1. Introduction

As an interference-free ultrasensitive analytical technique, surface-enhanced Raman spectroscopy (SERS) has been widely used in various biochemical assays at the single-molecule level [1–3]. As we all know, generally, the enhancement mechanism of SERS signal of adsorbed probe molecule includes electromagnetic mechanism (EM) and chemical mechanism (CM) [4,5]. Furthermore, EM can be ascribed to enhanced local electromagnetic fields (usually called hot spots) at the surface of coinage metal nanostructures, and CM is related to charge transfer or energy transfer between the analytes and the substrate. In addition, Electromagnetic enhancements caused by the light-excited surface plasmon resonance (SPR) are usually orders of magnitude higher than chemical enhancements caused by charge transfer or energy transfer process [6–10]. However, the development of SERS substrates composed solely of noble metals were limited due to the disadvantages of high cost and poor biocompatibility [11].

Simultaneously, semiconductor SERS substrates have also been widely developed [12,13]. Transition metal dichalcogenides (TMDs), which have already attracted the intensive interest of many researchers, have broad application prospects because of their distinctive optical and electrical properties [14–19]. Especially, TMDs play an important role in exploring the CM of SERS. As a new member of the TMDs family, Rhenium disulphide (ReS₂) possesses many extraordinary properties different from traditional TMDs materials such as MoS₂, MoSe₂, WS₂ and WSe₂, because of its distorted octahedral (1T) crystal structure [20–23]. ReS₂ maintains a direct band gap semiconductor owing to electronic and vibrational decoupling, whether it is single-layer, few-layer or bulk. As a consequence, ReS₂ possess huge development potential and is high-profile in the field of catalysis, energy storage, energy harvesting and solid-state electronics in recent years. Besides, Peng Miao et al. also reported that ReS₂ possesses the SERS effect due to the charge transfer process between ReS₂ and the analytes [24]. Nevertheless, the Raman enhancement effect of ReS₂ is greatly inferior to that of the noble metal nanostructures, which limits its practical applications for the detection of aromatic molecules, toxic substances and biomolecules, etc. Therefore, it is of great significance to optimize the SERS effect of ReS₂. Jihyung Seo et al. optimized the electronic structure of ReS₂ and adjusted the surface polarity by doping different concentrations of oxygen to improve the dipole-dipole interactions and charge-transfer resonance between the ReS₂ and the analytes. Moreover, the femtomolar detection level of R6G was achieved by constructing a graphene/ReO_xS_y vertical heterostructure [25].

Though, these works promoted the application of ReS₂ in SERS, integrating the ReS₂ with metal nanostructures has not been investigated yet. Furthermore, compared with the SERS substrate with two-dimensional planar hot spots, the SERS substrates with three-dimensional and cavity structure possess its unique advantages. Zewen Zuo et al. developed a kind of silver conical cavity array [26]. The highly sensitive SERS substrate was obtained by effectively coupling the surface plasmons of nanoparticles with the cavity mode. The cavity structure can not only increase the surface area contributing to molecular adsorption, but also expand the SERS region from the surface area to the hollow area within the nanocavity [27]. The concept of volume-enhanced Raman scattering (VERS) was proposed by Xingang Zhang et al. And their hollow nanocones substrate greatly improved the Raman signal of single virus [28].

Currently, combining two materials with different properties into hybrid structure as SERS substrate is an effective method to improve the overall Raman enhancement performance [29–33]. In this work, ReS₂ nanocavity (ReS₂NC) decorated with silver nanoparticles (AgNP) is firstly synthesized on pyramid silicon (PSi) modified by gold nanoparticles (AuNP) using hydrothermal technique and thermal evaporation technology. By simply adjusting the thickness of silver, the SERS performance of the substrate (called PSi/Au/ReS₂NC/Ag) is optimized. Benefiting from large surface area and dense hot spots, PSi/Au/ReS₂NC/Ag shows excellent detection performance for dye molecules and can be used to in-situ monitor the conversion of p-nitrothiophenol (PNTP) to p,p'-dimercaptobenzene (DMAB) driven by plasmon. In addition, the temperature-responsive behavior of the substrates in SERS were first investigated. We believe this research will promote the application of metal-modified semiconductor substrates in SERS and deepen our understanding of its underlying properties.

2. Experimental

2.1 Fabrication of the PSi/Au/ReS₂NC/Ag

The preparation procedures of the PSi/Au/ReS₂NC/Ag are schematically shown in Fig. 1. Simply put, Au and Ag are deposited on the substrate by vacuum thermal evaporation, and ReS₂ is grown on the substrate by hydrothermal method. As we have reported previously, pyramid silicon (PSi) was fabricated by wet etching technology. In a typical experiment, several pieces of PSi samples were ultrasonically washed by acetone, ethanol and deionized water (DI water) for 15 min separately. The gold film with 7 nm thickness was deposited on clean PSi using thermal evaporation method, the sample was then put in a quartz tube and annealed for 30 min at 700°C. 26.8 mg of ammonium perrhenate (NH₄ReO₄), 20.8 mg of hydroxylamine hydrochloride (NH₂OH·HCl) and 36.3 mg of L-cysteine ware ultrasonically dispersed in 10 ml ultrapure water (UP water) well, and then transferred to a 50 ml polytetrafluoroethylene-lined autoclave, followed by the PSi modified by AuNPs was put into the above autoclave, sealed and stored at 200°C with different reaction time of 2 h, 4 h and 6 h. Let the sample cool to room temperature naturally, take it out and rinse it well with UP water. Afterwards, the silver was deposited on prepared PSi/Au/ReS₂NC by vacuum thermal evaporation technology. In addition, by adjusting the thermal deposition thickness from 5 nm to 25 nm, the PSi/Au/ReS₂NC/Ag with the optimal SERS performance was obtained.

 

Fig. 1. Schematic diagram of the fabrication of PSi/Au/ReS₂NC/Ag.

Download Full Size | PDF

2.2 Characterization

Scanning electron microscope (SEM, Zeiss Gemini Ultra-55), transmission electron microscope (TEM JEM-2100) and energy dispersive spectrometer (EDS) were used to characterize the morphology of the fabricated samples. The content of elements was also measured by X-ray photoelectron spectroscopy (XPS Thermo Scientific Escalab 250Xi). The SERS spectrum was collected by the Raman spectrometer (Horiba HR Evolution) with the excited laser of 532 nm to study the SERS performance of the prepared PSi/Au/ReS₂NC and PSi/Au/ReS₂NC/Ag SERS substrates. The 50 × objective lens was used. The integration time and the diffraction grating were set to 4 s and 600 gr/nm respectively. For the plasmon-driven, surface-catalyzed reactions, the PSi/Au/ReS₂NC and PSi/Au/ReS₂NC/Ag SERS substrates were immersed in 10⁻³ M PNTP solution for more than 1 h. Then, the samples were rinsed with ethanol several times and dried in N₂ condition before the Raman test. A temperature controller equipped with a liquid nitrogen tank was used to controllably adjust the temperature of the SERS substrates. The temperature change rate was set to 50 °C per minute. SERS data was collected every 10 °C change.

3. Results and discussion

3.1 Characterization of the PSi/Au/ReS₂NC/Ag structure

A mass of tightly arranged AuNP ware uniformly attached to the PSi surface (Supplement 1, Fig. S1A). The histograms of the size distribution of AuNP diameter with average 30 nm and interparticle gap with average 11 nm are also shown in Supplement 1, Figs. S1B and S1C respectively. From Fig. 2(A), we can see clearly that ReS₂ layer has been grown successfully along the PSi/Au surface. ReS₂ nanosheets with 10 nm wall thickness were vertically overlaid on the AuNP to form dense nanocavities with 60 nm cavity width. The inset in Fig. 2(A) show boundary of the PSi/Au coated or uncoated ReS₂ nanosheets. Our experiments indicated that the existence of AuNP has a positive effect on the growth of ReS₂ on PSi [34]. As shown in Supplement 1, Fig. S2, ReS₂ nanocavity is difficult to directly grow on PSi surface without AuNP, and form nanospheres made up of nanosheets scattered sporadically in the valley of PSi, which is attributed to the weak interaction between ReS₂ and Si. Figure 2(B) shows the SEM image of the PSi/Au/ReS₂NC/Ag, and the illustration in Fig. 2(B) exhibits the side view of Au/ReS₂NC/Ag where its thickness of is about 70 nm. To further confirm the successful preparation of PSi/Au/ReS₂NC/Ag hybrid, HRTEM was also carried as shown in Fig. 2(C). It could be clearly seen that AuNP marked with yellow line is attached to ReS₂ nanosheets, and the inter-layer spacing of 0.61 and 0.24 nm correspond to the (100) plane of ReS₂ nanosheets and the (111) plane of AuNP, separately. In addition, the Ag with 0.23 nm inter-layer spacing in agreement with the (111) plane is distributed at the interface of ReS₂ nanosheets as presented in Fig. 2(D). It is worth noting that the AuNP is located in the center of ReS₂ nanosheets and the Ag located in the edge of ReS₂ nanosheets, which is consistent well with our experimental expectations. Moreover, as shown in Fig. 2(E), the elements composition of the PSi/Au/ReS₂NC/Ag was also measured by EDS elemental mappings, which clearly shows the presence of Au (cyan), Re (orange), S (green) and Ag (purplish red). In order to make a comparison, Si element mapping was also put into Fig. 2(E). Si element is mainly distributed in the valley of pyramid, while Au, Re, S and Ag elements are mainly distributed in the inclined plane and top of pyramid, which is related to the sparse distribution density of AuNP in the valley areas. This phenomenon is consistent with the experimental results discussed previously.

 

Fig. 2. (A) SEM image of PSi/Au/ReS₂NC. (B) SEM image of PSi/Au/ReS₂NC/Ag. (C) and (D) HRTEM image of PSi/Au/ReS₂NC/Ag. (E) EDS elemental mappings from Au, Re, S, Ag and Si on the PSi/Au/ReS₂NC/Ag.

Download Full Size | PDF

The elemental constituent and chemical states of the fabricated substrates ware confirmed by using X-ray photoelectron spectroscopy (XPS). It can be clearly observed that the two characteristic peaks of PSi/Au/ReS₂NC substrate are 42.47 and 44.85 eV in Fig. 3(A), corresponding to the level peaks of Re⁴⁺ 4f_7/2 and Re⁴⁺ 4f_5/2, respectively. It is interesting to note that the binding energies of 4f_7/2 and 4f_5/2 peaks of Re⁴⁺ in PSi/Au/ReS₂NC/Ag are negative shifted by 0.1 eV. In addition, the two Ag 3d peaks at 368.79 and 374.80 eV in the PSi/Au/ReS₂NC/Ag are also observed as shown in Fig. 3(B), corresponding to the level peaks of 3d_5/2 and 3d_3/2, separately. However, compared with the bulk Ag, the binding energies of the two Ag 3d have positively shifts around 0.29 eV [35]. The above results indicate that the electrons transfer from Ag to ReS₂ occurs. The existence of charge transfer between ReS₂ and Ag is suitable for improving the chemical enhancement and catalytic capability of SERS substrate. The high-resolution S 2p spectrum of the PSi/Au/ReS₂NC and PSi/Au/ReS₂NC/Ag substrate are shown in Supplement 1, Fig. S3, and the atomic percentage of Re and S is close to 1:2 (Supplement 1, Fig. S4), which confirms the elemental constituent and chemical states of the ReS₂. Figure 3(C) depicts the Raman signals of ReS₂ from the PSi/Au/ReS₂NC and PSi/Au/ReS₂NC/Ag. Because of its low symmetry, the Raman spectrum of ReS₂ display more characteristic peaks than other TMDs. The basic Raman modes (E_2g, A_1g and E_1g) are coupled with each other and acoustic phonons. The Raman bands at 157, 163 and 316 cm^-1 (Red curves) refer to the in-plane vibration, while the Raman band at 213 cm^-1 (Green curve) refers to the mostly out-of-plane vibration. The scanning Raman mapping of E_2g mode of the PSi/Au/ReS₂NC is also exhibited in Fig. 3(D). The low color difference indicates well uniformity of ReS₂.

 

Fig. 3. (A) XPS spectra of Re of PSi/Au/ReS₂NC and PSi/Au/ReS₂NC/Ag sample. (B) XPS spectra of Ag of PSi/Au/ReS₂NC/Ag sample. (C) The Raman spectra of PSi/Au/ReS₂NC and PSi/Au/ReS₂NC/Ag sample. (D) The scanning Raman mapping of E_2g mode of the PSi/Au/ReS₂NC sample.

Download Full Size | PDF

3.2 Optimization of SERS performance of the PSi/Au/ReS₂NC/Ag

As shown in Supplement 1, Fig. S5, Ag of changing from 5 to 25 nm was steamed onto the PSi/Au/ReS₂NC to optimize the SERS response of the PSi/Au/ReS₂NC/Ag. An enlarged view of the PSi/Au/ReS₂NC is displayed in Supplement 1, Fig. S5A. When the deposition thickness is 5 nm, the Ag distribut granularly and sparsely on the ReS₂ nanosheets (Supplement 1, Fig. S5B). When the deposition thickness reaches 10 nm, the diameter of the AgNP becomes larger (Supplement 1, Fig. S5C). By further increasing the deposition thickness to 15 nm, the diameter of the AgNP continues to increase, and they are densely and uniformly adsorbed on the ReS₂ nanosheets (Supplement 1, Fig. S5D). As the deposition thickness increased to 20 nm, AgNP began to contact each other on the ReS₂ nanosheets to form a continuous film, and many protrusions were produced at the same time (Supplement 1, Fig. S5E). The Ag film becomes thicker and completely and uniformly covers the ReS₂ nanosheet when the deposition thickness continues to increase to 25 nm (Supplement 1, Fig. S5F).

In order to assess the SERS performance of the above substrates, Rhodamine 6G (R6G) was chosen as the analytes to conduct SERS contrast detection. The observed vibrational bands from Fig. 4(A) are well matched with the previously reported works. The peak at 614 cm^-1 is attributed to the in-plane vibration of C-C-C deformation, the peak at 776 cm^-1 refers to the out-of-plane vibration of C-H bond deformation, and the peak at 1650 cm^-1 applies to the aromatic C-C stretching [36]. Figure 4(B) demonstrates the line graph of the peak intensities of the characteristic bands at 614 and 776 cm^-1 with the Ag thickness from 5 nm to 25 nm. With the increase of Ag thickness, the SERS signals of R6G molecules (10⁻⁵ M) was enhanced gradually and reached maximum at the 15 nm thickness. Note that when the Ag thickness continued to increase, the SERS signal strength declined instead. This is because the over-thick Ag completely covers the ReS₂ layer to form a flat film, resulting in a reduction in the density of hot spots.

 

Fig. 4. (A) SERS spectrum of R6G (10⁻⁵ M) collected on the PSi/Au/ReS₂NC/Ag with varying deposition thickness of Ag. (B) The broken line graphs of SERS intensities at 614 and 776 cm^-1 with Ag deposition thickness. (C)-(G) Simulated vertical electric field distributions of the ReS₂NC deposited different Ag thickness of (C) 5 nm, (D) 10 nm, (E) 15 nm, (F) 20 nm, (G) 25 nm. (H) The simulated enhancement factor (|E|²/|E₀|²) of the different hybrid structures.

Download Full Size | PDF

3.3 SERS enhanced mechanism of the PSi/Au/ReS₂NC/Ag system

To deeply understand the variation of the SERS intensity changes with the Ag thickness, the electric field distribution of different SERS substrates was simulated by commercial software COMSOL. Due to the thick ReS₂ layer, the electric field enhancement generated by the AuNP was ignored in the simulation. The refractive index of ReS₂ was set as n(ReS₂) = 3.99 + 0.17i under 532 nm laser irradiation [37]. According to the SEM images, the thickness of the ReS₂ nanosheets was set as 10 nm, the interval between the nanosheets was set as 60 nm, the diameter of the AgNP was set as 20 nm, 22 nm, 25 nm, and the thickness of the Ag film was set as 14 and 16 nm, respectively. As the particle size increases and the gap between particles decreases, the electric field intensity increases significantly, and reaches a maximum at 15 nm (Fig. 4(C)-(E)). The electric field is primarily distributed in the cavity and around the nanoparticles because of the coupling of the surface plasmons of the AgNP and the cavity mode. However, the number of hot spots is drastically reduced due to the formation of the silver film, resulting in a decrease in electric field intensity (Fig. 4(F) and (G)). Figure 4(H) shows the change trend of the enhancement factors (|E|²/|E₀|²) of the above five samples, which is consistent with the previous experimental results. Besides, ReS₂ nanosheets can be used as a superior adsorbent for analytes, which is helpful to improve the SERS response of the PSi/Au/ReS₂NC/Ag.

Apart from electromagnetic enhancement, the chemical enhancement of ReS₂ caused by charge transfer should also be considered in our samples as shown Fig. 5. It was reported that the conduction band minimum (CBM) and valence band maximum (VBM) of monolayer ReS₂ are -6.15 and -4.26 eV, respectively [17]. And the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) levels of R6G are -5.7 and -3.4 eV, separately [38]. The photoexcited electrons induced by surface plasmon resonance of AgNP were injected into the CB of ReS₂ when the visible laser was irradiated on the substrate, then transferred to LUMO of the probe molecules. Furthermore, from an energy point of view, the establishment of Schottky barrier between AgNP and ReS₂ facilitates the transport of the photoexcited electrons from AgNP to ReS₂.

 

Fig. 5. Schematic diagram of the proposed charge transfer mechanism.

Download Full Size | PDF

3.4 Surface catalytic reaction performance of the composite structure

Some chemical reactions can be driven by hot electrons induced by SPR on the surface of coinage metal nanostructures [39–43]. Up to present, the catalytic reactions driven by plasmon have been extremely critical in environmental protection, solar energy, and chemical industry for low-energy requirements and high efficiency [44,45]. Under normal circumstances, the extremely short hot electron lifetime greatly bounds the efficiency of hot electron-induced reduction reactions owing to the fast decay of SPR. The combination of metals and semiconductors materials is an effective method to transfer and extend the lifetime of hot electrons. Therefore, the PSi/Au/ReS₂NC/Ag SERS substrates were used to in-situ monitor the reduction reaction from PNTP to DMAB and verify the catalytic ability. The SERS spectra of Ag free substrate only shows the 1340 cm^-1 peak, representing the ${V_{({N{O_2}} )\; }}$ of PNTP, which is related to the thicker ReS₂ layer, as shown in Fig. 6(A). For the sample with a deposition thickness of 5 nm of silver, the peaks at 1397 and 1445 cm^-1 representing the ${V_{({N = N} )\; }}$ of DMAB molecules appeared after 10 seconds of reaction (Fig. 6(B)). However, the intensity of the DMAB band is weak compared to the intensity of the PNTP band, indicating poor conversion efficiency. Compared with the ${V_{({N{O_2}} )\; }}$ band, the intensity of the ${V_{({N = N} )\; }}$ bands are much stronger after the reaction for 10 s when the deposited thickness of Ag increased to 10 nm or more, indicating high conversion efficiency (Fig. 6(C)-(F)). As shown in Fig. 6(G), the catalytic efficiency of the substrates was evaluated using the time-dependent formula of ${I_{{V_{({N = N} )}}}}/{I_{{V_{({N{O_2}} )}}}}$, where ${I_{{V_{({N = N} )}}}}$ represents the SERS intensity of DMAB band at 1445 cm^-1, while ${I_{{V_{({N{O_2}} )}}}}$ represents the SERS intensity of PNTP band at 1340 cm^-1. The results show that the reaction rate is the fastest when the deposited thickness of Ag reaches 15 nm. It is worth noting that the PSi/Au/ReS₂NC/Ag with higher intensity of hot spots is expected to generate more “hot electrons” to accelerate this reaction. However, too much Ag deposited on the ReS₂ nanosheets would act as a barrier layer which prevents the molecules from contacting ReS₂, resulting the electron and hole recombination more easily, and ultimately reducing the reaction efficiency. In addition, the SERS strength of ${V_{({N = N} )\; }}$ band at 1445 cm^-1 after 60 s reaction collected from PSi/Au/ReS₂NC/Ag with different deposited thickness of Ag are exhibited in Fig. 6(H). This change trend is consistent with the SERS spectrum of R6G and the simulated results.

 

Fig. 6. (A)-(F) In-situ SERS monitoring the dimerization of PNTP to DMAB for PSi/Au/ReS₂NC/Ag with different deposited thickness of Ag. (G) Time dependence of ${I_{{V_{({N = N} )}}}}/{I_{{V_{({N{O_2}} )}}}}$ for PSi/Au/ReS₂NC/Ag with different Ag thickness under 532 nm laser excitation. (H) Ag thickness-dependent SERS intensities of the ${V_{({N = N} )\; }}$ band.

Download Full Size | PDF

3.5 Investigation of SERS activity, signal homogeneity and substrate repeatability

For the sake of detection of the SERS sensitivity of the PSi/Au/ReS₂NC/Ag with 15 nm Ag thickness, the SERS spectrum of R6G aqueous solution with changing concentrations (10⁻⁵ - 10⁻¹¹ M) was measured (Fig. 7(A)). Although the SERS intensity of R6G was reduced with the decreasing concentration, SERS signal of R6G could still be detected when the R6G concentration is as low as 10⁻¹¹ M. Besides, as shown in Fig. 7(B), Raman characteristic peaks of 614 and 776 cm^-1 were chosen for linear fitting to study the correlation between SERS signal intensity and R6G concentration. The value of fitting correlation coefficient (R²) of 614 and 776 cm^-1 reached 0.986. The fitting formula of intensity change with concentration is as follows: $\textrm{Log}\; I = A\; \textrm{Log}\; C + B$, where I is the SERS intensity of the probe molecules, A refers to the slope, C refers to the concentrations and B refers to the intercept. For 614 and 776 cm^-1 peaks, the fitting formulas are $\textrm{Log}\; I = 0.358\; \textrm{Log}\; C + 5.901$ and $\textrm{Log}\; I = 0.360\; \textrm{Log}\; C + 5.510$, respectively. Excellent linear relationship under the log scale enables quantitative detection of dye molecules. As a comparison, the SERS spectrum of R6G aqueous solution with various concentrations (10⁻⁴ - 10⁻⁷ M) were also measured in PSi/Au/ReS₂NC (Supplement 1, Fig. S6). Even if gold does not play a role in the direct Raman enhancement, the substrate still shows a good ability to detect low concentration molecules. In addition to the charge transfer effect, the nanocavity has a unique cavity structure, which can effectively capture light by concentrated electric field and improve the light capture ability of the substrate. The decrease of detection line in PSi/Au/ReS₂NC/Ag could be explained by the synergistic effect of physical enhancement and chemical enhancement, as mentioned above. Figure 7(C) shows the Raman mapping of R6G molecules at 614 cm^-1 from the same sample. The relatively uniform color distribution in maps indicates that the homogeneity of the sample is relatively good. In addition to sensitivity and uniformity, repeatability is also an indispensable indicator of the performance of SERS substrates. Figure 7(D) shows the SERS spectrum of R6G collected from ten samples obtained using the same preparation method. For the sake of clarity, as shown in Fig. 7(E) and (F), the 614 and 776 cm^-1 peaks were chosen to study the repeatability of SERS spectra, and the relative standard deviation (RSD) could be calculated using the followed equation:

 

Fig. 7. (A) SERS spectrum of R6G (10⁻⁵ - 10⁻¹¹ M) in PSi/Au/ReS₂NC/Ag. (B) The relation between the intensity of SERS signals and the concentration of R6G at 614 and 776 cm⁻¹ in logarithmic coordinates. (C) The scanning Raman mappings of characteristic peak of R6G (10⁻⁶ M) at 614 cm^-1. (D) SERS spectrum of R6G (10⁻⁶ M) collected on different substrates. SERS intensity at 614 cm^-1 (E) and 776 cm^-1 (F) obtained from Raman spectra in (D).

Download Full Size | PDF

3.6 Temperature response behavior of SERS properties of composite substrates

The regulation of hot spots is an important research direction in the field of SERS. Previous studies have shown that the morphology of ReS₂ nanosheets changes with ambient temperature [46]. For the extremely weak interlayer interaction of ReS₂, out-of-plane bending would occur in the ReS₂ layers and the bending could recover to initial state as the temperature drops. In this work, the SERS signals of PSi/Au/ReS₂NC and PSi/Au/ReS₂NC/Ag were investigated by changing the test temperature. As shown in Fig. 8(A) and (B), the SERS signals collected from the PSi/Au/ReS₂NC gradually decreased when the temperature was elevated from room temperature to 85 °C. While for the PSi/Au/ReS₂NC/Ag, the SERS intensity gradually increased when the temperature rose from room temperature and reached maximum at 45 °C shown in Fig. 8(C) and (D). Subsequently, the SERS intensity decreased as the temperature increase to 85 °C. Interestingly, the SERS intensity increased again when the temperature decreased, and also reached the maximum when the temperature decreased to 45 °C. This experimental phenomenon can be expounded by the out-of-plane bending of ReS₂ nanosheets with temperature changes as shown in Fig. 8(E). When the temperature rises, the ReS₂ nanosheets begin to tilt from the vertical state. For the PSi/Au/ReS₂NC, the analyte molecules absorbed on the bottom of the cavity could not be detected resulting in continuously decreased SERS signals as the temperature increases. While for the PSi/Au/ReS₂NC/Ag, the gap among nanosheets would decrease as the temperature increases, which is beneficial for the elevation of the hot spot intensity bringing advantages to the SERS. When the temperature further heightens, the nanocavities would disappear, and thus resulting in a weak SERS response.

 

Fig. 8. (A) SERS spectrum of R6G (10⁻⁵) in PSi/Au/ReS₂NC at 25 - 85 °C. (B) The broken line graphs of SERS signal intensity at 614 and 776 cm^-1 with temperature in PSi/Au/ReS₂NC. (C) SERS spectrum of R6G (10⁻⁵) in PSi/Au/ReS₂NC/Ag at 25 - 85 °C. (D) The broken line graphs of SERS signal intensity at 614 and 776 cm^-1 with temperature in PSi/Au/ReS₂NC/Ag. (E) Schematic diagram of the morphology of ReS₂ nanosheets changing with the temperature.

Download Full Size | PDF

3.7 Verification of detection ability for real samples of the PSi/Au/ReS₂NC/Ag

Malachite Green (MG), a toxic chemical of triphenylmethane, is an effective fungicide, which has been widely used in aquaculture. However, chronic overuse of MG is harmful to human health and has carcinogenic effect. Therefore, many countries have banned the use of MG in the field of pollution-free aquaculture. In order to verify practicability of PSi/Au/ReS₂NC/Ag substrate, we collected water from the local lake and used it as a solvent to prepare MG solutions with different concentrations. As shown in Fig. 9(A), the detection limit is as low as 10⁻⁹ M, reaching European and American standards. Similarly, the Raman characteristic peak at 1620 cm^-1 is chosen to further survey the relationship between signal intensity and concentration (Fig. 9(B)). The fitting formula is $\textrm{Log}\; I = 0.385\; \textrm{Log}\; C + 6.131$, and R² reaches 0.985, which could be used for quantitative detection of real samples. As a comparison, SERS signals of MG (10⁻⁸ M) in two different solvent of DI water and lake water were collected as shown in inset of Fig. 9(B). The SERS intensity of MG in lake water is weaker than that in DI water, which may be due to a large number of other impurities weakening the signal strength. Crystal Violet (CV), another effective bactericide, also has carcinogenic effect and is forbidden to be used in edible products. In likewise, lake water was also used as solvent to prepare CV solution for SERS measurement. Figure 9(C) exhibits the SERS spectrum of CV at different concentration. The fitting formula of intensity changing with concentration at 915 cm^-1 is $\textrm{Log}\; I = 0.461\; \textrm{Log}\; C + 6.352$. And the R² is 0.992, which can realize quantitative detection (Fig. 9(D)). As shown in the inset of Fig. 9(D), we again compared the SERS spectra of CV with the same concentration (10⁻⁸ M) in lake water and DI water. The result is similar to that of MG. Therefore, PSi/Au/ReS₂NC/Ag sensor shows a potential ability to detect harmful substances in aquatic products.

 

Fig. 9. (A) SERS spectrum of MG (10⁻⁴ - 10⁻⁹ M) in lake water solution collected on PSi/Au/ReS₂NC/Ag. (B) The relation between the intensity of SERS signals and the concentration of MG at 1620 cm⁻¹ in logarithmic coordinates. (C) SERS spectrum of CV (10⁻⁶ - 10⁻¹⁰ M) in lake water solution collected on PSi/Au/ReS₂NC/Ag. (D) The relation between the intensity of SERS signals and the concentration of CV at 915 cm⁻¹ in logarithmic coordinates.

Download Full Size | PDF

4. Conclusions

In short, we developed a three-dimensional SERS substrate based on ReS₂ nanocavity by using thermal evaporation and hydrothermal synthesis technology, and studied the SERS enhancement mechanism in detail, including physical and chemical mechanisms. The PSi/Au/ReS₂NC/Ag exhibits good sensitivity, uniformity and repeatability for detection of R6G, and can be used to in-situ monitor the reaction for PNTP to DMAB conversion driven by plasmon. Moreover, in this study, it is found for the first time that the SERS performance of ReS₂ nanocavity-based substrates show a strong temperature correlation, and the SERS effect achieve the best performance at 45 °C. We believe this result can be used as a new thought for the designing 2D materials/metal composite structures used in SERS effect with putting the thermal deformation as a vital factor. For actual application, MG and CV molecules in lake water were successfully detected, showing its development potential in the field of food safety.