Effects of Ag contents on the microstructure and SERS performance of self-grown Ag nanoparticles/Mo–Ag alloy films

Ag nanoparticles/Mo–Ag alloy films with different Ag contents were prepared on polyimide by magnetron sputtering. The effects of Ag contents on the microstructure of self-grown Ag nanoparticles/Mo–Ag alloy films were investigated using XRD, FESEM, EDS and TEM. The Ag content plays an important role in the size and number of uniformly distributed Ag nanoparticles spontaneously formed on the Mo–Ag alloy film surface, and the morphology of the self-grown Ag nanoparticles has changed significantly. Additionally, it is worth noting that the Ag nanoparticles/Mo–Ag alloy films covered by a thin Ag film exhibits highly sensitive surface-enhanced Raman scattering (SERS) performance. The electric field distributions were calculated using finite-difference time-domain analysis to further prove that the SERS enhancement of the films is mainly determined by “hot spots” in the interparticle gap between Ag nanoparticles. The detection limit of the Ag film/Ag nanoparticles/Mo–Ag alloy film for Rhodamine 6G probe molecules was 5 × 10 mol/L. Therefore, the novel type of the Ag film/Ag nanoparticles/Mo–Ag alloy film can be used as an ideal SERSactive substrate for low-cost and large-scale production.


Introduction
Surface-enhanced Raman scattering (SERS) technique, a highly sensitive fingerprint identification tool [1], has shown enormous application potential in various fields such as biomedical science [2], environmental monitoring [3], food and drug safety [4][5][6], catalysis [7], etc. Due to the localized surface plasmonic resonance effect, SERS provides a large Raman signal enhancement for molecules close to the plasma nanostructures [8,9]. SERS technology has been proved as a promising method for the rapid and sensitive detection of chemicals and biochemicals [10][11][12]. At the same time, it is crucial to prepare high-performance SERS substrates in order to achieve an efficient SERS detection.
The size, morphology and the composition of various kinds of nanoparticles can be controlled by physical and chemical methods in order to prepare SERS substrates with excellent properties [13][14][15]. Metal colloids are the first type of commonly used SERS substrates, mainly because of its effortless preparation and large SERS signals enhancement [16]. However, avoiding the spontaneous aggregation of colloidal metal nanoparticles is still a challenging problem. Nowadays, this challenge can be solved using some novel techniques, including inkjet printing [17], oblique angle deposition [18], lithography [19], vacuum thermal evaporation [20] and self-assembly techniques [21]. Researchers generally use these technologies to uniformly distribute metal nanoparticles on specific support substrates. In comparison with the brittle silicon and glass SERS substrates, it is worth noting that the flexible SERS substrates are easy to cut and suitable for test objects with complex shape surface [22]. Therefore, it is urgent to develop a sensitive, uniform and low-cost flexible SERS substrate, which can promote the practical application of SERS substrate. The commercial tape with colloidal gold nanoparticles can be used as a novel flexible and adhesive SERS substrate to effectively extract objects from complex surfaces [23], which is expected to bring the SERS technology closer to real applications. The Au@Ag/PMM/qPCR-PET film chip was applied as a stable and high-performance SERS chip for highly sensitive sensing of thiabendazole residues in fruit juices [24]. The flexible SERS substrates can be integrated into portable Raman spectroscopes for pointof-care diagnostics [25], which are conceivable to penetrate global markets and have important economic value and social significance.
A uniform, stable, highly ordered and controllable nanoparticle film can be prepared by self-assembly method [26], which can improve the reproducibility and sensitivity of SERS detection. A previous study reported that the annealed Ag-Zr film deposited on polyimide (PI) can produce a new type of particle/film structure [27], which is significantly different from the traditional granular film [28][29][30][31]. Considering that the Ag nanoparticles spontaneously formed on the surface of the annealed Ag-Zr alloy films and the nanophase effects in copper alloys [27,[32][33][34][35][36][37], we attempt to selfgrown Ag nanoparticles with different morphologies on the surface of the deposited Mo-Ag alloy films by adjusting the Ag content at room temperature. If this purpose can be achieved, a new type of SERS substrate with excellent performance can be obtained by combining the Ag nanoparticles with high SERS sensitivity and the Mo-Ag alloy films with good repeatability at room temperature. The SERS enhancement was characterized using Rhodamine 6G (R6G) probe molecule on the Ag nanoparticle/Mo-Ag alloy film. The effects of Ag contents on the microstructure and SERS performance of the Ag nanoparticles/Mo-Ag alloy film were investigated. The PI substrate was cleaned with acetone, deionized water and alcohol for 10 min, respectively, and the PI substrate was fixed on the substrate holder after drying. In order to reduce the influence of impurity gas in the chamber, the vacuum degree of the chamber is pumped to 5 × 10 −4 Pa. The argon flow rate, sputtering power and sputtering pressure were adjusted to 40 sccm, 100 W and 0.4 Pa, respectively. The impurities on the target surface were removed by presputtering for 5 min and then sputtered for 10 min to prepare Mo-Ag alloy films. Unless otherwise mentioned, all experiments were carried out at room temperature.

Characterization of the Ag
nanoparticles/Mo-Ag alloy film X-ray diffractometer (Bruker-AXS D8 Advance) and transmission electron microscopy (TEM, JEM-2100) were used to characterize the phase structure and the crystal structure of the Ag nanoparticles/Mo-Ag alloy films. X-ray diffractometer uses a Cu target (λ = 0.1541 nm), and the voltage and current parameters are set at 40 kV and 40 mA, respectively. The scanning speed is 6°/min, the step size is 0.02°and the range of angle is 20°-90°. The high voltage of the TEM is 200 kV. The surface morphology and the composition of the Ag nanoparticles/Mo-Ag alloy films were observed and confirmed using a field emission scanning electron microscope (FESEM, JSM-7800F) with energy dispersive spectroscopy (EDS). The electric field distributions were calculated using finite-difference timedomain (FDTD) analysis. The number and average size of the nanoparticles on the surface of the Mo-Ag alloy films were measured by ImageJ software. Figure 1 shows the schematic diagram of the Ag film/Ag nanoparticles/Mo-Ag alloy film formation and the detection process of SERS performance. The Mo-Ag alloy films were deposited on flexible PI substrate by DC magnetron sputtering (Figure 1-I). Considering that the Mo and Ag are immiscible at room temperature [38] and the initial Mo-Ag alloy film will be in a metastable state, the Mo-Ag alloy films form a stable structure through atoms diffusion along the surface and grain boundaries under the relaxation of residual stress and distortion energy [39]. As a result, a large amount of the Ag nanoparticles were self-formed on the surface of the Mo-Ag alloy film by adjusting the Ag content. Subsequently, a pure Ag target (99.99 %, Ø 50 mm × 4 mm) was used to further sputter a layer of Ag film on the surface of the Ag nanoparticles/Mo-Ag alloy films under the same sputtering parameters as the preparation of the Mo-Ag alloy film (Figure 1-II). The most common probe molecule, R6G, was selected to measure the performance of the SERS substrate [40]. The Ag film/Ag nanoparticles/ Mo-Ag alloy films were tailored into 5 × 5 mm pieces (Figure 1-III) and soaked in different concentrations of R6G in aqueous solutions (20 mL) for 60 min (Figure 1-IV). After the droplet has been evaporated naturally, samples were characterized using the laser wavelength 532 nm Raman spectrometer (Invia, Renishaw, UK) with 50× objective lens (N.A. = 0.80) (Figure 1-V). The laser excitation energy and spot, respectively, were 5 mW and 2 µm, respectively. The diffraction grid was 1,200 g/mm. The Raman spectra were collected in the range of 200-2,000 cm −1 with the acquisition time of 1 s.

Results and discussions
3.1 Surface morphology of the Ag nanoparticles/Mo-Ag alloy films  Figure 2(c-f). This is different from the surface morphology of pure Ag film and pure Mo film. According to the EDS spectra in Figure  2(c1-f1), the Ag contents of the Mo-Ag alloy films in Figure  2(c-f) are 18.0, 25.5, 38.5 and 51.2 at%, respectively. The cross-sectional morphology of the Mo-Ag alloy films is shown in Figure 2(c2-f2), and the relative position relationship between the self-formed particles and the film is obviously different from the particles embedded in the traditional film [26]. As shown in Figure 2  irregular "worm-like," which is different from the regular nanoparticles obtained on the surface of the annealed Ag-Zr alloy films [27]. The average nanoparticle size on the surface of the Mo-38.5 at% Ag film is increased by 43 nm compared with the nanoparticle size on the surface of the Mo-25.5 at% Ag film. However, the number of the nanoparticles on the Mo-38.5 at% Ag film is significantly lower than on the 25.5 at% film. Moreover, it is worth noting that some nanoparticles on the surface of the Mo-38.5 at% Ag film have changed from irregular "worm-like" to polyhedron as shown in Figure 2(e). This is significantly different from the morphological transformation from flower-shaped particles to polyhedral particles on the surface of the annealed Mo-Cu alloy film observed previously [41]. The driving force for the formation of the Ag nanoparticles decreases with the increase of Ag content, resulting in a decrease in the number and particle size of the Ag nanoparticles in the Mo-51.2 at% Ag film compared with the Mo-38.5 at% Ag film. Furthermore, the morphology of the nanoparticles on the surface of the Mo-51.2 at% Ag film has completely been changed into polyhedron as shown in Figure 2(f). The number and average size of the nanoparticles on the surface of the Mo-Ag alloy films were measured by ImageJ software, as shown in Figure 2(g) and (h). The nanoparticles were taken from the surface of the Mo-51.2 at% Ag films and proved to be single-crystal Ag nanoparticle by TEM characterization as shown in Figure  2(i) and (j), respectively. Based on the above analysis, the Ag atoms on and near the film surface can easily diffuse along the grain boundaries to the triple junctions and the vacancies to aggregate and nucleate continuously with the increase of the Ag content in the Mo-Ag alloy films. Meanwhile, the adjacent Ag nanoparticles may contact with each other and grow into one Ag nanoparticle on the surface of the Mo-Ag alloy film.

Ag nanoparticles/Mo-Ag alloy films as SERS substrate
The surface morphology and Raman spectra of the Ag nanoparticles/Mo-Ag alloy films are shown in Figure 4.
Despite the large number of Ag nanoparticles formed on the surface of the Mo-Ag alloy films, the SERS performance of the Ag nanoparticles/Mo-Ag alloy film is still not efficient due to the small size of the Ag nanoparticles [42,43]. Therefore, a 23-nm Ag film was further sputtered on the surface of the Ag nanoparticles/ Mo-25.5 at% Ag film, and its surface morphology is shown in Figure 4(a). The Ag nanoparticles/Mo-25.5 at% Ag film and the Ag film/Ag nanoparticles/Mo-25.5 at% Ag film were soaked in the 5 × 10 −6 mol/L R6G solution for 60 min and tested after air-dried naturally to compare the Raman intensity. As shown in Figure 4(b), the SERS performance of the Ag nanoparticles/Mo-25.5 at% Ag film after sputtering of Ag film is significantly enhanced. In order to better understand the enhancement mechanism of the Ag film/Ag nanoparticles/Mo-Ag film as an efficient SERS substrate, we have calculated the electric field distributions of the Ag nanoparticles/Mo-Ag film and the Ag film/Ag nanoparticles/ Mo-Ag film using FDTD analysis, as shown in Figure  4(c) and (d). It obviously shows that smaller gaps between adjacent nanoparticles on the surface of the Ag film/Ag nanoparticles/Mo-Ag alloy film can increase the amount of "hot spots" and improve the sensitivity of the SERS substrate [44]. Thence, a thin Ag film was further sputtered on the surface of the Ag nanoparticles/Mo-Ag alloy films with different Ag contents. The intensity of the Raman peaks increases and then decreases with the increase in the Ag content as shown in Figure 4(e). This is closely related to the number, size and shape of Ag particles. Moreover, the SERS performance of the Ag film/Ag nanoparticles/Mo-25.5 at% Ag film as SERS substrate is excellent. This can be ascribed to the small gaps between adjacent nanoparticles, the irregular nanoparticle shapes and the rough surface of the films. It is worth noting that the maximum number of particles on the surface of the Mo-25.5 at% Ag alloy film leads to the best SERS performance among Mo-Ag alloy films with different Ag content. This is consistent with the largest number of Ag particles on the annealed Ag-Zr alloy film surface leading to the strongest SERS enhancement [27]. The previous studies have also indicated that the particle shape of the annealed Mo-Cu alloy film surface is crucial for the SERS enhancement [41]. The Ag film/Ag nanoparticles/Mo-25.5 at% Ag films were tailored into 5 × 5 mm pieces and soaked in the R6G solutions of 5 × 10 −6 , 5 × 10 −8 , 5 × 10 −10 , 5 × 10 −12 and 5 × 10 −14 mol/L, respectively. As can be seen from Figure 4(f) that the intensity of characteristic peaks of the SERS spectra decreased with the concentration of R6G solution decreases. As the concentration of R6G solution is 5 × 10 −14 mol/L, weak characteristic peaks can be observed in the Raman spectra of the Mo-25.5 at% Ag films, which is enough to identify the existence of R6G probe molecules. Therefore, the number of particles per unit area and the particle average size on the surface of the Mo-25.5 at% Ag films are 418/µm 2 and 24 nm, respectively, and it can be applied as highly effective SERS substrates for the detection of 10 −14 mol/L R6G solution due to its abundant active "hot spots." To some extent, a large number of high-density "hot spots" are generated due to the small gap between adjacent particles, resulting in the SERS enhancement with good uniformity and reproducibility [45].
To evaluate the enhancement factor (EF) of the Ag film/Ag nanoparticles/Mo-25.5 at% Ag alloy films as SERS substrate, the average EF of the characteristic peak of 613 cm −1 for the R6G solution on the SERS substrates was calculated according to the following formula: For practical applications, an efficient SERS substrate should possess not only a large number of "hot spots" to ensure high sensitivity but also the uniform distribution of "hots pots" to achieve good signal reproducibility [46,47]. Therefore, as shown in Figure  4(g), the 15 randomly selected locations on the Ag film/ Ag nanoparticles/Mo-25.5 at% Ag films immersed in the 5 × 10 −8 mol/L R6G solution and naturally air-dried were measured to evaluate the uniformity and reproducibility of the SERS substrate. The strongest characteristic peaks of R6G are the 613 cm −1 peaks (C-C-C in-plane bending), the 774 cm −1 peaks (C-H out-plane bending), and the C-C stretching of 1,362, 1,508, and 1,650 cm −1 peaks [48]. The estimation results of the deviation of the Raman characteristic peaks of the Ag film/Ag nanoparticles/Mo-25.5 at% Ag films were shown in Figure 4(h), indicating that the films have good uniformity and reproducible SERS performance.

Conclusion
Mo-Ag alloy films with different Ag contents were fabricated on flexible PI substrate by DC magnetron sputtering. The addition of Ag can inhibit the growth of Mo grains in the Mo-Ag alloy films and can result in a large amount of Ag nanoparticles with different morphologies spontaneously formed on the surface of as-deposited Mo-Ag alloy films. Moreover, as the Ag content increases, the number of the Ag nanoparticles significantly increases first and then decreases, and the morphology of Ag nanoparticles has gradually evolved in the following way: spherical → "worm-like" → "worm-like" + polyhedral → polyhedral. Interestingly, the "worm-like" Ag nanoparticles/Mo-25.5 at% Ag films have abundant active "hot spots," and further sputtering of the Ag film can be used as high-efficient SERS substrates. The detection limit of the Ag film/Ag nanoparticles/Mo-25.5 at% Ag film as SERS substrate for R6G probe molecule is 10 −14 mol/L, and the EF is 2.69 × 10 12 . Therefore, the Ag film/Ag nanoparticles/Mo-25.5 at% Ag film can be used as an efficient SERS substrate for low-cost and large-scale production.