Wet Chemical Synthesis and Characterization of Au Coatings on Meso-and Macroporous Si for Molecular Analysis by SERS Spectroscopy

: Porous silicon (PS) is a promising material for nanostructure fabrication providing a precise control over its size, shape, and spatial distribution. This makes it an excellent candidate for constructing highly sensitive, reproducible, and low-cost platforms for surface enhanced Raman scattering (SERS) spectroscopy. In this work, we connect the PS structural parameters with the morphology of the gold nanostructures fabricated on its surface, placing the emphasis on the SERS response. Two different types of PS are considered here, namely meso-and macro-PS. The former is prepared by Si electrochemical etching, applying three different current densities: 100 mA cm − 2 , 60 mA cm − 2 , and 30 mA cm − 2 , while the technological parameters for the latter are selected to mimic metal nanovoids’ (Me NVs) conﬁguration. The gold-coated PS surfaces are produced via an electroless chemical deposition method for different time periods. By performing comprehensive structural, morphological, and optical characterization, we show the importance of the size and density of the PS pore openings, which govern the Au growth kinetics. The results presented in this work assure a simple yet ﬂexible approach for the fabrication of large-area plasmonic gold nanostructures, which are not only suitable for advanced SERS spectroscopy studies but can also serve for a wider range of plasmonic applications.


Introduction
Nowadays, the rapid development of biomedicine has been increasingly faced with a demand for the detection, identification, and structural study of different organic analytes in trace amounts. Several sensing techniques, such as X-ray crystallography [1], nuclear magnetic resonance [2,3], and circular dichroism spectroscopy [4], have been applied to meet these needs. However, their routine application outside the framework of laboratory usage remains challenging: the low speed, high cost, complexity, and lack of complete databases [5] are major limitations for their practical implementation.
On the other hand, vibrational spectroscopy is a well-established molecular characterization technique [6], which does not require complex sample preparation and is widely used to study organic substances, e.g., in medicine [7], biology [8,9], the food industry [10], and pharmaceutics [11][12][13], to name a few. Utilizing the latest advancements of the Surface Enhanced Raman Scattering (SERS) spectroscopy technique, it is also possible to perform the detailed structure assignment of intricate substances, such as peptides, in small concentrations [14]. This is because of the extremely high sensitivity of the SERS technique described elsewhere [43]. Briefly, an n + -and p − -Si wafer was used to grow meso-and macro-PS, respectively. Just before anodization, the native silicon dioxide was removed from a silicon sample by etching in a diluted hydrofluoric acid (4.5%) for 60 s. After that, the Si sample was placed in a horizontal electrolytic cell made of polytetrafluoroethylene. An aqueous-alcoholic solution made of HF (45%), deionized water, and isopropanol at a volume ratio of 1:3:1 was used to prepare the meso-PS. Dimethylsulfoxide (DMSO) mixed with HF (45%) at the volume ratio of 10:46 was used for the macro-PS fabrication. The prepared solutions were gently added to the electrolytic cell. The meso-PS samples were fabricated at 100 mA cm −2 , 60 mA cm −2 , and 30 mA cm −2 current densities applied for 20 s, 22 s, and 20 s, respectively. The macro-PS samples were produced at an 8 mA cm −2 current density applied for 4 min. All freshly prepared PS samples were rinsed with deionized water and air dried.

Gold Deposition and Samples' Characterization
Gold was deposited by the immersion of a PS sample in an aqueous solution of 0.01 M KAu(CN) 2 and 0.15 M HF for different time periods according to the slightly changed procedure reported before [44]. Subsequently, freshly prepared samples were carefully washed with deionized water and air dried.
The organic dye rhodamine 6G (R6G) was used as an analyte for the SERS study. The dye molecules were adsorbed on the gold-coated PS sample by dipping it in a 10 −6 M aqueous solution for 2 h. Each substrate was then thoroughly rinsed with deionized water for 30 s to remove the excess analyte molecules.
The morphology of the gold-coated PS was studied by scanning electron microscopy (SEM) with the help of an FESEM Hitachi−S4800 microscope. The crystallinity of the samples was studied with X-ray diffraction (XRD) using Cu K α radiation (X-ray wavelength λ = 0.15406 nm). The SEM images were analyzed using ImageJ software. Reflectance spectra measurements were performed in the 350-900 nm wavelength range with the help of a Shimadzu UV-3600 equipped with an integrating sphere module (ISR-2200). SERS spectra were acquired with the help of a WITec Scanning Near-field Optical Microscope alpha300 S equipped with a 532 nm laser diode and an 60x air objective (0.9 NA). The spectra were acquired with 1 mW laser power for 1 s sample surface exposition.

Morphology of the Au-Coated Meso-PS Substrates
The electroless (galvanic, corrosion) deposition of the Au films on monocrystalline Si surfaces followed the oxidation-reduction (redox) mechanism, which assumes the reduction of Au 3+ cations to atomic Au 0 by electron (e -) exchange with Si. It is known that the redox potential for Au 3+ /Au 0 is positive in the presence of Si atoms [45]. Hence, the Au deposition proceeded spontaneously as the result of the electron transfer from semiconductor valence band to metal cations: As the result of this process, the Si was simultaneously oxidized, since the above reduction reaction can be also viewed in terms of the holes (h + ) "injection" into the valence band of Si. This injection in the presence of H 2 O induced silicon dioxide formation: On a large timescale, the gold deposition is limited by the passivation of the Si surface with the oxide, reducing the process rate. However, with the presence of HF in the solution, the SiO 2 can be dissolved, producing new charge carriers and water soluble SiF 2-6 species: In these conditions, the Au electroless deposition on the Si is not limited by SiO 2 growth, and new Au nuclei form on the already adsorbed metal clusters and the available Si surface. Notably, the former is possible due to the galvanic cell formation between the noble metals and Si [46].
Considering a porous instead of monocrystalline Si surface, an important role for Au deposition is the presence of Si-H x groups, which are usually generated via the hydration of uncompensated Si bonds produced upon electrochemical Si etching. These hydrated bonds are very reactive and can be easily oxidized, releasing additional electrons, and thus accelerating Au deposition: 2Si Therefore, porous Si substrates can act not only as shape-forming platforms for plasmonic nanostructures due to the diversity of their surface configurations (meso-, macro-, and microporous surfaces of different porosity), but they also can control the redox reaction kinetics, which is an intriguing approach for producing even more sophisticated nanostructures, such as noble metal dendrites [47] or nanothorns [37]. Figure 1 shows the SEM images for the meso-PS templates prepared by different electrochemical etching regimes. Figure 1a    The fabricated meso-PS matrices were utilized for the following gold deposition step. Figure 2 represents the SEM images for different Au-coated meso-PS substrates. The first row shows the samples obtained after 30 min of Au deposition, while the second row represents the samples after 70 min of Au deposition. Each column in the figure corresponds to the gold nanostructure on the meso-PS etched at different current densities: 30 mA cm −2 , 60 mA cm −2 , and 100 mA cm −2 . The increase in the Au deposition time led to the gold layer densification, producing a void-free continuous film along with a second Au layer nucleation (observe the material excess on top of the formed layer in Figure 2a,d). The same morphology evolution of a gold layer was obtained when the porosity of the PS was reduced, while keeping the Au deposition time constant (Figure 2a-c). Notably, no significant layer alteration was revealed when the porosity of PS varied at higher Au deposition times (Figure 2d-f). The analysis of the SEM images indicated that the gold coverage level increased from ~83% to ~97% after 30 min and 70 min of Au deposition, respec- The quantitative characterization of the PS was carried out with the calculation of the mean pore opening diameter, pore density, and porosity. The mean pore diameter was calculated by averaging a Feret diameter distribution [48]. The pore density was calculated by dividing the number of pore openings by the visible surface area. The porosity of the meso-PS samples was estimated by gravimetric [49] and SEM image assessment methods. Both approaches resulted in the same porosities. When the applied etching current density was 30 mA cm −2 , we obtained an average pore diameter of 19 nm ± 8 nm, a pore density of 5.3 × 10 10 cm −2 , and 49% porosity. For the sample etched at 60 mA cm −2 , we obtained an average pore diameter of 28 nm ± 14 nm, a pore density of 3.1 × 10 10 cm −2 , and 59% porosity. For the highest etching current density of 100 mA cm −2 , we obtained an average pore diameter of 44 nm ± 32 nm, a pore density of 2.8 × 10 10 cm −2 , and 72% porosity. These data are in good agreement with the well-studied mechanism of PS formation [50,51] and reported experimental observations [52,53], albeit the density of the pores slightly decreased for the more porous samples. Based on the obtained results, we can attribute the fabricated samples to the mesoporous material [54].
The fabricated meso-PS matrices were utilized for the following gold deposition step. Figure 2 represents the SEM images for different Au-coated meso-PS substrates. The first row shows the samples obtained after 30 min of Au deposition, while the second row represents the samples after 70 min of Au deposition. Each column in the figure corresponds to the gold nanostructure on the meso-PS etched at different current densities: 30 mA cm −2 , 60 mA cm −2 , and 100 mA cm −2 . The increase in the Au deposition time led to the gold layer densification, producing a void-free continuous film along with a second Au layer nucleation (observe the material excess on top of the formed layer in Figure 2a,d). The same morphology evolution of a gold layer was obtained when the porosity of the PS was reduced, while keeping the Au deposition time constant (Figure 2a-c). Notably, no significant layer alteration was revealed when the porosity of PS varied at higher Au deposition times (Figure 2d-f). The analysis of the SEM images indicated that the gold coverage level increased from~83% to~97% after 30 min and 70 min of Au deposition, respectively, and did not alter significantly for different PS matrixes. Although the produced gold coverage represented a rather quasicontinuous film, there was still a meaningful number of Au nanoparticles (NPs). Hence, we estimated their mean diameter and standard deviation, which were varied within the 22-60 nm range for PS matrices after 30 min Au deposition, while the 70 min Au deposition produced 31-46 nm NPs ( Figure 3).
These results reveal the importance of the meso-PS matrix morphology for the accurate control over the Au layer fabrication process. The obtained Au layer was defined by a PS surface, which contained an array of densely packed voids on the silicon surface ( Figure 1a-c). Their density primarily depended on the etching current density and could be characterized by the porosity of the PS. As a rule, the higher the porosity, the denser the void package on the surface, and the lower the "free" silicon area. Consequently, the gold deposition rate, which followed the redox reaction mechanism between the silicon atoms and gold cations, was dramatically reduced for the highly porous matrices. The number of seed nucleation centers, as well as the surface diffusion of the gold atoms, were restricted by the voids, preventing the growth of a continuous Au film. Therefore, the Au coverage appeared in the form of isolated islands, whose sizes strongly depended on the porosity of the parent meso-PS matrix. It should be emphasized, however, that the situation changed in the later stages of Au film growth (Figure 2d-f). In these stages, the produced Au nanoislands became new nucleation centers, leading to the film growth over the pore openings. Eventually, the Au nanoislands collapsed, and the quasicontinuous Au film on top of the porous surface was achieved. Continuing the Au deposition induced the new Au layer formation. Interestingly, the appearance of this layer started with NP nucleation, suggesting that the Au film formation followed the Stranski-Krastanov mechanism of thin films' growth [55][56][57].  These results reveal the importance of the meso-PS matrix morphology for the accurate control over the Au layer fabrication process. The obtained Au layer was defined by a PS surface, which contained an array of densely packed voids on the silicon surface (Figure 1a-c). Their density primarily depended on the etching current density and could be characterized by the porosity of the PS. As a rule, the higher the porosity, the denser the void package on the surface, and the lower the "free" silicon area. Consequently, the   These results reveal the importance of the meso-PS matrix morphology for the accurate control over the Au layer fabrication process. The obtained Au layer was defined by a PS surface, which contained an array of densely packed voids on the silicon surface (Figure 1a-c). Their density primarily depended on the etching current density and could be characterized by the porosity of the PS. As a rule, the higher the porosity, the denser the void package on the surface, and the lower the "free" silicon area. Consequently, the  These data indicate the gold coverage growth along with the deformation of a Au elementary crystal cell by a Si surface with (100) crystallographic orientation, leading to the monocrystallization of the Au coverage. The appearance of the Si (400) peak, however, suggested a weak uniformity of the Au layer for both samples. The same XRD patterns were observed for all the Aucoated meso-PS samples, albeit with a minor intensity variation of the characteristic peaks, depending on the porosity of the meso-PS and the Au deposition time.
nucleation, suggesting that the Au film formation followed the Stranski-Krastanov mechanism of thin films' growth [55][56][57]. Figure 4 reports the XRD pattern for the Au-coated meso-PS after 30 min and 70 min of Au deposition on the PS etched at 60 mA cm −2 . As can be seen from Figure 4, both XRD patterns contained the characteristic peaks for the cubic cell Au crystals with orientations (111), (200), (220), and (311), confirming the polycrystalline nature of the Au coverage. Remarkably, the intensity of the (200) peak increased after 70 min of Au deposition, while the other peaks became less prominent, compare Figure 4a,b. These data indicate the gold coverage growth along with the deformation of a Au elementary crystal cell by a Si surface with (100) crystallographic orientation, leading to the monocrystallization of the Au coverage. The appearance of the Si (400) peak, however, suggested a weak uniformity of the Au layer for both samples. The same XRD patterns were observed for all the Au-coated meso-PS samples, albeit with a minor intensity variation of the characteristic peaks, depending on the porosity of the meso-PS and the Au deposition time.   Figure 5 reports the SEM images for the macro-PS template and the corresponding distribution of the pore openings' diameter. The SEM images confirmed the 1.1 μm thick PS layer formation with round pore openings (Figure 5a). The pore diameters' distribution for the macro-PS was rather nonuniform, covering the 400-1000 nm range (Figure 5b). These sizes were about one order of magnitude larger compared to the meso-PS ( Figure  1). We therefore attributed the obtained sample to the macroporous material [54]. These results were in good agreement with the established macro-PS formation models [51].  Figure 5 reports the SEM images for the macro-PS template and the corresponding distribution of the pore openings' diameter. The SEM images confirmed the 1.1 µm thick PS layer formation with round pore openings (Figure 5a). The pore diameters' distribution for the macro-PS was rather nonuniform, covering the 400-1000 nm range (Figure 5b). These sizes were about one order of magnitude larger compared to the meso-PS (Figure 1). We therefore attributed the obtained sample to the macroporous material [54]. These results were in good agreement with the established macro-PS formation models [51]. Figure 6 shows the SEM images for the Au-coated macro-PS substrates after 30-80 min of Au deposition. The morphology of the film altered dramatically at various Au deposition steps. An array of closely packed Au NPs was obtained after 30 min deposition, Figure 6a. These NPs coalesced after 40 min producing a quasicontinuous gold film, as shown in Figure 6b. The subsequent increase in the deposition time led to the film densification together with the new Au seeds' nucleation over the formed coverage, producing a multilayered Au film, as shown in Figure 6c-f. The analysis of the SEM images depicted in Figure 7 reported the average Au NP diameter after different deposition times. No noticeable dependence of NP size on the deposition time was observed. The average Au NP diameter for all samples varied from 50 nm to 60 nm. The statistical data for the samples after 30 min and 40 min of Au deposition is not reported due to the difficulties in distinguishing individual Au NPs.   Figure 6 shows the SEM images for the Au-coated macro-PS substrates after 30-80 min of Au deposition. The morphology of the film altered dramatically at various Au deposition steps. An array of closely packed Au NPs was obtained after 30 min deposition, Figure 6a. These NPs coalesced after 40 min producing a quasicontinuous gold film, as shown in Figure 6b. The subsequent increase in the deposition time led to the film densification together with the new Au seeds' nucleation over the formed coverage, producing a multilayered Au film, as shown in Figure 6c-f. The analysis of the SEM images depicted in Figure 7 reported the average Au NP diameter after different deposition times. No noticeable dependence of NP size on the deposition time was observed. The average Au NP diameter for all samples varied from 50 nm to 60 nm. The statistical data for the samples after 30 min and 40 min of Au deposition is not reported due to the difficulties in distinguishing individual Au NPs.

Morphology of the Au-Coated Macro-PS Substrates
The observed kinetics of the Au film growth may be explained considering the presence of a thin microporous layer [58][59][60] over the entire macro-PS surface. This layer had a fractal-like geometry, which acted as the main source of the charge carriers for the redox reaction between the gold cations and the silicon wafer. This layer was also highly porous and composed of dense tiny voids. The growth of a continuous Au film on such a surface was even more confined by the voids compared with the meso-PS, leading to the formation of a dense Au NP layer in an early deposition stage ( Figure 6a). As the redox reaction progressed with time, the thin microporous layer dissolved, leaving the bare macro-PS surface opened. At this growth stage, the number of Au seed nucleation centers increased, and the surface diffusion of gold atoms became possible. As the result, the Au NPs coalesced, and the quasicontinuous film was produced. For the rest of the deposition time, the gold layer grew similarly to the Au film on the meso-PS.       (Figure 4a,b), no characteristic Si peak was observed in the XRD pattern for the gold-coated macro-PS, as shown in Figure 8a. This confirmed the introduced mechanism of Au film growth, where the coverage became uniform after the complete dissolution of the micro-PS layer. Apart from that, no noticeable difference between the Au-coated meso-and macro-PS samples was observed from the XRD; compare Figures 8b and 4a  The observed kinetics of the Au film growth may be explained considering the presence of a thin microporous layer [58][59][60] over the entire macro-PS surface. This layer had a fractal-like geometry, which acted as the main source of the charge carriers for the redox reaction between the gold cations and the silicon wafer. This layer was also highly porous and composed of dense tiny voids. The growth of a continuous Au film on such a surface was even more confined by the voids compared with the meso-PS, leading to the formation of a dense Au NP layer in an early deposition stage (Figure 6a). As the redox reaction progressed with time, the thin microporous layer dissolved, leaving the bare macro-PS surface opened. At this growth stage, the number of Au seed nucleation centers increased, and the surface diffusion of gold atoms became possible. As the result, the Au NPs coalesced, and the quasicontinuous film was produced. For the rest of the deposition time, the gold layer grew similarly to the Au film on the meso-PS. Figure 8 shows the XRD pattern for the Au-coated macro-PS after 50 min and 80 min of Au deposition. The patterns reported peaks associated with (111), (200), (220), and (311) Au crystal orientations, confirming the polycrystalline nature of the samples. Interestingly, in contrast to the XRD patterns for the Au-coated meso-PS (Figure 4a,b), no characteristic Si peak was observed in the XRD pattern for the gold-coated macro-PS, as shown in Figure 8a. This confirmed the introduced mechanism of Au film growth, where the coverage became uniform after the complete dissolution of the micro-PS layer. Apart from that, no noticeable difference between the Au-coated meso-and macro-PS samples was observed from the XRD; compare Figure 4a,b and Figure 8b.
We believe that the gold-coated macro-PS platform developed in this work can serve as an efficient nanoscale platform for SERS due to the strong plasmonic enhancement of the Raman signal. This can be better explained by closer inspection of the acquired SEM images in Figures 5 and 6. These images suggest that the obtained samples repeated the familiar nanoscale configuration, named metal nanovoids (NVs), which is one of the major candidates for mass utilization in SERS [61,62]. These NVs usually represent 500-2000 nm holes coated with various coinage metals and ensure the Raman signal enhancement by at least two basic optical effects: a localized plasmon mode with a strong rim component and multiple reflections of excitation light rays through the void. Moreover, the periodic morphology of the Me NVs guarantees a good spot-to-spot reproducibility of the SERS signal. The standard approach for Me NVs fabrication involves the deposition of a seed metallic layer, the coating with polymer spheres, the electrodeposition of a metal, and finally the dissolution of the polymer spheres. This procedure, however, is rather complex and limited by the small surface area covered with Me NVs. In contrast, our approach involved only two simple steps, electrochemical etching and chemical deposition, and can be extended up to wafer scale. It should be emphasized that the porosity and the thickness of the macro-PS in this work were optimized to meet the dimensions of the standard Me NV configurations, while the optimum Raman signal enhancement was determined by the variation in the Au deposition time in the 30-80 min range. riodic morphology of the Me NVs guarantees a good spot-to-spot reproducibility of the SERS signal. The standard approach for Me NVs fabrication involves the deposition of a seed metallic layer, the coating with polymer spheres, the electrodeposition of a metal, and finally the dissolution of the polymer spheres. This procedure, however, is rather complex and limited by the small surface area covered with Me NVs. In contrast, our approach involved only two simple steps, electrochemical etching and chemical deposition, and can be extended up to wafer scale. It should be emphasized that the porosity and the thickness of the macro-PS in this work were optimized to meet the dimensions of the standard Me NV configurations, while the optimum Raman signal enhancement was determined by the variation in the Au deposition time in the 30-80 min range.
(a) (b)  Figure 9 depicts the reflectance spectra for the Au-coated meso-PS ( Figure 9a) and macro-PS (Figure 9b) samples. These spectra clearly depict the reflectance minima around 480-495 nm for all samples, which could be associated with the (localized) surface plasmon resonance ((L)SPR). Considering that the highest Raman signal enhancement was achieved upon this (L)SPR condition, we therefore utilized the closest to this resonance and available for us laser wavelength of 532 nm. It should be emphasized that the (L)SPR observation for gold in the blue region of visible light is rather uncommon. It is known that isolated spherical Au nanoparticles demonstrate the LSPR band in the 520-580 nm wavelength region [63]. In some cases, however, it can shift towards the blue part of the spectrum, which is usually explained by the plasmons coupling between different nanoparticles upon shortening their spacing distance [64]. Therefore, our nanostructures were most likely to be affected by such a coupling effect, since they represented collections of densely packed Au nanoparticles comprising quasicontinuous films.  Figure 9 depicts the reflectance spectra for the Au-coated meso-PS ( Figure 9a) and macro-PS (Figure 9b) samples. These spectra clearly depict the reflectance minima around 480-495 nm for all samples, which could be associated with the (localized) surface plasmon resonance ((L)SPR). Considering that the highest Raman signal enhancement was achieved upon this (L)SPR condition, we therefore utilized the closest to this resonance and available for us laser wavelength of 532 nm. It should be emphasized that the (L)SPR observation for gold in the blue region of visible light is rather uncommon. It is known that isolated spherical Au nanoparticles demonstrate the LSPR band in the 520-580 nm wavelength region [63]. In some cases, however, it can shift towards the blue part of the spectrum, which is usually explained by the plasmons coupling between different nanoparticles upon shortening their spacing distance [64]. Therefore, our nanostructures were most likely to be affected by such a coupling effect, since they represented collections of densely packed Au nanoparticles comprising quasicontinuous films. Figures 10 and 11 show the SERS spectra for 10 −6 M R6G adsorbed on either the Au-coated meso-or macro-PS samples, respectively. To simplify the results representation and subsequent discussion, all the samples produced in this work and the corresponding fabrication regimes are summarized in Table 1. As can be seen from the figures, all SERS spectra exhibited the same characteristic Raman bands for R6G, namely 613 cm -1 (C-C-C ring in-plane bending); 772 cm -1 (C-H out of plane bending); 1185 cm -1 (C-H in-plane bending); 1311 cm -1 (C-O-C stretch); 1363 cm -1 , 1510 cm -1 , 1575 cm -1 , and 1650 cm -1 (aromatic C-C stretch). These results are in good agreement with the data reported elsewhere [65]. Figure 10 indicates that the intensity of the R6G SERS spectrum obtained for the best Au-coated macro-PS (sample 6) was two times higher than for the best Au-coated meso-PS (sample 2 , Figure 9). This confirms the stronger Raman signal enhancement provided by a configuration mimicking metal NVs compared with the usual metal NPs. The variation in the SERS spectra intensity between different nanostructures prepared on the same type of PS may be explained in terms of the stage of Au film growth. For example, the weak intensity of the R6G SERS spectra corresponded to the samples with a (quasi)continuous Au film (samples 2 and 5 in Figures 9 and 10, respectively; their morphology is highlighted in Figures 2a and 6b, respectively), while stronger SERS signals originated from the Au NPs (samples 2 and 6 in Figure 9, respectively; their morphology is highlighted in Figures 2c and 6c, respectively). These results can be correlated with the number of so-called "hot spots", tiny gaps in a metal nanostructure where the localization of a giant electromagnetic field occurred. The number of these spots was greater for the nanostructures exhibiting nanoscale roughness compared with the smooth continuous surfaces, leading to a higher probability for a molecule to be located within it and contribute to the enhancement of Raman scattering. With only this in mind, the reasons for the weak SERS response for sample 4 ( Figure 10) are rather unclear. One may expect comparable intensities of R6G SERS spectra for both samples 4 and 6 because of the similar nanostructure morphology (compare Figure 6a,c). However, the correct results interpretation also requires a particle size comparison. The calculated average particle diameter for sample 6 was 51 nm ± 30 nm, which was close to the optimum diameter of 60 nm for an isolated Au NP, demonstrating the maximum SERS enhancement [66,67]. On the other hand, the individual Au NPs for sample 4 could not be resolved from the SEM images, suggesting their sizes were far smaller than the known optimum. A similar analysis can be conducted for the other Au-coated samples prepared on the same type of PS matrixes. This means that the careful choice of the initial PS morphological parameters, as well as the Au deposition regimes, can help to obtain the optimum Au nanostructure with the corresponding NP size and spacing for maximum SERS enhancement.  Figures 10 and 11 show the SERS spectra for 10 −6 M R6G adsorbed on either the Aucoated meso-or macro-PS samples, respectively. To simplify the results representation and subsequent discussion, all the samples produced in this work and the corresponding fabrication regimes are summarized in Table 1. As can be seen from the figures, all SERS spectra exhibited the same characteristic Raman bands for R6G, namely 613 cm -1 (C-C-C ring in-plane bending); 772 cm -1 (C-H out of plane bending); 1185 cm -1 (C-H in-plane bending); 1311 cm -1 (C-O-C stretch); 1363 cm -1 , 1510 cm -1 , 1575 cm -1 , and 1650 cm -1 (aromatic C-C stretch). These results are in good agreement with the data reported elsewhere [65]. Figure 10 indicates that the intensity of the R6G SERS spectrum obtained for the best Au-coated macro-PS (sample 6) was two times higher than for the best Au-coated meso-PS (sample 2′, Figure 9). This confirms the stronger Raman signal enhancement provided by a configuration mimicking metal NVs compared with the usual metal NPs. The variation in the SERS spectra intensity between different nanostructures prepared on the same type of PS may be explained in terms of the stage of Au film growth. For example, the weak intensity of the R6G SERS spectra corresponded to the samples with a (quasi)continuous Au film (samples 2 and 5 in Figures 9 and 10, respectively; their morphology is highlighted in Figures 2a and 6b, respectively), while stronger SERS signals originated from the Au NPs (samples 2′ and 6 in Figure 9, respectively; their morphology is highlighted in Figures 2c and 6c, respectively). These results can be correlated with the number of so-called "hot spots", tiny gaps in a metal nanostructure where the localization of a  the Au deposition regimes, can help to obtain the optimum Au nanostructure with the corresponding NP size and spacing for maximum SERS enhancement.    the Au deposition regimes, can help to obtain the optimum Au nanostructure with the corresponding NP size and spacing for maximum SERS enhancement.

Conclusions
In summary, we demonstrated that an effective SERS platform may be fabricated by Au electroless immersion deposition on meso-and macro-PS. SEM, XRD, and SERS analyses were carried out for sample characterization. In this work, a macro-PS template was fabricated to reproduce the Me NVs' structure [61,62], while the meso-PS matrixes were produced with three different porosities. SERS measurements were performed for the R6G molecules adsorbed on different Au-coated meso-and macro-PS samples at a 10 −6 M concentration. It should be emphasized that the maximum SERS enhancement was achieved for the Me NVs' structures fabricated by a simple two-step procedure, while the current Me NVs' fabrication procedures involve at least four complex nanotechnology techniques, and the final structure is limited by a low surface area. We also demonstrated the correlation of the R6G SERS spectra intensity with the morphology and structural properties of Au coverage on the PS surface. Notably, the SERS response from the Aucoated PS samples may be tuned by the accurate choice of the initial PS morphological parameters, as well as the Au deposition regimes, which is explained by the formation of Au NPs with different sizes and spacings. Overall, both types of developed substrates demonstrated a sufficiently strong SERS response and show great promise for simple and versatile large-area plasmonic nanostructures fabrication.