Ag/Au Nanoparticle-Loaded Paper-Based Versatile Surface-Enhanced Raman Spectroscopy Substrates for Multiple Explosives Detection.

We present a systematic study on the fabrication, characterization of versatile, and low-cost filter paper-based surface-enhanced Raman spectroscopy (SERS) substrates loaded with salt-induced aggregated Ag/Au nanoparticles (NPs). These were demonstrated as efficient SERS substrates for the detection of multiple explosive molecules such as picric acid (5 μM), 2,4-dinitrotoluene (1 μM), and 3-nitro-1,2,4-triazol-5-one (10 μM) along with a common dye molecule (methylene blue, 5 nM). The concentrations of the dye and explosive molecules in terms of mass represent 31.98 pg, 11.45 ng, 1.82 ng, and 13.06 ng, respectively. Silver (Ag) and gold (Au) colloidal NPs were prepared by femtosecond laser (∼50 fs, 800 nm, 1 kHz) ablation of Ag/Au-target immersed in distilled water. Subsequently, the aggregated nanoparticles were achieved by mixing the pure Ag and Au NPs with different concentrations of NaCl. These aggregated NPs were characterized by UV-visible absorption and high-resolution transmission electron microscopy techniques. The SERS substrates were prepared by soaking the filter paper in aggregated NPs. The morphologies of the paper substrates were investigated using field-emission scanning electron microscopy technique. We have achieved superior enhancements with high reproducibility and sensitivity for filter paper substrates loaded with Ag/Au NPs mixed for an optimum concentration of 50 mM NaCl.

Introduction

Surface-enhanced Raman spectroscopy (SERS) is a simple and nondestructive technique which promises numerous applications such as the trace detection of drugs, explosives, food contaminants, biological entities, and environmental hazardous gases.[1−8] In SERS technique, the intensity of inelastically scattered photon from the molecule can be enhanced using the long-range electromagnetic (EM) field and short-range chemical effects. EM enhancement is attributed to the localized surface plasmon resonance (LSPR) in the near-field metallic surface, and the chemical enhancement is because of the charge transfer (CT) mechanism between the substrate and the analyte. Generally, the enhancement because of the CT is 1–3 orders of magnitude smaller than that of EM.[9] The interaction of the incident EM field with metal nanoparticles (NPs) possessing negative real and small positive imaginary (absorption) dielectric constant induces a collective and coherent electron oscillations, called plasmons, in the vicinity of the NP. At the overlap of the near-field regions between adjacent NPs called “hot-spots”, the local field enhancement is superior.[10,11] Several techniques have been employed to fabricate SERS substrates (metal nanostructured patterns) such as electron beam lithography, focused ion-beam lithography, atomic layer deposition, and ultrafast pulsed laser ablation producing a variety of patterned nanostructures.[12−16] These experimental techniques are expensive and also complex procedures need to be followed in the preparation of SERS platforms. In practical SERS applications, the prepared traditional rigid substrates (Si wafer, glass, and metal) have numerous drawbacks because of the difficulty in collection efficiency and manipulation of the solid samples.

Ideally, SERS substrates should not only have multiple numbers of hotspots and capability of low-detection limit but also should be versatile in many aspects such as uniformity, reproducibility, scalability, and cost. The main challenge for the SERS community is to develop a low cost, easy to handle, flexible, and sensitive practical SERS substrates, especially for the detection of trace hazardous materials.[17−19] A common filter paper (FP) loaded with metal NPs serves as the versatile SERS substrate. The cellulose-rich paper has attracted growing interest in various applications, such as food safety, sensing devices, cancer screening, solar cell, and paper electronics such as to build a various electronic component (diode, transistor, antenna, superconductors, etc.).[20,21] The tremendous attributes of cellulose paper such as the thinness of cellulose fibers, porosity, sample storage capability, lightweight, disposable, handy, flexible (rolled or folded), and annual renewability serve as the versatile SERS substrate. Different methods are available for the fabrication of simple and low-cost paper-based SERS substrates such as inkjet printing, screen printing, and wax printing NPs on cellulose paper.[22−28] These techniques require the formulation of NPs ink as an appropriate liquid carrier for proper viscosity and surface tension. NPs made of noble metals, such as silver (Ag) and gold (Au) are extensively utilized materials in the preparation of SERS substrates because of their fantastic properties such as LSPR in the visible region relative to other metals. Polavarapu et al. recently demonstrated a “pen-on-paper” approach for creating SERS substrates and obtained average enhancement factors (EF) of 2 × 105 and 1.5 × 105 at excitation wavelengths of 532 and 785 nm, respectively, utilizing the Ag NP substrates for malachite green (MG-1 × 10–6 M) dye.[29]

Oliveira et al. reported high reproducibility with good uniformity and high SERS enhancement (EF ≈ 107) with office paper SERS substrates using Ag nanostars drop-casted in wells and patterned in the paper using printed wax.[26] Ngo et al. reported the performance of robust and recyclable Au NP-loaded FP SERS substrate prepared with the simple approach of dip-coating and detected the molecule 4 NTP (∼1 nM).[24] Herein, two typical routes have been pursued to improve the efficacy of paper-based SERS platforms by inducing the aggregation of NPs: (1) salt-induced aggregation and[30] (2) paper-induced aggregation.[31] These NPs tend to aggregate under specific chemical conditions (in the presence of ionic liquids) and are termed as salt-induced aggregation. The three-dimensional network of cellulose fibers on paper enables the transition of NPs in colloidal solution to solid state with aggregation of NPs, called paper-induced aggregation. Multiple metal NPs when brought together to form aggregates yield high field strength and many orders of magnitude greater than the fields at the surface of individual NP. This could be attributed to the coupling of their transition dipoles, that is, each particle’s enhanced field interferes coherently at the junction between the aggregated NPs.[32,33] Several reports have shown the significant increment in SERS signals from intentionally aggregated Ag and Au NPs. Xu et al. observed large EFs (∼1010) between two Ag NPs separated by 1 nm.[34] Kneipp et al. explained the effect of SERS signal on NaCl (10–2 M)-activated Ag colloidal aggregates in different stages and observed an increase in EF by 7–8 orders of magnitude through the formation of aggregates.[35] Theoretical estimates also demonstrated the enhancement of local field in the interparticle gaps of closely spaced NPs.[36] In the present work, paper SERS substrates were prepared by soaking the FP in aggregated NPs achieved with different concentrations of NaCl. The aggregation effects improved the number of hotspots (interstitial gaps formed between the NPs to entrap the analyte molecule which enhance the SERS signal).[11] The synthesized active plasmonic paper SERS substrate is an ideal platform for building portable miniature for the detection of adsorbed molecules with high specificity and sensitivity, especially for explosive detection in countering terrorist extortions, trace amounts of contaminants, pollution of soil, and disease diagnostics.[37,38] To fabricate noble metal NPs, numerous techniques were investigated and implemented including chemical methods, citrate reduction method, electrochemical methods, and laser ablation in liquids (LAL).[39−43] In the present case, plasmonic NPs were fabricated through LAL technique, which has the advantages of fast, easy, and green approach,[44] devoid of long reaction times and multistep chemical processes.[45] During the last few years, the laser ablation of bulk targets (metals, semiconductors, etc.) immersed in liquid has augmented in popularity because of the tremendous advantages over the chemical routes. High surface-purity NPs are achieved in a chosen solvent without any counterions and residuals of the reducing agents on the surface. Such NPs produced without any capping agent are very useful in SERS studies. Au and Ag NPs were fabricated by laser ablation of Ag/Au in liquid. Later, the aggregated NPs were achieved by mixing the pure Ag/Au NPs in different concentrations of NaCl. Plasmonic SERS substrates were prepared by soaking the FP in metal (Ag/Au) NP aggregates which are self-assembled in solution[46] by the addition of the NaCl molecule. These SERS substrates were enabled to detect the dye-like methylene blue (MB—5nM) and explosives such as picric acid (PA—5 μM), 2-4-dinitrotoluene (DNT—1 μM), and 3-nitro-1,2,4-triazol-5-one (NTO—10 μM) using a portable Raman spectrometer.

Characterization of the NPs

UV–Visible Absorption Spectra

To achieve the aggregated NPs, initially pure Ag/Au NPs were suspended in different concentrations of NaCl (from 1 mM to 1 M) solutions and their absorption studies were investigated by placing 3 mL of colloidal solution in 1 cm quartz cuvette using the UV–visible (PerkinElmer, LAMBDA 750) absorption spectrometer in the wavelength range of 300–800 nm. Figure a,b illustrates the absorption spectra of pure Ag and Au NPs mixed with different concentrations of NaCl.

Figure 1

UV–visible absorption spectra of (a) nonaggregated spherical Ag NPs with λmax = 404 nm; aggregated spherical Ag NPs with different NaCl concentrations and (b) nonaggregated spherical Au NPs with λmax = 520 nm; aggregated spherical Au NPs with different NaCl concentrations.

For pure Ag NPs, SPR position was located at ∼404 nm. It is obvious that the addition of NaCl solution resulted in decreased absorption, possibly because of the aggregation effect. A broadening along with a redshift in the plasmon band (∼512 nm) was also observed which is clearly depicted in the inset of Figure a. In the case of pure Au NPs, the SPR peak is located at ∼520 nm. By adding various concentrations of NaCl (up to 50 mM), a slight broadening and a small red shift in SPR peak were noticed which could be attributed to the formation of small aggregates of colloidal Au NPs. However, in the case of higher concentrations of NaCl (from 100 mM), more broadening and red-shift (SPR peak near 678 nm) was observed. The possible reason for this is the replacement of smaller aggregates into larger aggregates at higher concentrations. Because of the aggregation effect, a shift in the plasmon peak position, broadening of SPR peak, and color change of the colloidal solution was noticed in earlier works, which was attributed to the chemical interface damping.[47] The results observed in this study matched well with the theoretical and experimental investigations of Mehrdel et al.[47] The change of the SPR peak from shorter to longer wavelengths is evident from the change in the color of colloidal solution (Figure S1). The absorption of pure Ag/Au NPs on the FP (Figure S2) was also recorded, and a broadening in the SPR peak on FP was noticed compared to that of NPs in solution, which could be ascribed to the paper-induced aggregation of NPs (shown in Figure S3).

Transmission Electron Microscopy (TEM) Analysis of Colloidal NPs

Particle sizes and morphology of the NPs were investigated using transmission electron microscopy with an electron-accelerating voltage of 200 kV, whereas selected-area electron diffraction (SAED) pattern was utilized to identify the structural characterization and constituents of NPs. The shape of Ag/Au NPs was almost spherical (Figure a,c). The average size of Ag and Au NPs was ∼10.3 and ∼6.5 nm (insets of Figure a,c). The interplanar spacing (d111) obtained through high-resolution TEM (HRTEM) images was 0.24 nm for Ag and 0.232 nm for Au NPs (Figure b,d). SAED patterns revealed the polycrystalline nature of fabricated Ag/Au NPs (insets of Figure b,d). The aggregation of Ag/Au NPs after the addition of NaCl (50 mM) was also confirmed from the TEM image data presented in Figure a,b.

Figure 2

TEM and HRTEM images of (a,b) Ag NPs and (c,d) Au NPs in distilled water at lower and higher magnifications. Inset of (a,c) and (b,d) depicts their size distribution and SAED patterns, respectively.

Figure 3

TEM images of (a) Ag NPs and (b) Au NPs mixed with NaCl solution at a concentration of 50 mM. FESEM analysis of the SERS substrates.

Before field-emission scanning electron microscopy (FESEM) characterization, a thin conductive layer of gold was sputtered on the FP to facilitate the lower magnification images because of a nonconductive nature of FP substrate. FESEM (Carl Zeiss model Merlin Compact 6027) measurements were conducted with a beam voltage of 30 kV to investigate the morphology of the aggregated NPs on the FP surface. Interweaved cellulose fibers of porous bare FP with the high surface area are shown in Figure S4. It can provide a fine adsorption capability using capillarity action for the NPs and the analyte molecule.[22] For comparison, the FP loaded with Ag/Au NPs without NaCl at lower and higher magnification images are provided in Figure S5.

The FPs loaded with aggregated Ag NPs at different concentrations of NaCl (1 mM, 10 mM, 50 mM, 100 mM, 500 mM, and 1 M) are illustrated in Figure a–f. Similarly, FPs are loaded with the aggregated Au NPs, as shown in Figure a–f, for concentrations ranging from 1 mM to 1 M. In the case of FP loaded with NPs at very lower concentrations of NaCl (1 and 10 mM), fairly few aggregates are observed and are distributed on the cellulose fibers. At higher concentrations of NaCl (500 mM and 1 M), a large number of NP aggregates are observed because of an excessive amount of NaCl (Figures e & 5e). The excess aggregation of NPs may possibly limit the number of hotspots or help in forming the precipitate on the cellulose fibers which may also inhibit the formation of hotspots. Therefore, a higher concentration of NaCl solution will induce a significant impact on aggregation of Ag/Au NPs and forms clusters on the fibers. The higher magnification images revealing the distribution of aggregated Ag/Au NPs on the paper substrates could be because of the salt-induced aggregation in addition to the paper-induced aggregation. From Figures S6 and S7, the distribution of NPs and over-precipitate of aggregated NPs on the FP fibers are evident. However, SEM images of FP loaded with Ag/Au NPs at 50 mM NaCl concentrations clearly demonstrated the distribution of high-density NPs on the cellulose fibers. The energy-dispersive X-ray spectrum further confirmed that the presence of Ag (weight ∼80%) was higher compared to Na and Cl on the FP surface (presented in Figure S8). We also noticed the other elements such as carbon (C) and oxygen (O), arising from the inherent elements of cellulose fibers. The FP loaded with Ag/Au NPs at 50 mM NaCl concentration depicted higher density of the NPs embedded, leading to plenty of EM hotspots evenly distributed, and will act as efficient SERS substrate compared to other substrates.

Figure 4

FESEM images of FP loaded with Ag NPs with NaCl at different concentrations: (a) 1 mM, (b) 10 mM, (c) 50 mM, (d) 100 mM, (e) 500 mM, and (f) 1 M with the same magnification 10k×. The scale bars in (a,d,e,f) are 1 μm, whereas in (b) and (c), it is 2 μm.

Figure 5

FESEM images of FP loaded with Au NPs with NaCl at different concentration (a) 1 mM (b) 10 mM (c) 50 mM (d) 100 mM (e) 500 mM (f) 1 M with the same magnification 10k×. The scale bars in (a,d,e) is 1 μm while in (b,c,f) it is 2 μm.

Acquisition of Raman Spectra and SERS Enhancement Factor

All of the Raman spectra were acquired with a portable Raman system (BWTEK) in the spectral range of 170–2000 cm–1. A solid-state laser operating at 785 nm with power scalable in the range of 3–300 mW was used as the excitation source. The Raman system is equipped with a micropositioning system for fine XYZ adjustments. SERS data were collected with a laser power of ∼25 mW from an average of three spectra, and the integration time used was 5 s. Before the Raman measurements were initiated, the intensity of the Raman peak at 520 cm–1 from silicon was verified for each data acquisition. Initially, the Raman spectra were acquired on plane FP, and the peak noticed at ∼1094 cm–1 is attributed to the C–O–C bending mode of cellulose fibers.[48,49] High scattering was observed from the FP because of its porous structure and the random cellulose matrix. After soaking, a reduction in the Raman intensity of the 1094 cm–1 mode was observed which could possibly be due to the distribution of Ag/Au NPs over the cellulose fibers; the Raman spectra of FP with and without NPs are shown in Figure S9. The EF was estimated using the relation EF = (ISERS × CR)/(IR × CSERS), where ISERS is the Raman signal intensity with NPs on FP, IR is the Raman intensity on FP (without NPs), CSERS is the concentration of sample on NP substrates (low concentration), and CR is the concentration of sample (0.1 M) which produces the Raman signal (IR). The bulk Raman spectra of PA, DNT, NTO, and MB were recorded with a concentration of 0.1 M on plane FP without loading the NPs and were acquired at the same conditions (Figure S10).

Results and Discussions

Salt Aggregation Effect

To reduce the preparation time of paper-based SERS substrates, the FP was soaked in aggregated NPs, which decreased the soaking time from few hours to few minutes (in the present study, it was ∼30 min) compared to earlier reports. Because of the aggregation of NPs, generated hotspots will be more, which will play a vital role in the enhancement of SERS.[30,50,51] Jana et al. proposed that the SERS signal will be affected based on salt-induced particle aggregation.[52] They suggested three reasons: (a) an increase in the EM field at the junction between the particles, (b) CT also possibly contributes to the enhancement, and (c) absorption/reorientation of the analyte is likely because of the induced anion. Initially, the SERS spectra of MB were collected in the presence of NaCl. MB is a common dye mainly used in textiles, bacteriology, and also causes eye and skin irritations and is toxic to microorganisms. Different SERS substrates were investigated to understand the effect of aggregation: (i) FP soaked in aggregated Ag/Au NPs solution (with NaCl) and (ii) pure colloidal NPs (without NaCl). The photo of soaked FP with and without NaCl are mentioned in the schematic (inset of Figure ). The color of aggregated NPs (mixed with NaCl) loaded on FP was much darker than that of the pure NPs (Figure ). It indicates that NaCl has greatly improved the aggregation capacity of NPs. The SERS spectra can be achieved from coherent oscillations of conduction electrons near the surface of noble metal NPs, which is known as the LSPR. The LSPR in NPs is sensitive to dielectric environment, making it an excellent candidate for obtaining huge enhancements.[9] The aggregation of the metal NPs was influenced by environmental parameters (pH and ionic strength), and external factors such as light and heat also effect NP aggregation.[27] The SERS spectra of MB with and without the addition of NaCl are illustrated in Figure S11. Because of the scaffold region of porous and fibrous FP, the NPs were immobilized leading to SERS signal without a salt solution. Furthermore, the presence of NaCl induced the aggregation of NPs rendering a narrow interparticle gap of adjacent NPs, which cooperates with plasmons in improved SERS enhancement. Nobel metal NP aggregates on FP support a large number of plasmonic hotspots within the wide range of hotspots which yielded highly enhanced SERS signals.[53] We could observe ∼three times increment in the Raman signals from the salt-induced aggregation of the plasmonic NPs (Ag: 2451, Au: 889) than that of the paper-induced aggregation of NPs (595) for the 1620 cm–1 significant Raman peak of MB. Strongest SERS signals were observed in the case of FP loaded with aggregated Ag NPs compared to Au NPs, and this could be attributed to the strong optical excited plasmonic effect offered by the Ag NPs.

Figure 6

Schematic representation of explosive detection by paper-based SERS substrates (Ag/Au-aggregated NP-loaded FP). Inset shows the photograph of bare and Au NP-loaded FP.

To optimize the NaCl concentration, the achieved SERS substrates with Ag/Au NPs mixed with various concentrations of NaCl solution were investigated with MB as a probe molecule. Initially, FP embedded with aggregated Ag NPs prepared using various concentrations of NaCl (1 mM, 10 mM, 50 mM, 100 mM, 500 mM, and 1 M) was evaluated by collecting the SERS signals of MB (5 μM), and the spectra are presented in Figure a. It is well-known that the intensity of the SERS signal is affected by the plasmonic coupling of the metal interparticle and also depends on the concentration of NaCl. While increasing the NaCl concentration from 1 mM to 1 M range, the lowest EF (∼1.9 × 104) was observed at 1 mM NaCl, and this is expected because the NP aggregation is the least. We could expect more hotspots with increasing NaCl concentration but interestingly, the maximum EF (∼4.3 × 105) observed for 50 mM NaCl, an intermediate state of aggregation. The decrease in the EF to ∼7 × 103 when the NaCl concentration increases, from 50 mM to 1 M, might be because of the over-aggregation of NPs. The over-aggregated Ag NPs would have affected the homogeneity of the substrate and played a key role in achieving inferior SERS signal. Furthermore, at higher concentrations, the over-aggregation probably decreased the number of SERS-active hotspots. Figure b presents the EFs of the characteristic mode of MB (1620 cm–1) with respect to the NaCl concentration. The homogeneity of the SERS signal is a significant parameter in evaluating the practicality of the SERS substrates. To test the reproducibility of this optimized SERS substrate (FP + AgNPs + NaCl 50 mM), more than 10 spectra were acquired at randomly selected sites on the paper substrate for MB (5 μM). All of the typical vibrational modes observed in this study are in good agreement with the results reported previously.[54] In certain cases, a small peak shift was noticed and exhibited selective enhancement, and this could be because of the orientation of molecules on the paper substrates. The results of the calculated relative standard deviation (RSD) values for three characteristic peaks of 449 cm–1 (skeletal deformation of C–N–C), 1031 cm–1 (in-plane bending of C–H), and 1620 cm–1 (ring stretching of C–C) were 7.97, 13.08, and 11.22%, respectively. The single-peak intensities and their RSD histograms are shown in Figure S12. From the observed results, we concluded that FP soaked in Ag NPs (50 mM NaCl) demonstrated highly sensitive and reproducible SERS response. It is to be noted that at higher concentrations (of NaCl), there is a possibility of AgCl formation on the surface of NPs, thereby affecting the efficiency of EM enhancement, and this is reflected in the observed lower EFs at higher concentration (Figure b). However, detailed studies will be essential to confirm this conjecture. Similarly, the other highly efficient paper-based SERS substrates were fabricated using gold NPs. The SERS response of MB was evaluated for the FP loaded with the aggregated Au NPs achieved at various concentrations of NaCl. Figure a represents the SERS spectra of MB (5 μM) detected using FP loaded with Au NPs with different NaCl concentrations. The EF was calculated for the 1620 cm–1 mode and plotted as a function of concentration, data of which are shown in Figure b. At low concentrations (1 and 10 mM), the obtained EF was ∼1 × 104. While increasing the NaCl concentration to 0.5 and 1 M, low SERS signal was observed and the corresponding EF was ∼103. The maximum EF (∼7.6 × 104) was achieved at 50 mM NaCl, and low SERS signals were observed when the NaCl concentration exceeded 50 mM and finally reached 1 M. The high salt-concentration factor could easily lead to the aggregation and precipitation of NPs on the FP. Thus, moderate addition of NaCl concentration promotes the aggregation of NPs. Sufficient amount of halide ions is required to achieve higher number of hotspots, that is, less amount of chloride ions would not overcome the electrostatic repulsion between the NPs and more amount induces excessive aggregation or precipitation of the NPs.

Figure 7

(a) SERS spectra of MB (5 μM) acquired from FP loaded with Ag NPs at different concentrations of NaCl and (b) corresponding EF at 1620 cm–1 as a function of concentration.

Figure 8

(a) SERS spectra of MB from Au NP-based FP with different concentrations of NaCl and (b) corresponding EF at 1620 cm–1 as a function of concentration.

The reproducibility of SERS substrate (FP + Au NPs + NaCl 50 mM) was examined by collecting the SERS spectra from more than 10 randomly selected sites on paper substrate for MB of 5 μM. More ordered and uniform Au NPs are observed in FESEM images, enabling high SERS performance and lower RSD values. The calculated RSD values for the prominent peaks of 446 and 1620 cm–1 were 5.78 and 7.37%, respectively (Figure S13). It can be concluded that FP soaked in Au NPs with 50 mM NaCl demonstrated strong SERS signal and relatively lower RSD values. The optimum condition for the best SERS performance was observed at 50 mM NaCl concentration for both Ag/Au NP paper-based substrates. The optimized Ag substrates presented in this work depicted higher EFs (∼4.3 × 105) than the Au substrate EFs (∼7.6 × 104). It is worth mentioning that all of the SERS spectra were recorded with a portable Raman spectrometer, useful for on-field applications. Therefore, it is reasonable that increasing the excitation spot size potentially ensures the possibility for rapid, on-site detection of explosives. The sizes of the nanoaggregates on the paper are much smaller compared to the Raman system laser beam size and, hence, the SERS signals are an average over thousands of NPs aggregates.

SERS: Detection of Dye (MB) Molecule

As a result of great sensitivity and good reproducibility of aggregated Ag/Au NPs (mixed with 50 mM NaCl), FPs with that particular concentration were chosen as appropriate SERS substrates for detection of explosive/dye molecules. A systematic study was performed to record the spectra of MB (concentrations ranging from 5 μM to 5 nM) using aggregated Ag NPs (mixed with 50 mM NaCl) incubated on FP, and the data are shown in Figure a. A good correlation was observed between the characteristic peak intensity of 1620 cm–1, and the amount of MB was observed (5 × 10–6 to 5 × 10–9 M). Figure b shows the plot where y-axis corresponds to the Raman peak intensity at 1620 cm–1 and X-axis corresponds to the MB concentration (in log scale) and the R2 retrieved was ∼0.98. Thus, the intricate fiber network with the decoration of aggregated plasmonic NPs enabled the formation of high-density “hotspots”. A similar study was performed using aggregated Au NPs (mixed with 50 mM of NaCl) loaded on FP for MB (5 μM to 5 nM), and the data are depicted in Figure S14a. The corresponding linear calibration curves were constructed by monitoring the intensity of the strong 1620 cm–1 spectral feature as a function of analyte concentration, and the data are shown in Figure S14b. The log–log plot also exhibited a good linear relationship with an R2 value of 0.997, demonstrating the potential application of paper-based SERS substrates. The greater SERS sensitivity and highest EF of Ag NPs (EF ≈ 3.4 × 107) compared to Au NPs (EF ≈ 7.9 × 106) for MB 5 nM can be attributed to the superior plasmonic properties and scattering cross sections. Real-time detection of trace explosives is a major concern all over the world. Paper-based SERS targets for explosive[49] detection combined with a portable Raman instrument could provide a low-cost and fast on-site solution. The SERS performance of optimized FP SERS substrate (aggregated Ag NPs at NaCl 50 mM) was examined with the explosive molecules such as PA, DNT, and NTO (nitroaromatic explosives), which exists in trace amounts in water and soil. DNT is environmentally a more stable decomposition component of TNT, which was used in buried landmines for military and terrorist activities. NTO has a similar explosive performance to RDX, but it is an insensitive explosive.

Figure 9

(a) SERS spectra of MB with various concentrations from 10–6 to 10–9 M and (b) corresponding linear calibration curves constructed by monitoring the intensity of the strong 1620 cm–1 using FP loaded with Ag NPs (50 mM NaCl).

SERS: Detection of Explosives

The SERS spectra of PA at different concentrations, 5 mM, 50 μM, and 5 μM, were recorded with optimized Ag paper substrate, and the data are shown in Figure a. It clearly shows the C–H bending mode at 821 cm–1 and NO2 symmetric stretching mode at 1344 cm–1 of PA.[70] Even at a low concentration (5 μM), the characteristic mode of 821 cm–1 was clearly visible. The EF was estimated to be ∼2.5 × 104 which is comparable to that PA detection with Ag nanotriangle-loaded FP.[60] Similarly, the as-prepared Ag NPs loaded paper substrate verified with the other explosive molecule DNT. The SERS spectra of DNT (5 mM, 50 μM, and 1 μM) shown in Figure b. The characteristic peak of DNT 858 cm–1 was chosen to calculate the EF and estimated as ∼1.9 × 104. We next examined with another explosive molecule NTO (1 mM, 100 μM, and 10 μM) shown in Figure c. The characteristic peaks of NTO 845, 1305, and 1383 cm–1 are assigned to NO2 symmetric stretching and deformation.[71] The EF was estimated for NTO using Ag paper-based SERS substrate is ∼2.1 × 104 for the 1383 cm–1 mode. The reproducibility measurements were also carried for NTO (10 μM) by recording the SERS spectra at 16 different randomly selected points on the optimized paper substrate, shown in Figure a. The Raman peak shifts and assignments of all of the probe molecules are summarized in Table S1 (Supporting Information) and the observed Raman modes are in good agreement with the previous reports.

Figure 10

(a) SERS spectra of explosives: (a) PA [(i) 5 mM, (ii) 50 μM, and (iii) 5 μM], (b) DNT [(i) 5 mM, (ii) 50 μM, and (iii) 1 μM], (c) NTO [(i) 1 mM, (ii) 100 μM, and (iii) 10 μM] by using FP embedded with Ag NPs aggregated by optimized concentration of NaCl (50 mM).

Figure 11

(a) SERS spectra of explosive (NTO) from different spots within paper substrate with Ag NPs at 50 mM NaCl and (b) histogram of the peak intensity at 1383 cm–1 with RSD of 9.78%.

The SERS signal intensity at 1383 cm–1 mode from different spots given an RSD of 9.78% revealing high spot to spot reproducibility, which may result from the uniformity of hotspots. The histogram of SERS intensity (1383 cm–1) is shown in Figure b. Surprisingly, no significant variations in the EF for PA (∼2.4 × 104), DNT (∼2 × 104), and NTO (∼2.1 × 104) were observed by employing the same optimized Ag substrate (FP + Ag NPs + NaCl 50 mM), and the data are shown in Figure S15. Similarly, FP loaded with Au NPs at NaCl 50 mM concentration substrate was also used to detect an explosive DNT (10 μM), as shown in Figure S16. The observed EF was ∼1.5 × 104. The reproducibility of the SERS signal was also verified, the data of which are shown in Figure S17a with an RSD value of 15.58% for the characteristic peak 855 cm–1 (illustrated in Figure S17b). In the case of explosive detection too, better EF was observed for Ag paper substrates than Au substrates, and this could possibly be attributed to the UV broadening [(Ag: 400–700 nm); (Au: 500–600 nm)] as well as the higher mean size. Polavarapu et al.[29] recently demonstrated the effect of Raman excitation wavelength (532, 633, and 785 nm), size, and shape of different NPs (Au NP, Au NR, and Ag NP inks on paper) on SERS intensities and observed the superior EFs for the substrate loaded with Ag NPs. Further detailed studies are required to comment on the wavelength effect in the Au case. We strongly feel that the wavelength effect in the Au case was overpowered by the size effects and the number of hotspots.

In real-world scenarios, the detection of multicomponent analyte has a great value and there is urgent need for the development of low-cost SERS substrates for such detection capabilities.[72] Here, we present our initial results demonstrating the capability and versatility of the fabricated paper substrate to detect complex molecules (a mixture of PA and MB). The mixture of dye and explosive molecule, consisting PA with 5 μM and MB with 50 nM concentrations in equal volumes, was prepared, and the SERS spectrum was recorded using the paper substrates with Ag NPs. The SERS signature of the explosive molecule is dominated by dye molecule signature, as shown in Figure . However, we could clearly distinguish both (dye and explosive molecule) the Raman peaks. The sensitivity of the low-cost flexible paper substrate can be further improved by adopting different shapes and sizes of the plasmonic NPs to achieve superior enhancements toward the detection of various explosive mixtures or combination of explosives and common substances. The EFs obtained in this study are superior to some of the reported literature (summarized in Table ), whereas some of them demonstrated better performance than ours. Nevertheless, it is to be noted that we have tested four different analyte molecules (including three different explosives) and utilized a portable Raman spectrometer for our studies providing ample evidence for the versatility of our substrates. Additionally, the preparation methodology of our SERS substrates is simple, robust, and green, unlike some of the reported techniques, which are quite complicated and involve multiple steps. Our future endeavor would be to optimize these SERS targets by incorporating higher density of NPs, testing these with different sized (probably smaller sized) NPs, and also test them with alloy NPs. Our earlier studies have clearly advocated that alloy NPs have substantially improved the SERS signals by an order of magnitude.[54,70,73] We also plan to test different paper grades for efficient loading of these NPs.

Figure 12

SERS spectra complex molecule (PA 5 μM + MB 50 nM) obtained with FP + Ag NPs + NaCl 50 mM.

Table 1

Summary of the Results Obtained from Paper-Based SERS Substrates Reported in the Literature

s. no.	NPs & size	SERS substrate	fabrication method	analyte	EF	spectrometer (wavelength)	refs
1	Ag	Whatman chromatography paper (Grade 1 CHR)	silver mirror reaction	Rhodamine (Rh6G)—10–11 M p-aminothiophenol (p-ATP)—10–9 M	p-ATP—105	HORIBA Jobin Yvon (532 nm and 632 nm)	(55)
2	Ag (60 nm)	GSM grade paper	pipetting NPs	RH6G—0.1 nM, Malachite green (MG)—1 nM	MG—107	Table Top Raman spectrometer (785 nm)	(18)
3	Ag	Fisher brand chromatography paper	inkjet printer	Rh6G—10 nM	105	HORIBA Jobin Yvon HR-VIS Raman microscope (632.8 nm)	(56)
5	Ag	FP	silver mirror reaction	Crystal Violet—10–9 M Thiram—10–7 M		microscope Raman system—DeltaNu	(57)
6	Ag	Whatman No 3 grade FP	soaking in NaCl	seminal plasma Rh6G	107	Micro Raman System (785 nm)	(58)
7	Ag	brushing	microfluidic paper	Rh6G—1 nM MG—10 nM	107	DXR Raman microscope (632.8 nm)	(59)
8	Au NRs (60 nm long and 18 nm dia.)	Whatman FP grade 1	drop cast	trans-1,2-bis(4-pyridyl)ethane (BPE)—0.5 nM	5 × 106	Raman system (785 nm)	(25)
9	Ag NPs & Ag NSs (stars)	Whatman FP grade 1	drop cast	Rhodamine (Rh6G)	107	HORIBA Jobin Yvon (633 nm)	(31)
10	Ag nano triangles	FP	immersion	PA—10–6 M PATP—10–8 M		Titan Electro Optics Raman spectrometer (532 nm)	(60)
11	Ag nanowires	Millipore-mixed cellulose membrane	filtrating of Ag NS through hydrophilic filter	CV—1.0 × 10–8 mol·L–1 BPE—5.0 × 10–9 mol·L–1	7.7 × 106 1.4 × 106	Renishaw InVia Raman microscope (532, 633 and 785 nm)	(61)
12	Au nanorods (5–100 nm)	Whatman FP grade 1, swab	immersion	1,4-benzenedithiol BDT—0.1 nM	5 × 105	HORIBA confocal Raman spectrometer (785 nm)	(49)
13	Au NPs (5–100 nm)	Whatman FP grade 4	drop cast	trans-1,2-bis(4-pyridyl)ethylene	3 × 104	HORIBA confocal Raman Spectrometer (785 nm)	(62)
14	Au NPs	Whatman chromatography FP grade 1	ink jet printing	carboxymethyl cellulose & AuNPs (7:1), Rhodamine (RH6G)	1.8 × 104	SENTERRA confocal Raman spectrometer—(785 nm)	(63)
15	anisotropic Au NPs (20 nm)	FP	soaking	benzenedithiol (BDT)—1 μM steronoin—100 nM		WITec alpha300 Raman spectrometer (785 nm)	(51)
16	Au NPs	FP	dropped	mucin—1 nM		confocal Raman spectrometer (Renishaw—633 nm)	(64)
17	Au nano rods	FP	dipping	cancer screening		Raman microscope (Renishaw—785 nm)	(65)
18	Au NPs	Whatman FP grade 1	thermal inkjet technology	2,4,6-trinitrotoluene—94 pg DNT—7.8 pg, 1,3,5-trinitrobenzene—0.89 pg		portable spectrometer (R-3000QE, Agiltron)—785 nm	(66)
19	Au @ Ag 30 nm Au core & 7 nm Ag shell	Whatman 1 qualitative FP	printing	Thiram—10–9 mol L–1		micro Raman system (785 nm)	(67)
20	GO @ Ag NPs @ paper	Whatman 1 chromatography paper	immersion and dip coating method	RH6G—10–19 M CV—10–20 M		HORIBA HR evolution 800 Ar excitation laser (532 nm)	(68)
21	Au NPs (15–120 nm) Au NRs Ag NPs (50–80 nm)	A4 paper FP	pen-on-paper	LOD < 10 attomoles-dye < 10 ppb—pesticide		Micro-Renishaw InVia reflex system (532, 633, and 785 nm)	(29)
22	Au porous nanospheres (55 nm)	laboratory FP	soaking	Rhodamine (Rh6G)—10 nM	2.3 × 106	Renishaw Invia Raman (633 nm)	(69)
23	aggregated Ag/Au NPs	Whatman FP grade 1	soaking	MB (5 nM), PA (5 μM), DNT (1 μM), NTO (10 μM)	3.4 × 107, 2.4 × 104, 2 × 104, 2.1 × 104	portable (BWTEK) 785 nm	present study

Conclusions

We have successfully fabricated ecofriendly SERS substrate by using low-cost FP as a base and is embedded with NPs that were also fabricated by a green technique LAL. The common FP with aggregated Ag/Au NPs provided a simple and robust approach for SERS substrate preparation. Specifically, spherical Ag and Au NPs were loaded for an optimally aggregated configuration to yield maximum signal enhancements in the SERS measurements. Our results from detailed experiments presented here indicate that the Ag/Au NPs with 50 mM NaCl concentration is helpful for optimal SERS performance. From the result obtained, we conclude the paper-based Ag substrate performance better than paper-based Au substrate. The optimized active SERS paper substrate is helpful for the detection of four adsorbed molecules (MB-5 nM, PA-5 μM, DNT-1 μM, and NTO-10 μM) with high specificity, sensitivity, and reproducibility. These concentrations in terms of mass represent 31.98 pg, 11.45 ng, 1.82 ng, and 13.06 ng, respectively. Overall, we report here a simple, inexpensive technique for creating plasmonic substrates that are readily integrated with portable spectrometer for rapid SERS analysis. It offers to decrease the cost of analysis using paper-based SERS targets and utilizing a portable Raman instrument, which is portable to the point of diagnostics of interest. Furthermore, the NPs produced by laser ablation are proven to be stable for long periods of time.[44] However, the stability of these NPs when loaded in the FP is being investigated thoroughly. We strongly believe that following further optimization studies, the combination of low-cost and high-performance paper SERS devices will be an ideal platform to enter on-site analysis, beyond the research laboratories, for practical applications related to explosives detection.

Experimental Details

Materials

Whatman Grade 1 FP with porosity 11 μM and thickness 180 μM, silver (>99% pure) and gold targets (>99% pure) of thickness 1 mm, and NaCl compound (molar mass 58.44 g/mol) was used for the preparation of different concentrations; methanol (reagent grade), acetone (reagent grade), and MB (C16H18ClN3S·H2O) were purchased from Sigma-Aldrich. The explosive molecules of 2,4,6-trinitrophenol (PA, C6H3N3O7), DNT (C7H6N2O4), and NTO (C2H2N4O3) were provided by HEMRL (Pune, India) for SERS detection.

Synthesis of Ag/Au NPs

Ag/Au NPs were synthesized by LAL technique. Pure Ag and Au substrates were cleaned with acetone, ethanol, and deionized water in an ultrasonic cleaner for 15 min to remove any organic dopants present on the surface. The samples were then ablated with a regenerative fs (Ti:sapphire) amplifier (∼50 fs, 800 nm, 1 kHz). The incident laser pulses were focused through a planoconvex lens of focal length 100 mm onto the target surface submerged in 10 mL of distilled water and placed in a glass beaker. The focus was adjusted coarsely by using a low-power laser and later finely by monitoring the cracking sound and the formation of the bright plasma at the solid–liquid interface. The beaker was mounted on a computer-controlled translational stage and scanned with 100 μm/s speed along X and Y directions, so as to avoid repeated ablation at a single spot.[45,73] The incident laser energy was adjusted to ∼500 μJ and the ablation time was typically 40 min for both of the cases. When the laser pulses are incident on the target surface, a series of processes will undergo such as ejection of electrons, plasma plume formation, and the creation of cavitation bubble within a short duration.[54,74,75] Finally, the particles are generated through the nucleation and growth process and suspended in liquid as colloids. This is a “green” method used to produce NPs in liquid without stabilizing molecules or ligands.[45,76]

Salt-Induced Aggregation of Synthesized Ag/Au NPs

Alkali halides[77] (NaF, NaCl, NaBr, NaI, KCl, KBr, etc.), NaNO3, and K2SO4 boost the aggregation in colloidal NPs.[52,78] In the present work, we have used NaCl as an aggregation agent. Furthermore, the concentration of aggregation agent and ratio with NPs have been demonstrated to play a vital role in the SERS intensities. In the present case, different concentrations of NaCl (1 mM, 10 mM, 50 mM, 100 mM, 500 mM, and 1 M) mixed with 1:1 ratio of pure Ag/Au NPs. After addition of NaCl, we noticed the change in the color of NPs [Figure S1 (left) shows the FP before soaking].

SERS Substrates Preparation

A Whatman Grade 1 FP composed of almost (∼98%) α-cellulose, which is abundant in nature, biodegradable, biocompatible, and recycled. The large fibers were made of micro- and nanocellulose fibers interweaving with each other and were normally porous in nature providing a large surface area for localizing the plasmonic NPs. The penetration of NPs into the paper depends on various factors such as the porosity (pore size), thickness, and hydrophobicity.[18] FP was cut into small strips (1 cm2) and then soaked in the prepared NP colloids. The association of the NPs on to the paper was obvious because the color of NPs colloidal solution decreased and became colorless, which means they essentially have got transferred to the FP. The color of the colloidal NPs [Figure S1 (right) after soaking FP] and the soaked FP loaded with Ag/Au NPs at different concentrations of NaCl were also modified (Figure S2). The density of the NPs will change after aggregation (though the number of total NPs remains the same in each case) and further soaking the FP in these solutions (optimized concentration of NaCl) have led to increased darkness in the observed color. At higher concentrations, the loading of NPs was lower because of overaggregation. The time of soaking was ∼30 min uniformly for all of the substrates studied. However, this time also affects the loading of NPs and has been studied extensively by Huang et al.,[58] wherein they observed that soaking for 1 h procured excellent SERS results. The absorption of NPs on a substrate will enhance the EM field in sensing because of its ability achieved through the excitation of LSPR. The soaked FP substrate was cut into four pieces (i.e., 2.5 mm2) and allowed to dry. Subsequently, 10 μL of the volume of analyte was pipetted on to the paper. Schmucker et al., from their detailed studies, suggested that the NPs immobilized on the paper remain stable under a variety of complex environmental conditions.[79] A summary of literature reports on paper SERS substrates fabrication techniques and their sensing applications are presented in Table .