Plasma‐Etched Nanograss Surface without Lithographic Patterning to Immobilize Water Droplet for Highly Sensitive Raman Sensing

The development of reliable, cost‐effective molecular detection at the attomolar level on analyte‐immobilizing surfaces fabricated without lithographic patterning remains a major challenge in chemical sensing technology. This issue is addressed using custom‐designed adhesive superhydrophobic silicon nanograss surfaces produced via plasma etching. When applied to ultrasensitive surface‐enhanced Raman scattering, the nanograss surface enables effective immobilization of water droplets containing Ag nanoparticles and R6G target molecules. Upon water evaporation, the R6G analytes are confined at the edge of the self‐organized coffee‐ring‐like stains with the plasmonic hot spots of the Ag nanoparticles, thus providing a reliable Raman scattering platform for detecting trace analytes. Even at an ultralow concentration of 10−16 m, the corresponding relative standard deviation is 17.57%. A novel plasma‐enabled approach for precise interface nanostructuring, potentially leading to unprecedented capabilities in molecular‐level sensing technologies, is presented.


Introduction
Raman scattering has been widely used to analyze molecular structures via inelastic scattering between incident lasers and molecules.However, because of the extremely small scattering cross-section of most molecules, the Raman signal must be amplified, specifically when detecting low concentrations of molecules. [1]sing metallic nanostructures, Raman scattering can be amplified through two mechanisms: an enhanced local electric field (near field) generated by the collective oscillation of excited electrons on the metal (electromagnetic enhancement, EM) and electron transfer between the metal and the target molecule (chemical enhancement, CHEM). [2,3]These surface-related strategies significantly amplify detection signals, enabling the detection of single molecules, [1,4,5] making surface-enhanced Raman scattering (SERS) an exceptionally sensitive analytical technique.However, trapping target molecules close to the plasmonic hot spots generated by nanostructures prolongs the accumulation time required to detect molecules in dilute solutions. [6]To address this limitation, several methods for enhancing the effectiveness of the interactions between molecules and hot spots have been developed by designing nanostructures, such as nanospheres, [7] nanogaps, [8] nanopores, [9] nanopillars, [10−12] and nanotubes. [13]Furthermore, several advanced techniques including tip-enhanced Raman scattering, [14] shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS), [15] and liquid interface-assisted SERS (LI-SERS) have been employed in SERS. [16]Recently, the focus on enhancing the sensitivity, reliability, and sensing efficiency of SERS has been increasing.The use of superhydrophobic surfaces that exhibit water-repellent properties is a promising approach.−21] However, typical state-of-the-art superhydrophobic surfaces are slippery with low sliding angles for droplets.Therefore, they need to be positioned using lithographic patterning processes [18,19,21] (i.e., optical lithography, [22] e-beam lithography, [6] laser writing, [23−25] laser printing, [26] and particle assembly) [27] to locate the droplets, which renders the fabrication process complex, expensive, and time-consuming.Although most superhydrophobic surfaces are slippery, sticky (adhesive) superhydrophobic surfaces have also been reported.Capillary force, [28,29] van der Waals' force, [30] and partial hydrophilicity [31,32] have been suggested to explain the unique sticky properties.Xu et al. revealed that the stickiness of droplets is determined by the density of the three-phase contact line between the solid, liquid, and gas. [33]We previously demonstrated that a superhydrophobic surface can exhibit either slippery or sticky behavior by adjusting the length of the three-phase contact line. [34]Herein, we employed silicon nanograss, [35] which demonstrates exceptional uniformity, nanoscale feature size, and high packing density of nanostructures for SERS measurements.These characteristics enabled a high-density three-phase contact line, thereby enhancing the adhesion of water droplets.
In addition to utilizing superhydrophobic surfaces that offer a reduced contact area, another effective approach for further minimizing the surface area is to use the coffee-ring effect.When a water droplet containing analytes evaporates on a hydrophobic surface, the analytes can be drawn toward the perimeter of the contact spot by the outward capillary flow of the interior water.The effectiveness of this approach has been demonstrated in enhancing the detection sensitivity of analytes in various applications, [36−38] including drop coating deposition Raman spectroscopy (DCDRS). [39]However, creating a coffee ring on a superhydrophobic surface remains challenging because positioning the droplets is difficult, and accumulation tends to be confined to a single spot rather than forming a coffee-ring stain. [40]he nanograss exhibits exceptionally sticky superhydrophobic characteristics, making it ideal for SERS applications.We previously investigated the SERS properties of nanowires prepared via chemical etching. [41]To the best of our knowledge, no chemically derived nanowire film-based research has reported similar SERS properties based on sticky superhydrophobicity to concentrate analytes on the hot spots distributed along the small coffee-ring edge.Accordingly, we utilized this sticky superhydrophobic silicon nanograss to immobilize droplets without a lithographic patterning process.The Ag nanoparticles and analytes can be concentrated on the edge of the coffee-ring stain by evaporating the droplet at 110 °C, which increases the possibility of finding trace molecules, thereby enhancing the reproducibility and reliability of SERS measurements.The trace (10 −16 m) R6G molecules were precisely detected with an acceptable relative standard deviation (RSD) of 17.57%, and the maximum analytical enhancement factor (AEF) was higher than 10 11 , indicating strong potential for detecting analytes with single-molecule concentrations.

Nanograss with High-density Three-phase Contact Line
To create Si nanograss on a Si wafer, we used an inductively coupled plasma (ICP) system with a mixture of hydrogen and argon gases to etch the Si wafer.Before loading the sample into the reaction chamber, the chamber was cleaned using CF 4 and O 2 plasma.As no additional process was required for mask fabrication, plasma-polymerized fluorocarbon was suggested to serve as a self-mask for nanograss formation. [35]Figure 1a shows a crosssectional scanning electron microscopy (SEM) image of the silicon nanograss.Each rod-like nanotip had an average diameter of ≈30 nm and a height of ≈230 nm, displaying excellent height uniformity.The top-view SEM image and the corresponding contour image are shown in Figure 1b,c, respectively.The density of the nanograss was determined to be higher than 934 nanotips per square micrometer based on contour images.The total perimeter of the nanotips within an area of 800 nm × 800 nm was 51.44 μm, which indicates that the topography of the nanograss significantly increases the density of the three-phase contact line, increasing the resistance for droplet movement.The surface profile and roughness were measured through atomic force microscopy (AFM), and the 2D and 3D images are shown in Figure 1d,e, respectively.The average roughness (R a ) is only 19 nm, demonstrating exceptional height uniformity.This indicates that all the top surfaces of the nanograss were in contact with the water droplets to retard their movement.Figure S1 (Supporting Information) shows that the total reflectance of the nanograss structure with a height of 230 nm was reduced from 37% to 3% on average over a wavelength range of 400-900 nm, and the reflectance was reduced tenfold by the nanograss compared to the flat silicon wafer at the excitation wavelength (532 nm).This low-reflective surface enables a high absorption of incident light.After modification with F 13 -TCS silane, the nanograss surface was transformed from superhydrophilic to superhydrophobic, which allowed a small footprint to be left from a 2.5 μL water droplet, as shown in Figure 1f.Furthermore, because the surface is sticky, the droplets can be manipulated anywhere on the surface, as shown in the inset of Figure 1f, where the droplets are arranged in an array without rolling off the surface.The enlarged view shows Ag nanoparticles distributed on the top surface of the silicon nanograss (Figure 1g).The inset in Figure 1g shows that the particle diameter and the gap between the particles were 100 and 10 nm, respectively.Ag nanoparticles with a diameter of 100 nm have a maximum absorption at 488 nm, as shown in Figure S2 (Supporting Information), which is close to the wavelength of incident light and therefore facilitates the excitation of surface plasmon resonance (SPR).In addition, a gap between the nanoparticles was created by the sodium citrate monolayer, which provides essential hot spots for SERS.For nanoparticles with a 100 nm diameter (D) and a 10 nm gap (d), the local electrical field strength can be increased by a factor of (D/d + 1) 4 = 146541. [3]Therefore, localized SPR (LSPR) enables ultrahigh-sensitivity Raman detection if molecules can be deposited near the hot spots generated by Ag nanoparticles.

Sticky Superhydrophobic Surface
After functionalization to decrease its surface energy, the nanograss exhibited a superhydrophobic surface.However, owing to the high-density triple-phase contact line, the nanograss exhibited remarkable droplet anchoring ability.Figure 2a shows the dependence of the water contact angle on the tilt angle for various droplet sizes.As the tilt angle of the sample stage increases through rotation, the gravitational force causes the droplet to move downwards, increasing the advancing angle and decreasing the receding angle.The smallest water droplet (1.5 μL) was attached to the surface, even when the tilt angle approached 90°( Figure 2b).Figure 2c shows that the sliding angle () increased as the droplet size decreased.The coefficient of static friction (μ s ) evaluated by tan(), shown in Figure 2c, also indicates that smaller droplets had larger μ s (the value for a 1.5 μL droplet is not shown here because the droplet adhered to the vertical surface), which can be explained by the greater Laplace force imposed by the larger curvature of the smaller droplets. [42]When a water droplet is impaled by nanograss structures, the increased surface area may cause the water droplets to become pinned.Figure 2d shows the variation in the contact angle and contact diameter of a 2.5 μL droplet during evaporation.The contact diameter changed only slightly at the beginning of the evaporation, indicating that the water droplet was pinned to the nanograss.The gradually decreasing contact angle suggests that the water droplets were firmly anchored to the surface, which differs from a typical superhydrophobic surface that maintains a high contact angle throughout the water evaporation process. [23]he ability to anchor a droplet to a superhydrophobic surface enables precise positioning without complex lithographic patterning processes.As shown in Figure 3a, a two-step positioning method was used to create a plasmonic active region for analytes, wherein a 2.5 μL droplet containing 100 nm Ag nanoparticles (0.02 mg mL −1 ) was first placed on the surface.The droplet was then baked to evaporate the water, following which a second 2.5 μL droplet containing R6G molecules was placed in the same position as the first droplet.A coffee-ring stain was formed by heating the droplet, as shown in Figure 3b and Movie S1 (Supporting Information).Owing to the limited presence of Ag NPs and the maintenance of nonwetted regions in the superhydrophobic state, subsequent droplets containing R6G molecules could be placed on the same spot without any spreading.This approach reduced the consumption of Ag nanoparticles and increased the probability of finding hot spots on the ring edge.

Reliable SERS Measurements with Ultrahigh Sensitivity
Figure 4a shows that the perimeter of the ring stain from the water droplet after evaporation is influenced by the heating temperature.As the temperature increased, the width of the perimeter decreased, indicating that the heat accelerated the outward flow of the interior liquid, causing the accumulation of nanoparticles and analytes on the ring edge where the droplet was pinned by the three-phase contact line.Owing to the small size of the laser spot (≈800 nm in diameter, which was determined by calculating the diameter of the Airy disk using the formula 1.22 × /NA, where  represents the laser wavelength (532 nm) and NA is the numerical aperture of the objective lens, 0.8 for the 100× lens), detecting trace molecules is challenging.For example, there are only 150 molecules in a 2.5 μL droplet with a concentration of 10 −16 m.However, by applying heat to the immobilized water droplet, we limited the plasmonic active region from the entire contact area under the droplet to only its perimeter, where the Ag   nanoparticles had aggregated.Thus, trace molecules are more likely to be detected.
The Raman spectra measured from the ring edge are shown in Figure 4b,c.A sample of 10 −4 m R6G on a Si wafer without Ag nanoparticles was also recorded as a reference, and its signal was significantly lower than that of the specimen containing Ag nanoparticles, indicating that the plasmonic active site was significant for the enhancement of Raman scattering.The spatial Raman mappings of R6G molecules with 10 −12 and 10 −16 m concentrations extracted from the band at ≈605 cm −1 are presented in Figure 4d.According to the intensity distribution, the aggregation of Ag nanoparticles could be responsible for the enhancement of the Raman signals.We also conducted a simulation to demonstrate near-field enhancement using our in-house developed rigorous coupled-wave analysis. [43]Figure 4e shows the near-field distribution around Ag nanoparticles with a diameter of 100 nm arranged in a hexagonal pattern with a gap size of 10 nm.Air was assumed as the environmental medium.Light ( = 532 nm) was introduced at normal incidence with polarization along the x-and y directions.A significant LSPR enhancement was observed in the gaps between the nanoparticles, which contributed to the advancement of the Raman signal enhancement.
When the concentration of R6G was relatively high (>10 −8 m), the characteristic signal of R6G was easily identified by SERS, as shown in Figure 4b.−46] Interestingly, when the concentration decreased significantly, such as in the range of 10 −8 −10 −16 m, distinctive signals were still detected (Figure 4c).The strongest peaks ≈605 cm −1 for these samples are presented in Figure 4f, each of them was obtained by measuring the sample at 10 different positions, showing that the repeatability was high.To verify the detection of trace molecules, we conducted Raman mapping (605-611 cm −1 ) at the ring edge, as shown in Figure 4g.The presence of a ring pattern in the 10 −16 m sample reveals that the molecules were accumulated at the ring edge.Because the stability of molecules near the available hot spots determines the fluctuations in SERS intensity, [47] we propose that the formation of a small coffee ring following the two-step evaporation process allowed lower fluctuations by keeping molecules and hot spots in the same location.Moreover, we used this platform to assess the SERS spectrum of 10 −16 m malachite green (MG) molecules, as shown in Figure 4h.The identifiable peak observed at 1616 cm −1 , which corresponds to the ring C─C stretching within the molecule, [48] further verifies the remarkable sensitivity of our approach.
The explanation for this ultrahigh sensitivity is the high probability of molecules being close to the hot spots.Some regions of the non-wetting nanograss within the footprint were exposed (Figure 1g), indicating that the actual area of the Ag nanoparticle distribution was further reduced inside the footprint.In addition, the exposed hydrophobic nanograss may repel molecules into the region with the dispersion of more hydrophilic Ag nanoparticles, which also helps narrow the range of molecules near the hot spots.As the Ag nanoparticles were significantly larger than the silicon nanograss and the voids between them, the particles accumulated on the top surface of the nanograss.As shown in Figure 1g, the voids between the nanograss were sealed under the Ag nanoparticles such that the droplets containing R6G molecules could be placed on top of the Ag nanoparticles without penetrating the substrate, resulting in a strong LSPR for SERS.
We can evaluate the quantity of SERS using AEF, which is defined as [49] AEF = I SERS ∕C SERS I RS ∕C RS (1)   where I SERS and C SERS represent the intensity and concentration of R6G accompanied by Ag nanoparticles, respectively; I RS and C RS represent the Raman scattering intensity and concentration of R6G without Ag nanoparticles on a flat silicon wafer, respectively.Calculating the AEF from the 10 −16 m and reference (10 −4 m) samples, the AEF is >10 11 , which is significantly higher than most reported values, [50] indicating the ultrahigh sensitivity of the newly developed sensing platform.

Reliability of SERS Measurements
To further assess the reliability of the SERS measurement, we present the variation in SERS intensity at 605 cm −1 , obtained from 10 different spots on each sample, as shown in Figure 5a.The SERS results with Ag nanoparticles obtained from the blank Si wafer and the hydrophobic Si wafer (coated with silane molecules) are also shown in the inset of Figure 5a for comparison.Notably, the Raman signals were significantly weaker and the RSDs were high (0.540 and 0.434 for the blank and hydrophobic wafers, respectively) in the absence of superhydrophobic nanograss.Figure 5b depicts the relationship between the intensity and concentration, and the corresponding RSD across various concentrations.The plot reveals two regions with a linear correlation.The linear correlation coefficients (R 2 ) were 0.983 and 0.965 for zone (I) and zone (II), respectively.This indicates a predictable connection between the concentration and Raman intensity at different concentration ranges.−54] In contrast, our study showed good correlation, allowing us to obtain reliable signals even from low-concentration samples.Additionally, we compared the RSD values with those reported in other studies at varying concentrations, [20,55−61] and our results were lower than those of most reported records.Generally, the analytes with higher concentrations tended to exhibit lower RSDs. [56,60,62]Although ultrahigh-sensitive SERS methods have been reported previously, most of them provide RSD values only for samples with higher concentrations.In contrast, our method provided an acceptable RSD value of 17.57% even at concentrations as low as 10 −16 m, indicating that the method is suitable for detecting analytes at ultralow concentrations.Three regions are shown in Figure 5b; therefore, we propose a mechanism based on the relative number of R6G molecules and Ag nanoparticles to explain the dependence of SERS intensity on molecule concentrations, as shown in Figure 5c.The Raman scattering intensity increases linearly with the concentration of R6G in region II, indicating that R6G is uniformly distributed on the surface with active hot spots.However, when the concentration of R6G is extremely high (greater than 10 −4 m), the number of available adsorption sites on the nanoparticle surface becomes limited.This leads to the accumulation of too many R6G molecules, leaving some without SERS activity because of their distance from the hot spots.This results in a reduction in the concentration effect on the intensity, which is observed in zone III.However, in zone I, the number of silver nanoparticles and their hot spots is significantly higher than the number of R6G molecules, which makes it easier for R6G molecules to be positioned near the hot spots, resulting in a reliable measurement of the SERS effect, even when the analyte concentration decreases significantly.When the concentration is low, the relatively high number of hot spots near the molecule accounted for the slowing down of intensity decrease as the concentration decreased.Consequently, the linear relationship exhibited a smaller slope.These proposed mechanisms demonstrate that the superhydrophobic anchoring effect allows the reliable detection of molecular concentrations at very low ranges, even down to the single-molecule level (Figure S4, Supporting Information).

Conclusion
This study addressed a challenge in molecular-level sensing by introducing a novel patterning-free technique for anchoring droplets and positioning both Ag nanoparticles and analytes at the ring edge of a coffee-ring-like stain to enable detection at the single-molecule level.Sticky superhydrophobic silicon nanograss was produced via plasma etching, which allowed water droplets containing Ag nanoparticles and R6G analytes to be precisely positioned on the same small surface area without additional lithographic patterning processes.A coffee-ring stain was created by heating the droplet and the plasmonic region was concentrated on the ring edge, providing high sensitivity and reproducibility for the detection of trace molecules.In contrast to the lower sensitivity and repeatability observed in the samples without nanograss, the R 2 coefficients for the relationship between SERS intensity and concentration were 0.983 and 0.965, respectively, within the concentration ranges of 10 −8 −10 −16 m and 10 −6 −10 −8 m.Additionally, the RSD for the 10 −16 m concentration was 17.57%, and the AEF was greater than 10 11 .This easy-to-locate and highly sensitive SERS platform could facilitate the precise detection of environmental contaminants and disease biomarkers through trace analysis.
Preparation of Silicon Nanograss Substrates: Silicon nanograss samples were fabricated on a silicon wafer using an ICP system operated at a base pressure of 10 −5 Torr. [35]Prior to etching, the wafer was cleaned using a standard cleaning process, and the chamber was cleaned with CF 4 and O 2 plasma for 15 min.The flow rate of both CF 4 and O 2 was 200 sccm, and the working pressure was set to 50 mTorr.The parameters for the nanograss fabrication were radio frequency (13.56 MHz), power of 500 W, bias power of 300 W, working pressure of 50 mTorr, time of 30 min, and a temperature of 400 °C.
Functionalization of Surface with a Silane Monolayer: The as-prepared silicon nanograss exhibited a superhydrophilic surface.To transform the surface from a superhydrophilic to superhydrophobic state, F 13 -TCS was employed to reduce the surface energy by forming C-F groups on the surface. [63,64]The procedure involved placing the sample in a dish and covering it with a smaller dish.The F 13 was then injected into the sample through a small hole between the two dishes.Subsequently, the samples were heated to 210 °C and incubated for 2 h.The excess F 13 -TCS was removed by washing with n-hexane. [65]The presence of a terminal -CF 3 group on the surface contributed to the desired hydrophobicity.
Preparation of R6G and Ag Nanoparticle Solutions: For the R6G solution, a 50 mL solution with 10 −2 m concentration was initially prepared by dissolving 0.2395 g R6G in water.Serial dilutions were performed to obtain highly diluted solutions (Figure S3 and Movie S2, Supporting Information).First, a 2.5 μL droplet containing Ag NPs was deposited on the silicon nanograss.After drying by baking at 90−110 °C for 10 min, another 2.5 μL droplet containing R6G molecules was deposited on the same site.
Surface Characterization: The morphology of the silicon nanograss and the distribution of the deposited Ag nanoparticles were examined through field-emission SEM (6500F, JEOL).The wetted area was evaluated using a digital microscope (VHX-6000, Keyence).The surface roughness of the nanograss was measured through AFM (Dimension Icon, Bruker).The number and density of silicon nanotips in the nanograss were calculated using the ImageJ software package.The total reflection from the silicon nanograss was measured using a spectrometer equipped with an integrating sphere.The absorption of the Ag nanoparticles was measured using a UV-vis spectrometer (Hitachi, U-3900).
Contact Angle Measurements: The contact and sliding angles of the water droplets were evaluated using a contact-angle goniometer (Phoenix MT, SEO).Water droplets with volumes of 1.5−6.5 μL were utilized for the measurements, and the contact angles were determined using the tangent line method.
Raman Spectra Measurements: SERS measurements were performed using a confocal Raman microscope system (UniDRON, CL).The following measurement conditions were applied: wavelength: 532 nm, power: 0.01 mW, objective lens magnification: 100×, exposure time: 2 s, and 10 repetitions of exposure.The Raman measurement performed on other nanograss substrates etched for 30 min reveals that the nanograss formed on different substrates has minimal impact on the measurements (Figure S5, Supporting Information).2D Raman mapping was performed using a confocal Raman microscope (Nanofinder 30, Tokyo Instruments) with He-Ne laser excitation (633 nm).The objective lens was 100×, the scanning area was 15 μm × 15 μm, and large-area Raman mapping was performed using another confocal Raman system (JadeMat, Southport Corporation) with 532 nm wavelength.The laser power was 20 mW, the objective lens was 10×, and the scanning area was 500 μm × 500 μm.
Statistical Analysis: The data presented in Figure 5b were obtained from 10 measurements for each concentration.The standard deviation was calculated using Excel, and the plot was generated using OriginLab software.

Figure 1 .
Figure 1.Surface morphology of the silicon nanograss with the corresponding contour after positioning a 2.5 μL drop of Ag solution on the surface.a,b) Cross-sectional and top-view SEM image of the silicon nanograss.Scale bar: 100 nm.c) Surface contour of the silicon nanograss.d,e) 2D and 3D AFM images of the silicon nanograss.Scale bar: 200 nm.f) SEM image showing a footprint from a 2.5 μL water droplet containing Ag nanoparticles.Scale bar: 0.5 mm.Inset shows that the droplets can be arbitrarily deposited on the surface without additional patterning.g) Enlarged view of (f), showing the distribution of 100 nm Ag nanoparticles on the surface.Scale bar: 1 μm.Inset shows the particle size and the nanogap associated with the Ag nanoparticles.Scale bar: 100 nm.

Figure 2 .
Figure 2. Effects of the tilt angle on the contact angle.a) Variation of advancing, receding, and average contact angle with the tilt angles.b) Images of 1.5 μL droplets stuck on an inclined surface.c) Sliding angle and coefficient of static friction for various droplet sizes.d) Change in the water contact angle and contact diameter during evaporation, showing that the droplet was pinned by the surface because the contact angle decreased gradually.Inset shows the overlap images of 2.5 μL water droplets during evaporation.

Figure 3 .
Figure 3. Two-step positioning of R6G on Ag nanoparticles.a) Schematic of the two-step positioning process.Benefitting from the presence of sticky nanograss, R6G-containing droplets can be precisely positioned on top of the Ag nanoparticles.b) Pictures to show a 2.5 μL droplet before and after evaporation at 110 °C.Coffee-ring stain was left after evaporation, which allowed the laser spot to find the plasmonic active region.Scale bar: 0.5 mm.

Figure 4 .
Figure 4. Effects of the analyte concentration on Raman scattering.a) Optical images of water footprint heated at varying temperatures.Scale bar: 100 μm.b) Raman scattering spectra of R6G molecules recorded from samples with higher concentrations (≥10 −7 m).The stars indicate the spectral fingerprints of the R6G molecule.c) Raman scattering spectra of the R6G molecules recorded from samples with lower concentrations (≤10 −8 m).d) 2D Raman spectral mapping of 10 −12 and 10 −16 m R6G molecules on the silicon nanograss.e) Near-field distribution around Ag nanoparticles with a diameter of 100 nm arranged in a hexagonal pattern with a gap size of 10 nm.f) Enlarged view of (c) at ≈605 cm −1 .g) Raman mapping (≈605−611 cm −1 ) for 10 −16 m sample.Scale bar: 100 μm.h) Raman spectra of MG molecules with concentrations of 10 −10 and 10 −16 m, respectively.

Figure 5 .
Figure 5. Reliability of ultrasensitive SERS measurement.a) SERS intensity measurements acquired at 10 different spots on each sample.Inset shows SERS measured on blank and hydrophobic Si wafers.b) SERS intensity and RSD values as a function of R6G concentration.The standard deviations were calculated from the data presented in (a).The stars indicate the RSD values reported in other studies.c) Schematic showing the effects of the relative numbers of R6G and Ag on the SERS sensitivity.