stretching, CH₂ wagging and CH bending of R6G.⁴² Fig. 4(B) showed that the intensity of the Raman signal increased with the plating time, and the maximum SERS signals happened at a plating time of 20 min. The SERS signals decreased at the times over 20 min. This might be because of electromagnetic coupling of the double metallic layers.⁵² When the diameter of the gold core was 18 nm, the Au@Ag/TiO₂ NTs with a silver shell thickness of 3.5 nm could generate the largest electromagnetic coupling effect between the gold core and the silver shell. Versus single Au NPs, the Au@Ag NPs and Au@Ag/TiO₂ flat surface (Au@Ag/TiO₂ plate), the optimized Au@Ag/TiO₂ NTs had improved SERS peak intensities for the C–C–C bending (613 cm⁻¹) of 20, 2.3 and 1.6 times enhancement, respectively (Fig. 5). This was because the electromagnetic coupling between the bimetallic layers of the Au@Ag NPs could generate a greater SERS enhancement relative to single metal NPs. The interaction and charge transfer behaviour between TiO₂ and Au@Ag NPs could effectively facilitate charge transfer from the Au@Ag NPs to R6G.⁵³ Moreover, the large specific surface area of the 3D tubular structure could load more Au@Ag NPs versus the 2D flat structure.

Fig. 4 SERS spectra of 10⁻⁵ M R6G within microfluidic SERS chips with integrated Au@Ag/TiO₂ NTs with different chemical plating times: (a) 5 min; (b) 10 min; (c) 20 min; (d) 30 min; (e) 60 min, with (A) Raman shift in the range 500–1700 cm⁻¹; (B) Raman shift in the range 590–620 cm⁻¹.

Fig. 5 (A) SERS spectra of 10⁻⁵ M R6G within the microfluidic chips with different SERS substrates: (a) Au NPs; (b) Au@Ag NPs; (c) Au@Ag/TiO₂ plate; (d) Au@Ag/TiO₂ NTs. (B) Relationship between the SERS signal intensity and SERS enhancement substrates within microchannels at the same size of microchannel.

3.2 SERS performance of the microfluidic SERS chip with integrated Au@Ag/TiO₂ NTs

Reproducibility and sensitivity are two important indexes to evaluate microfluidic SERS chips. To evaluate the SERS detection reproducibility of the microfluidic SERS chip, 20 SERS spectra were collected from 20 different spots within the microchannel, where the space between the adjacent detection spots was 50 μm (Fig. 6(A)). The results showed that the intensities of all the spectra were very close and the RSD of 613 cm⁻¹ was 10.02%, which indicated that the microfluidic SERS chip with an integrated SERS substrate had good reproducibility in a larger range. In a smaller space, the homogeneity of the substrate was further studied via 2D Raman mapping at 613 cm⁻¹, as shown in Fig. 6(B). The mapping area was 40 × 40 μm², and the scan step was 8 μm. Dark areas represented a higher intensity of SERS signals. It is indicated in that figure that the aggregation of several Au@Ag NPs on the surface of TiO₂ NTs can cause a greater SERS enhancement in some areas. These SERS intensities showed a relatively narrow distribution with only a 9.8% deviation from the mean, indicating a homogeneous SERS response throughout the entire surface. Next, different concentrations of R6G solution were used (Fig. 7), and the SERS signal intensity of R6G decreased with decreasing concentrations. The peak at 613 cm⁻¹ remained, even at 10⁻¹⁰ M.

Fig. 6 (A) SERS spectra of 10⁻⁵ M R6G collected from 20 different detection spots on the as-prepared Au@Ag/TiO₂ NTs within the microchannel. (B) Raman mapping of the 613 cm⁻¹ peak (R6G) in the lateral direction over a 40 μm × 40 μm area of the Au@Ag/TiO₂ NTs substrate. The range of the relative Raman signal intensity, Z = 9000–11 000 au.

Fig. 7 SERS spectra of R6G with different concentrations within the microfluidic SERS chip: (a) 10⁻⁶ M; (b) 10⁻⁷ M; (c) 10⁻⁸ M; (d) 10⁻⁹ M; (e) 10⁻¹⁰ M.

Additionally, the enhancement factor (EF) is an important index to evaluate the SERS enhancement effect of the SERS enhancement substrate. It was calculated by using the following eqn (3):⁵⁴

where I_SERS and I_Raman are the corresponding SERS and normal Raman intensity of the R6G at the peak of 613 cm⁻¹, respectively, and C_SERS and C_Raman are the concentration of molecules contributing these intensities, respectively. As a result, the EF was calculated to be 1.15 × 10⁸ for the Au@Ag/TiO₂ NTs SERS substrate.

3.3 In situ monitoring of the photocatalytic degradation of R6G by using SERS

The recyclability of the substrate is another important feature of the microfluidic SERS chip. In this case, the photocatalytic activity of the Au@Ag/TiO₂ NTs substrate could be exploited to remove adsorbed R6G molecules and make the substrates ready for a new detection cycle. In a typical experiment, 10⁻⁵ M R6G solution was filled in the microchannel with integrated Au@Ag/TiO₂ NTs, which were then kept in the dark for 30 min. Then, the suspended R6G molecules were washed away from the microchannel. After that, the change process of the SERS signal intensity of the adsorbed R6G molecules with irradiation time were monitored using a confocal Raman microscope system (Fig. 8). The SERS intensity of the R6G decreased but no new Raman band appeared with prolonging the UV irradiation time. No Raman peak of R6G was found with an irradiation time of 40 min. This indicated that the R6G had been oxidized to a small molecule without Raman activity.

Fig. 8 SERS spectra of 10⁻⁵ M R6G adsorbed on the surface of Au@Ag/TiO₂ NTs under different UV irradiation times.

To evaluate the recyclability of the microfluidic SERS chip with integrated Au@Ag/TiO₂ NTs, we collected the six SERS spectra of the adsorbed R6G on Au@Ag/TiO₂ NTs (no UV irradiation). These were collected with six cycles of reuse (Fig. 9). In each cycle, the Au@Ag/TiO₂ NTs surface was recovered after the R6G was completely photodegraded and could be employed again for subsequent use.

Fig. 9 SERS spectra of R6G adsorbed on Au@Ag/TiO₂ NTs within a microchannel with six cycles of reuse.

4 Conclusions

In this paper, a novel and recyclable microfluidic SERS chip with integrated Au@Ag/TiO₂ NTs substrate was designed, and uniform Au@Ag NPs were successfully fabricated on the mouth surface and the pore walls of TiO₂ NTs without clogging by a facile self-assembly-chemical plating composite method. By controlling the PDDA concentration during the self-assembly process of Au NPs and the chemical silver plating time, the SERS enhancement effect of the SERS substrate was optimized, and the detectable concentration of R6G decreased to 10⁻¹⁰ M, with great detection repeatability. Versus single Au NPs, Au@Ag NPs and the Au@Ag/TiO₂ plate, the optimized Au@Ag/TiO₂ NTs had improved SERS peak intensities for the C–C–C bending (613 cm⁻¹) of 20, 2.3 and 1.6 times enhancement, this was possibly because of the synergistic effect of the electromagnetic coupling between the bimetallic layer of the Au@Ag NPs, the charge transfer of TiO₂ and there being more adsorbed molecules in the 3D NTs structure with the high specific surface area. In particular, the prepared substrate was highly reproducible. The adsorbed molecules could be rapidly degraded and removed from the surface of SERS substrate under UV irradiation, which makes the microfluidic SERS chip promising for applications in the detection of a variety of molecular species.

Acknowledgements

Authors would like to thank Dr Xueqiang Qi and Xiangnan Gong for the help during Raman measurement. This work was financially supported by the National Natural Science Foundation of China (NFSC: 21375156) and National High Technology Research and Development Program of China (863 Projects) (No. 2015AA021104); Frontier Research Key Projects of Chongqing Science and Technology Committee, [cstc2015jcyjBX0010], Scientific and Technical Innovation Projects for People's Livelihood of Chongqing Science and Technology Committee [cstc2015shms zx00014] and Key Project of Central University Basic Scientific Research Business Expenses (No. 106112015CDJZR225512).

References

1. L. Yang, P. Li and J. Liu, RSC Adv., 2014, 4, 49635.
2. H. Zhou, D. Yang, N. P. Ivleva, N. E. Mircescu, S. Schubert and R. Niessner, et al., Anal. Chem., 2015, 87, 6553.
3. C. De Bleye, E. Dumont, C. Hubert, P. Y. Sacre, L. Netchacovitch and P. F. Chavez, et al., Anal. Chim. Acta, 2015, 888, 118.
4. X. Lin, W. L. Hasi, S. Q. Han, X. T. Lou, D. Y. Lin and Z. W. Lu, Phys. Chem. Chem. Phys., 2015, 17, 31324.
5. H. Mao, W. Wu, D. She, G. Sun, P. Lv and J. Xu, Small, 2014, 10, 127.
6. C. C. Lin and C. W. Chang, Biosens. Bioelectron., 2014, 51, 297.
7. S. Wang, C. Liu, H. Wang, G. Chen, M. Cong and W. Song, et al., ACS Appl. Mater. Interfaces, 2014, 6, 11706.
8. T. H. Anne Marz, D. Cialla and M. Schmitta, Lab Chip, 2011, 11, 3567.
9. M. Fan, P. Wang, C. Escobedo, D. Sinton and A. G. Brolo, Lab Chip, 2012, 12, 1554.
10. H. Lee, L. Xu, D. Koh, N. Nyayapathi and K. W. Oh, Sensors, 2014, 14, 17008.
11. C. Lim, J. Hong, B. G. Chung, A. J. deMello and J. Choo, Analyst, 2010, 135, 837.
12. Q.-L. Li, B.-W. Li and Y.-Q. Wang, RSC Adv., 2013, 3, 13015.
13. J.-A. Huang, Y.-L. Zhang, H. Ding and H.-B. Sun, Adv. Opt. Mater., 2015, 3, 618.
14. H. X. Xu, J. Aizpurua, M. Kall and P. Apell, Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top., 2000, 62, 4318.
15. A. Liu, T. Xu, J. Tang, H. Wu, T. Zhao and W. Tang, Electrochim. Acta, 2014, 119, 43.
16. Y. Liu, J. Zhou, B. Wang, T. Jiang, H. P. Ho and L. Petti, et al., Phys. Chem. Chem. Phys., 2015, 17, 6819.
17. A. Garcia-Leis, A. Torreggiani, J. V. Garcia-Ramos and S. Sanchez-Cortes, Nanoscale, 2015, 7, 13629.
18. P. Guo, D. Sikdar, X. Huang, K. J. Si, W. Xiong and S. Gong, et al., Nanoscale, 2015, 7, 2862.
19. W. Lai, J. Zhou, Z. Jia, L. Petti and P. Mormile, J. Mater. Chem. C, 2015, 3, 9726.
20. F.-H. Ko, M. R. Tai, F.-K. Liu and Y.-C. Chang, Sens. Actuators, B, 2015, 211, 283.
21. X. Wang, W. Shi, G. She and L. Mu, Phys. Chem. Chem. Phys., 2012, 14, 5891.
22. Y. Wang, W. Ruan, J. Zhang, B. Yang, W. Xu, B. Zhao and J. R. Lombardi, J. Raman Spectrosc., 2009, 40, 1072.
23. B. H. Loo, J. Electroanal. Chem., 1982, 136, 209.
24. A. Kudelski, W. Grochala, M. Janik-Czachor, J. Bukowska, A. Szummer and M. Dolata, J. Raman Spectrosc., 1998, 29, 431.
25. Y. Zhou, J. Chen, L. Zhang and L. Yang, Eur. J. Inorg. Chem., 2012, 19, 3176.
26. A. Otto, I. Mrozek, H. Grabhorn and W. Akemann, J. Phys.: Condens. Matter, 1992, 4, 1143.
27. W. H. Park and Z. H. Kim, Nano Lett., 2010, 10, 4040.
28. A. Lamberti, in Encyclopedia of Nanotechnology, 2015, vol. 7, p. 1.
29. A. L. Giraldo, G. A. Penuela, R. A. Torres-Palma, N. J. Pino, R. A. Palominos and H. D. Mansilla, Water Res., 2010, 44, 5158.
30. K. Pathakoti, S. Morrow, C. Han, M. Pelaez, X. He, D. D. Dionysiou and H. Hwang, Environ. Sci. Technol., 2013, 47, 9988.
31. C. W. Dunnill and I. P. Parkin, Dalton Trans., 2011, 40, 1635.
32. V. B. R. Boppana and R. F. Lobo, J. Catal., 2011, 281, 156–168.
33. C. Sahoo, A. K. Gupta and I. M. S. Pillai, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 2012, 47, 1428.
34. S. Nie and S. R. Emory, Science, 1997, 275, 1102.
35. K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. R. Dasari and M. S. Feld, Phys. Rev. Lett., 1997, 78, 1667.
36. X. Li, G. Chen and L. Yang, et al., Adv. Funct. Mater., 2010, 20, 2815.
37. G. H. Gu and J. S. Suh, J. Phys. Chem. C, 2010, 114, 7258.
38. N. Ji, W. D. Ruan, C. X Wang, Z. C. Lu and B. Zhao, Langmuir, 2009, 25, 11869.
39. G. Guo, B. Yub and P. Yuc, et al., Talanta, 2009, 79, 570.
40. K. Roguska, et al., Appl. Surf. Sci., 2011, 257, 8182.
41. Y. Chen, G. Tian, K. Pan, C. Tian, J. Zhou, W. Zhou, Z. Ren and H. Fu, Dalton Trans., 2012, 41, 1020.
42. C. Wang, Y. Xu, H. Zhao, C. Gang, C. Lai, X. Liao and R. Wang, Appl. Surf. Sci., 2015, 353, 750.
43. H.-Z. Zhao, Y. Xu, C.-Y. Wang, R. Wang, S.-T. Xiang and L. Chen, RSC Adv., 2016, 6, 14105.
44. G. Frens, Nature Physical Science, 1973, 241, 20.
45. A. Kudelski, M. Pisarek, A. Roguska, M. Hołdyński and M. Janik-Czachor, J. Raman Spectrosc., 2012, 43, 1360.
46. S. Satyanarayana, R. N. Karnik and A. Majumdar, J. Microelectromech. Syst., 2005, 14, 392.
47. A. Wood, M. Giersig and P. Mulvancy, J. Phys. Chem. B, 2001, 105, 8810.
48. F.-H. Ko and Y.-C. Chang, RSC Adv., 2014, 4, 26251.
49. Z. Yi, X. Xu, J. Luo, X. Li, Y. Yi, X. Jiang, Y. Yi and Y. Tang, Phys. B. Condens. Matter., 2014, 438, 22.
50. Y. Ma, J. Zhou, W. Zou, Z. Jia, L. Petti and P. Mormile, J. Nanosci. Nanotechnol., 2014, 14, 4245.
51. R. Contreras-Caceres, C. Dawson, P. Formanek, D. Fischer, F. Simon, A. Janke, P. Uhlmann and M. Stamm, Chem. Mater., 2013, 25, 158–169.
52. S. Xu, B. Zhao, W. Xu and Y. Fan, Colloids Surf., A, 2005, 257–258, 313.
53. L. B. Yang, X. Jiang, W. D. Ruan, J. X. Yang, B. Zhao, W. Q. Xu and J. R. Lombardi, J. Phys. Chem. C, 2009, 113, 16226.
54. J. Parisi, Q. Dong and Y. Lei, RSC Adv., 2015, 5, 14081.

---