1. Introduction

Raman spectroscopy, a kind of molecular vibration spectroscopy, can provide vibration information related to molecular structure [1]. However, the conventional Raman signals of most molecules are usually very weak due to the little light scattering cross-sections [2,3]. The weak signals and low sensitivities limit the application of Raman spectroscopy. In 1974, Fleischmann et al. [4] fond that a rough silver electrode surface can effectively enhance the Raman scattering signals of pyridine molecules. Afterwards, Jeanmaire and Van Duyne [5] verified this phenomenon and called it as a surface enhancement effect. Since then, surface-enhanced Raman scattering (SERS) technologies have been extensively studied and developed, and the enhancement factors (EFs) and detection sensitivities of SERS signals have been greatly improved. Nowadays, the SERS have been widely used in drug analysis [6,7], food safety [8,9] and biomedical sensor [10–12].

Generally, the metals, such as gold (Au), silver (Ag), copper (Cu) and aluminum (Al), have been used for manufacturing nanostructure substrate for SERS application [13]. Moreover, the enhancement effect of SERS substrate can be also modulated by the size and shape of metal nanostructures [14–18]. Thus, it is critical to prepare a fine and high EF metal nanostructure for SERS application. At present, many types of metal nanostructures, for example, rough films [19], nanoparticles (NPs) [20], nanowires [21], nanotubes [22], nanocages [23], nanoshell-core microspheres [24] and nano arrays [25], etc. have been as SERS substrates. The methods of fabricating these metal structures include wet chemical method [26], electron beam lithography method [27], interference lithography method [28], laser ablation method [29], electrochemical deposition (ECD) method [30], inkjet or screen printing method [31], and electrospinning method [32], etc.. In these methods, the ECD method requires a complicated process. Moreover, the reference electrode can also reduce the surface cleanliness of the substrate [33]. The inkjet printing method requires to add the metal NPs into the ink, which has a poor stability and reproducibility for quantitative analysis [34]. The electron beam lithography method can finely control geometry of substrate with high repeatability, but there are some defects, such as expensive cost, un-suitable for large-area manufacture [35]. In general, the ideal SERS substrates should have a good uniformity, reproducibility, stability, economy and surface cleanliness [36]. Nevertheless, most of the above methods have a complex operation process, and it is difficult to prepare a large number of products. Therefore, the development and preparation of SERS substrates which are low cost, high efficiency and high stability are still a very difficult issue in the practical application of SERS.

Polyvinyl alcohol (PVA) has a well water-solubility, biocompatibility, stability and biodegradability, which has been widely applied as host materials in the process of preparing metal NPs [37]. Kong et al. [38] developed a SERS substrate based on PVA/Au NPs colloid to detect the thiram residues in fruits and vegetables. Zhang et al. [39] prepared PVA/Ag nanofibers via electrospinning technique, which made a significant contribution in the detection of 4-mercaptophenol molecules with high sensitivity. Xue et al. [40] obtained a highly conductive PVA/Ag hybrid using thermal annealing PVA/silver nitrate (AgNO₃) film. In addition, the Ag/PVA/zinc oxide (ZnO) nanofiber film which displayed a good antibacterial activity against Escherichia coli and Klebsiella Spp. Bacteria was synthesized by Hamza et al. [41].

Here, we used the AgNO₃ and PVA mixture solution to prepare a uniform composite film on ITO glass surface via spin-coated method, and the large-area uniform Ag nanostructure films were formed by a heating reduction method for in situ detecting melamine molecules. Compared with other pathways of prepared Ag NPs SERS substrates, it didn’t need complex experimental conditions and chemical process. This research also provides a new pathway for the application of low-cost and large-area SERS substrates.

2. Experiments and methods

2.1 Experimental materials

PVA powder (Polymerization degree of 1700 and alcoholysis degree of 98%, Aladdin Reagent Co., Ltd.) was used as a reducing agent. AgNO₃ powder (AR, Sinopharm Chemical Reagent Co., Ltd.) was used for reducing Ag NPs. Melamine powder (AR, Sinopharm Chemical Reagent Co., Ltd.) was acted as the target detection molecules. All reagents were used without further purification.

2.2 Preparation of Ag nanostructure films

Figure 1. showed the preparation process of Ag nanostruture films in SERS application. The preparing process were fellow as: (1). 5 g PVA powder was dissolved in 45 ml deionized water and then was stirred at 500 rpm for 1 h at 90 °C to obtain a 10 wt% PVA aqueous solution. (2). AgNO₃ powder (0, 0.1, 0.2, 0.3, 0.4, and 0.5 g) was respectively dissolved in 1 ml deionized water to prepare a series of AgNO₃ solutions of 0, 8, 17, 23, 28, and 33 wt%. (3). The AgNO₃ solutions were uniformly mixed with 5 ml PVA solution respectively, keeping for 1 hour. (4). Afterwards, 0.2 ml PVA/AgNO₃ composite solutions with different AgNO₃ contents were evenly dropped on the surface of ITO glass (20 mm × 20 mm × 0.55 mm, South China Science & Technology Co., Ltd.). For the same ITO glass, it needed to be spin-coated by a spinner (CB-1B, Jinshengdao, China) at three different rotation speeds with different time. The sequence of spin-coating was as follows: the rotation speeds of 500 rpm for 10 s, then the rotation speeds of 1000 rpm for 10 s and finally the rotation speeds of 4000 rpm for 40 s. (5). These films were laid on the surface of aluminum plate of 80 mm × 80 mm × 3 mm, and then were heated for 10 minutes by an alcohol lamp to reduce Ag nanostructure films. Before the heat treatment, the plate had been heated for 10 minutes to approach and maintain about 500 °C. Here, the uniformity of the temperature for the heating plate was 500 ± 3°C.

 

Fig. 1. Preparation process of Ag nanostruture films.

Download Full Size | PDF

2.3 Characterization of Ag nanostructure films

The surface morphologies were recorded by a material microscope (Lecia, DM-2700) equipped with a charge coupled device (CCD, Du934P, Andor). The thickness of Ag nanostructure films were measured by a contact step meter (ET150, Kosaka Lab, Japan). The Scanning electron microscope and energy dispersive spectrometer (SEM-EDS, SU8100, Hitachi, Japan) were used for measuring the morphologies and elemental analyses of Ag nanostructure films. The surface chemical compositions of Ag nanostructure films were measured by X-ray diffractometer (D8 Advance, Bruker, Germany) and Fourier transform infrared spectrometer (V-70+H1000, Bruker, Germany). Raman spectra were measured by a micro-Raman spectrometer (SR-500I-B1, Andor, North Ireland) with the excitation source of a 532 nm continuous laser (Torus, Laser Quantum, England) via a material microscope.

2.4 SERS detection of melamine molecules

We prepared 6 pieces of Ag nanostructure films which the content of AgNO₃ was respectively 0 wt%, 8 wt%, 17 wt%, 23 wt%, 28 wt%, and 33 wt% to test their EFs.

50 mg melamine powder was dissolved in 500 ml methanol solution, and was stirred at 300 rpm and 60 °C for 10 minutes to prepare a concentration of 100 µg/ml melamine-methanol solution. Then 100 µg/ml melamine-methanol solution were diluted to 1000, 100, 60, 30, 20, and 10 ng/ml melamine-methanol solutions for test.

All the films were respectively immersed in 20 ml melamine solutions with different concentrations for 8 h, and then were taken out to dry for SERS test.

3. Results and discussions

3.1 Morphologies of Ag nanostructure films

The morphologies of Ag nanostructure films were shown in Fig. 2. The color of PVA/AgNO₃ films via heat treatment for 0, 2, 4, 6, 8, 10 minutes in a row was shown in Fig. 2(a). The surfaces of all films were smooth and the color distribution of single film was uniform and consistent. Moreover, each film had a large area of 3.24 cm². When the films were not heated, they were colorless. After the heat treatment, the films with AgNO₃ were silver white, while the film without AgNO₃ was hazel. And the 28 wt% AgNO₃ film was yellow, light black, dark gray and off-white at heating time of 2, 4, 6, 8 minutes. After the heat treatment, there were no particles on the surface of non-AgNO₃ film as shown in Fig. 2(b). But in Fig. 2(c), there were uniform and dense particles on the 28 wt% AgNO₃ film. This indicated that under the heating condition, PVA could easily reduce silver ions (Ag⁺) in AgNO₃ to Ag NPs, which guaranteed that the uniform Ag nanostructure films could be prepared. To measure the thickness of Ag nanostructure film, a groove without Ag nanostructure was scratched by the tip of a tweezer on the 28 wt% AgNO₃ film, as shown in Fig. 2(d). It can be seen that the width of groove was 60 µm and the depth of groove was 100 nm which was also the thickness of Ag nanostructure film in Fig. 2(e).

 

Fig. 2. (a) Color of PVA/AgNO₃ films with increasing time of heat treatment. (b) Surface images without AgNO₃. (c) Surface images of the 28 wt% AgNO₃ film. (d) Surface images of the 28 wt% AgNO₃ film for thickness mesurement. (e) Profile picture of thickness from (d).

Download Full Size | PDF

The thermal redox process of PVA/AgNO₃ composite films was shown in formula (1) [42]:

In the room temperature, the strong hydrogen bonds in the PVA molecular chains were broken when AgNO₃ solution was mixed with PVA solution. And a coordination structure was formed because the lone pair electrons on free hydroxyl groups of PVA entered the d or f orbitals of Ag⁺. In the heat treatment, with the gradually increasing temperature of the film, the PVA/AgNO₃ chelate was drastically decomposed and the multiple pendant hydroxyl groups were oxidized to carbonyls to reduce Ag⁺ and stabilize Ag NPs via the electron donating hydroxyl functionalities. Thus, large numbers of Ag NPs were formed by the aggregation of Ag⁰ on the film.

The content of AgNO₃ directly determined the morphology of the Ag nanostructure on the film. When the AgNO₃ content was 9 wt%, the average size of reduced Ag NPs was 200 nm, as shown in Fig. 3(a). On the 17 wt% AgNO₃ film, the peanut-like nanoribes formed with Ag NPs had an average length of 800 nm and an average width of 300 nm in Fig. 3(b). While the content of AgNO₃ was increased to 23 wt%, the peanut-like nanoribes connected with each other on the film in Fig. 3(c). As shown in Fig. 3(d), on the 28 wt% AgNO₃ film surface, the uniform labyrinth-shaped Ag nanostrutures had an average skeleton size of 400 nm and an average skeleton gap of 300 nm. When the content of AgNO₃ was reached 33 wt%, the Ag nanostruture on the film was very uneven, and a lot of bulges and depressions were generated on the film surface as shown in Fig. 3(e). Additionally, the outline curves of the skeleton structures were clear in the subgraphs, indicating that the nanostructures were formed with the fused Ag NPs. Figure 3(f) also showed the large ratio of Ag nanostructures to film area from (a) to (e). With the increase of AgNO₃ content on the film, the more Ag NPs were generated, and the area of Ag nanostructures was also increased. When the AgNO₃ content was 28 wt%, the filling rate could reach 63%.

 

Fig. 3. SEM images of the films with (a) 9 wt%, (b) 17 wt%, (c) 23 wt%, (d) 28 wt%, (e) 33 wt% AgNO₃. (f) The ratio of Ag nanostructures to film area from (a) to (e).

Download Full Size | PDF

According to literature reports [43], the formation process of the skeleton structure could be described as follows: the coordination structure was formed when the PVA and AgNO₃ solutions were mixed well. In the initial heating process, the Ag NPs were gradually reduced on the composite film surface. The Ag NPs were not aggregated owing to the stabilization of PVA, and they were evenly distributed on the surface, presenting a metallic luster. With the further degradation of PVA, PVA was difficult to act as a stabilizer, thus the Ag particles were fused with each other and grew along the original path, eventually forming the skeleton structure.

3.2 EDS, XRD and FTIR of Ag nanostructure films

Here, in order to determine the composition of Ag nanostructure film, the EDS, XRD and FTIR spectra of the uniform labyrinth-shaped Ag nanostructure film of 28 wt% AgNO₃ after heat treatment were measured. As shown in Fig. 4(a), the film was mainly composed of Ag elements, indicating that the Ag nanostructures were relatively pure. Figure 4(b) showed the SEM image of EDS scanning region. As shown in Fig. 4(c), the distribution of Ag elements was relatively uniform, illustrating that the film had a good uniformity. As shown in Fig. 4(d), The XRD pattern for the Ag nanostructure film has four broad peaks at 38.2°, 44.5°, 64.6° and 77.5° corresponding to Ag(111), (200), (220) and (311), respectively. Furthermore, this four peaks can be well-indexed to Ag (JCPDS 04-0783), and the other peaks can be well-indexed to In₂O₃ (JCPDS 06-0416) and SnO₂ (JCPDS 06-0395), which were associated with the ITO layer. Figure 4(e) showed the FTIR spectra of the 28 wt% AgNO₃ film before and after the heat treatment. Before the heat treatment, the broad and strong absorbing peak observed in the range of 3600-3000 cm^-1 was the O-H stretching vibration of free hydroxyl and hydrogen bonds. Besides, the bands at 2942 and 2918 cm^-1 peaks were respectively regarded as the asymmetric stretching vibration and symmetric stretching vibration of the -CH₂ groups. Moreover, the bands at 1702 and 1655 cm^-1 peaks were respectively attributed to the C = O stretching vibrations in esters and ketenes. Furthermore, the bands at 1424 and 1377 cm^-1 peaks were attributed to the bending vibration of the C-H groups [44]. Nevertheless, there was no obvious absorption peak on the film after the heat treatment, indicating that the PVA was mostly degraded by the heat treatment.

 

Fig. 4. (a). The EDS of Ag nanostructure film with 28 wt% AgNO₃. (b). The SEM image of EDS fast mapping. (c). Distribution of Ag elements. (d). The XRD pattern of Ag nanostructure film with 28 wt% AgNO₃. (e) The FTIR spectra of 28 wt% AgNO₃ film before and after heat treatment.

Download Full Size | PDF

3.3 EFs of Ag nanostructure films

Figure 5(a) showed the Raman spectra of the 28 wt% AgNO₃ film and the melamine powder. Obviously, the film had no Raman peaks, but the melamine powder had a sharp Raman peak at 677 cm^-1 caused by the ring breathing vibration of melamine molecules [45]. Therefore, the 677 cm^-1 could be selected as the characteristic peak of melamine molecules for analysis.

 

Fig. 5. (a) Raman spectra of the 28 wt% AgNO₃ film and the melamine powder on the SiO₂. (b) Raman spectra of melamine molecules from Ag nanostructure films with differential AgNO₃ contents. (c) Histograms of relative standard deviation of Raman intensity at 677 cm-1. (d) EFs of SERS intensity of the films with differential AgNO₃ contents.

Download Full Size | PDF

The Raman spectra of 100 µg/ml melamine molecules on the films with differential AgNO₃ via SERS detection were shown Fig. 5(b). Obviously, there was not the Raman peak without Ag nanostructure film, while that of Ag nanostructure films had strong sharp peaks of SERS signals. With the increase of AgNO₃ contents, the intensities of SERS peaks were also increased. Especially, the intensity of SERS signal on the 28 wt% AgNO₃ film was increased up to more than 20000 units. But on the 33 wt% AgNO₃ film, it was decreased. To clearly show a reproducibility of SERS signals, the histograms of relative standard deviation (RSD) of SERS intensity at 677cm^-1 were shown in Fig. 5(c) in differential AgNO₃ contents. And The RSD data was calculated from the peak intensity of 5 different measurement points at random. The RSD of the 28 wt% AgNO₃ film was the smallest of 8%, showing a good reproducibility, while the film with 33 wt% AgNO₃ was 30% because of the uneven pore with size and distribution on the surface.

According to Eq. (2) [46], we calculated the EFs of the films with differential AgNO₃ contents.

Here, ${I_{SM}}$ and ${I_{RM}}$ were respectively the SERS intensity of melamine at 677 cm^-1 from the samples and the reference sample. ${C_S}$ and ${C_R}$ were respectively the concentrations of melamine molecules from the samples and the reference sample. The calculated EFs were shown in Table 1 and Fig. 5(d).

Figure 5(d) showed the EFs of SERS intensity of the films with differential AgNO₃ contents. With the increase of AgNO₃ on the films, the EFs were also increased. The EF of the 28 wt% AgNO₃ film was up to 1149. However, the EF of the 33 wt% AgNO₃ film was declined.

Table 1. EFs of SERS intensity of the films with differential AgNO₃ contents.

View Table

The enhancement effect of SERS signals were depended on the excitation of high local electromagnetic fields near metal nanostructures. When the surface Plasmons (SPs) of metal nanostructures were excited by laser of appropriate frequencies, the localized surface Plasmon resonances (LSPR) was formed by the collective oscillation of the conduction electrons in the metal nanostructures. Then the coupling of LSPR and incident light caused the electromagnetic fields to be concentrated and enhanced on a very small area of the nanostructure. Molecular polarization was greatly amplified in this region of enhanced electromagnetic fields. This amplification was further enhanced LSPR between adjacent plasma nanostructures to form the so-called electromagnetic hot spots, which was the key to obtain a strong SERS signals [47].

Here, when the films of Ag nanostructures were excited by 532 nm laser, the hot spots effect was generated between them to enhance the local electromagnetic fields for enlarging the SERS signlas. With the increase of AgNO₃ contents, the more Ag NPs were produced and aggregated on the film surface, the stronger hot spots effects were caused, which could greatly enhance the SERS signals of melamine molecules. Nevertheless, the 33 wt% AgNO₃ film had a reduced SERS signals because the uneven fused skeleton structures and the bumpy surface had reduced the hot spots effects. Furthermore, the bulges and depressions had also a great influence on the focus of the instrument, which reduced the collection of the SERS signals.

3.4 SERS sensitivity and LOD of Ag nanostructure films

Here, we selected the 28 wt% AgNO₃ films with the maximal EF to detect melamine molecules with the differential concentration. The Raman spectra were shown in Fig. 6(a). And the partial Raman spectra were enlarged in Fig. 6(d). The SERS intensities and standard deviations at 677 cm^-1 were calculated as shown in Fig. 6(b). With the decrease of melamine molecule concentrations, the SERS peak value and standard deviations were also gradually decreased. When the concentration was 1000 ng/ml, the SERS peak at 677 cm^-1 was 12702 units and the standard deviation was 445 units. While the concentration was 10 ng/ml, they were respectively 543 and 65 units, which also showed a good identification for melamine molecules. The standard deviation displayed a good stability of the film for melamine molecules detection.

 

Fig. 6. (a) Raman spectra of melamine molecules with differential concentration. (b) Histogram of Raman intensity of melamine molecules at 677 cm^-1. (c) The relation between the melamine concentration and SERS intensity. (d) The enlarged partial view of (a). (e) the LOD of melamine detection.

Download Full Size | PDF

According to Fig. 6(b), the linear fit calibration curve with error bars between the melamine concentration and SERS intensity was illustrated in Fig. 6(c), which showed a good capability and reproducibility of the films for the quantitative detection of melamine molecules. The good linear response of SERS with a determination coefficient of 0.9983 was obtained from 10 to 100 ng/ml. In addition, the slope of equation was 76.91, illustrating that the films had a high sensitivity to detect low concentration melamine molecules. Especially, the LOD of melamine detection shown in Fig. 6(e) was 3.85 ng/ml calculated by the intersection between the equation and the 3 times of noise level.

4. Conclusions

We have presented a simple method to prepare an active film of large-area uniform Ag nanostructure for SERS application. By using the thermally reducing PVA/AgNO₃ composite films, the film based on a labyrinth-shaped Ag nanostructure with the average skeleton size of 400 nm and the average skeleton gap of 300 nm was obtained. Moreover, the film had a large area of 3.24 cm², a thickness of 100 nm and a filling rata of Ag nanostructures of 63%. Melamine molecules were as a probe, the film displayed an excellent SERS activity with a EF of 1149 and a minimum detected concentration of 3.85 ng/ml. On the Ag nanostructure films, a good linear relationship between SERS peak intensity and melamine concentration was obtained from 10 to 100 ng/ml. We believe this SERS film can be widely applied in food, drug and bio-medicine, and its preparing method can provide a new pathway for the application of low-cost and large-area SERS substrates.