A green approach to nanoplastic detection: SERS with untreated filter paper for polystyrene nanoplastics

Abstract

Plastic pollution at the nanoscale continues to pose adverse effects on environmental sustainability and human health. However, the detection of nanoplastics (NPLs) remains challenging due to limitations in methodology and instrumentation. Herein, a “green approach” for surface-enhanced Raman spectroscopy (SERS) was exploited to detect polystyrene nanospheres (PSNSs) in water, employing untreated filter paper and a simple syringe-filtration set-up. This SERS protocol not only enabled the filtration of nano-sized PSNSs, which are smaller than the pore size of the ordinary filter paper, but also offered SERS enhancement by utilizing quasi-spherical-shaped silver nanoparticles (AgNPs) as the SERS-active substrate. The filtering of NPLs was accomplished by adding an aggregating agent to the nanoparticle mixture, which caused the aggregation of NPLs and AgNPs, resulting in a larger cluster and more hot spots for SERS detection. The optimal aggregating agent and its concentration, as well as the volume ratio between the AgNPs and NPLs, were also optimized. This SERS method successfully detected and quantified PSNSs of various sizes (i.e., 100, 300, 460, 600, and 800 nm) down to a limit of detection (LOD) of about 0.31 μg mL⁻¹. The method was also validated against the presence of several interferents (i.e., salts, sugars, amino acids, and surfactants) and was proven practical, as evidenced by the detection of 800nm PSNSs in drinking and tap water (LODs of 1.47 and 1.55 μg mL⁻¹, respectively).

---

Introduction

The breakdown of plastic garbage in the environment has released an enormous volume of small plastic particles into the surrounding environment.¹ Plastics are broken down into microplastics (less than 5 mm–1 μm)^2–4 and nanoplastics (less than 1 μm–1 nm)^5,6 by mechanical,^7,8 photo-oxidation,⁹ and hydrolysis processes.^7,9 Because of their low density and resistance to degradation, micro/nanoplastics can spread globally via atmospheric wind and ocean currents,¹⁰ as evidenced by micro/nanoplastics being detected in remote areas far from humans, such as Mount Everest,^11,12 the Mariana Trench,¹³ the Arctic region,¹⁴ and the Antarctic region.¹⁵ Nowadays, micro/nanoplastic contamination is a serious issue in a variety of environmental settings.¹⁶ Because of their small size, micro/nanoplastics are readily swallowed by living species and accumulate throughout the food chain at all trophic levels, causing various problems in aquatic habitats.¹⁷ Nanoplastics (NPLs) are typically smaller than microplastics. Therefore, their cytotoxicity, bioavailability, and health risks have more adverse impact than the microplastics.¹⁸ Micro/nanoplastics may pass across the biological barrier and translocate among the tissues,¹⁹ eventually accumulating in the cell and causing malfunction and damage to the tissue.²⁰ Microplastics are typically eliminated through the gastrointestinal system, while NPLs have a relatively small diameter, which allows them to readily pass through vessels into the bloodstream and pose a serious threat to human health.²¹

Several analytical methodologies have been developed to detect NPLs in laboratory and real-world environments. Optical microscopy (OM),²² scanning electron microscopy (SEM),²³ and transmission electron microscopy (TEM)²⁴ are visualization techniques that are always used to monitor the size, shape, and distribution of NPLs. These visualization techniques cannot differentiate NPLs from plastic-like particles due to their failure to provide the chemical composition of NPLs,²⁵ which is why the visualization technique is generally used alongside a chemical identification technique. The thermo-analytical methods such as pyrolysis gas chromatography-mass spectrometry (Py-GC-MS)²⁶ and thermal desorption-gas chromatography-mass spectrometry (TD-GC-MS)²⁷ are the techniques that typically provide the chemical composition of NPLs for identification. Still, the drawback is that this technique destroys the sample analytes,²⁸ which is impractical for real NPL characterization. Therefore, vibrational spectroscopic methods such as μFTIR^29–31 and Raman spectroscopy^32–36 are widely used because they provide information about chemical compositions for qualitative analysis. These techniques in combination with microscopic techniques also offer morphological information, and give high spatial resolution;³⁷ they are categorized as a non-destructive technique. Practically, μFTIR effectively detects and identifies plastic particles in the laboratory with a diameter down to 10–20 μm.^38–40 Raman spectroscopy is more frequently used for the analysis of NPLs because it can detect plastic particles below the limit of detection of μFTIR (>10–20 μm)^41,42 and is also suitable for the analysis of contaminants such as pigments⁴⁰ and (in)organic⁴³ and biological compounds.⁴⁴ However, detecting plastic particles lower than 1 μm by ordinary Raman spectroscopy is challenging because tiny plastic particles exhibit an intrinsically low Raman scattering signal.⁴⁵

Recently, the surface-enhanced Raman scattering technique (SERS) has been increasingly used due to its recognition as an ultra-sensitive technique for NPL detection.^46–58 The weak Raman scattering signal of tiny plastics, which limits detection by ordinary Raman spectroscopy, can be amplified by the SERS substrate to be detected. Utilizing a metallic nanostructure as the SERS active substrate, the weak signal of NPLs that is absorbed onto the rough metal surface will be enhanced by the phenomenon called electromagnetic SERS enhancement.^59,60 As reported in the literature, the developed SERS substrate-based Ag-coated gold nanostar is allowed to detect PSNSs of 400 nm with an LOD of about 50 μg mL⁻¹.⁵⁰ The LOD can be further improved using a commercial Karité SERS substrate, which has an ordered structure of holes or pits where a robust electromagnetic field benefits the SERS enhancement. This method exhibited an LOD down to 26 μg mL⁻¹ for PSNSs at 360 nm.⁵⁵ Si-O₂ self-assembly sputtered with a gold film is another approach that increases the sensitivity for NPL detection. This approach created a nanowell where NPLs can be deposited and exhibit a strong SERS effect. Utilizing this approach, NPLs with a diameter between 100 and 500 can be sensitively detected down to 5 μg mL⁻¹.⁶¹ Moreover, a glass cover slide sputtered with gold can easily fabricate the SERS active substrate for detection. This approach can detect NPLs down to 100 nm by dissolving them in organic solvent and dropping them onto the prepared substrate. The detection limit provided by this SERS approach can be down to 0.1 μg mL⁻¹.⁶² From the research above, it has been confirmed that the SERS method is sensitive enough to detect NPLs. However, the challenges for NPL detection remain, especially in terms of green, easy to use, low cost, and rapid detection requirements.

Herein, we developed a practical filter paper-based SERS measurement using a simple, low-cost, and single-step syringe-filtration protocol that combines the advantages of colloidal and nanostructured solid systems to detect and rapidly quantify NPLs in water. Filter paper was selected as the SERS substrate carrier in this work because of its biodegradability, portability, eco-friendly material, and availability in the laboratory. It is hypothesized that adding an aggregating agent or salt plays a vital role in SERS enhancement by generating abundant hot spots through salt-induced aggregation. Moreover, the formed aggregates (AgNPs/PSNSs) were captured by the plain, water-absorbent, cellulosic filter paper through its three-dimensional network of fiber pores during the filtration process and, at the same time, separate interference affecting the SERS measurement. Toxic chemicals are not employed, and there is no treatment or modification on the filter paper. This green approach is both efficient and secure in detecting nanoplastics in a variety of water conditions. Moreover, employing salts as an aggregating agent helps uniformity in the distribution of particles on the filter paper surface for SERS detection. The overall sample preparation and Raman acquisition procedure are shown in Scheme 1.

Scheme 1 Schematic illustration of NPL analysis using syringe filtration and SERS techniques.

Materials and methods

Chemicals and materials

Polystyrene (PS) latex beads (Sigma-Aldrich) with various particle sizes (i.e., 100, 300, 460, 600, and 800 nm), silver nitrate (AgNO₃, Merck), sodium citrate (Na₃C₆H₅O₇·2H₂O, Sigma-Aldrich), sodium chloride (NaCl, Merck), potassium chloride (KCl, Carlo Erba Reagents), magnesium chloride (MgCl₂, Acro Organics), calcium chloride (CaCl₂, Carlo Erba Reagents), potassium iodide (KI, Carlo Erba Reagents), potassium bromide (KBr, Carlo Erba Reagents), sodium fluoride (NaF, Ajax Chemicals), sodium nitrate (NaNO₃, Sigma-Aldrich), sodium carbonate (Na₂CO₃, Carlo Erba Reagents), sodium sulfate (Na₂SO₄, Carlo Erba Reagents), aluminum sulfate (Al₂(SO₄)₃, Ajax Chemicals), glucose (C₆H₁₂O₆, Ajax Chemicals), fructose (C₆H₁₂O₆, Ajax Chemicals), sucrose (C₁₂H₂₂O₁₁, Ajax Chemicals), glycine (C₂H₅NO₂, Sigma-Aldrich), sodium dodecyl sulfate (NaC₁₂H₂₅SO₄, Kemaus), dodecyl trimethylammonium bromide (C₁₅H₃₄BrN, Sigma-Aldrich), Triton X-100 (C₁₆H₂₆O₂, Sigma-Aldrich) were used as received. Qualitative filter paper (2.7 μm pore size, 125 mm diameter) was supplied by Advantech, Inc. Ultrapure water (18.2 MΩ cm⁻¹ resistivity at 25 °C) was prepared by a Milli-Q gradient system (Massachusetts, United States). Tap water was obtained from the Faculty of Science, Chulalongkorn University, Thailand. Bottled drinking water was commercially provided by Nestlé Co., Ltd. Water from the Chao Phraya River was collected at Wat Arun ferry pier, Bangkok, Thailand.

Apparatus and equipment

The morphologies of the PSNSs and AgNPs were studied using transmission electron microscopy (TEM, JEOL JEM-2100) at an electron accelerating voltage of 120 kV and a filament current of 61 μA. The particle size and zeta potential were determined by dynamic light scattering (DLS, Zetasizer Nano Malvern Panalytical). The UV-visible spectra and the retention efficiency of plain, untreated filter paper for standard PSNSs were evaluated using a UV-visible spectrophotometer (Thermo ScientificTM Genesys 10S). The distribution of the aggregates deposited on the filter paper was studied using an optical microscope (OM, Zeiss Axio Scope A1). SERS spectra were acquired using a Raman microscope (Thermo ScientificTM DXR).

Synthesis of silver nanoparticles (AgNPs)

A greyish-yellow colloid of AgNPs was prepared as described in our previous report using the citrate-reduction method.⁶³ Briefly, a 500 mL aqueous solution of AgNO₃ (0.018%) was heated to boiling. Then, dropwise, 10 mL of a 1% solution of sodium citrate was added. The mixture was vigorously stirred and kept boiling for an hour. The resulting colloid was cooled to room temperature and then adjusted to 500 mL. The final colloid (∼110 ppm) was stored at 4 °C until further use.

Sample preparation and SERS detection in pure water

In a small glass vial, 1 mL of PSNS suspension (100 μg mL⁻¹ of deionized water) with a size of 800 nm (PS800) was added to 1 mL of the as-prepared AgNPs. The resulting mixture was thoroughly vortexed to ensure that it was homogeneously mixed. Then, 1 mL of an aggregating agent (with an initial concentration of 0.15 mol L⁻¹) was added to the mixture and the mixture was vortexed thoroughly. Thereafter, the 3 mL sample mixture was allowed to flow through the cut filter paper (13 mm diameter) inserted in a removable syringe filter head. After the mixture was filtered, the filter paper was removed and air-dried prior to the SERS measurement. The SERS performance was optimized by varying the type of aggregating agents (i.e., NaCl, KCl, MgCl₂, CaCl₂, KI, KBr, NaF, NaNO₃, Na₂SO₄, Na₂CO₃, and Al₂(SO₄)₃) and concentrations (i.e., 0.05, 0.10, 0.15, 0.20, 0.25, and 0.50 mol L⁻¹) as well as the volume ratio between AgNPs and PSNSs (i.e., 1 : 2, 1 : 1, 2 : 1, 4 : 1, and 6 : 1). Standard calibration curves (working range: 0.5, 1, 2, 4, 6, 8, 10, and 12 μg mL⁻¹) for PSNSs of various sizes (i.e., 100, 300, 460, 600, and 800 nm) were established using the spectra collected from the optimized SERS conditions.

For SERS measurement, the excitation source was a diode laser with a wavelength of 532 nm (with a maximum power of 10 mW) focused on a spot of approximately 2.1 μm diameter on the sample surface using a 10× magnification objective (numerical aperture = 0.25) with a 25 μm pinhole. The acquisition time was set to 32 s for all measurements, and each acquisition was repeated three times at several random sites on the sample surface.

SERS detection in spiked (real environmental) samples

This work studied various real environmental samples: deionized water, tap water, bottled drinking water, and river water. Subsequently, large particles in the river water were removed by a pre-filtration step with filter paper. A 800 nm PSNS (PS800) suspension was prepared and treated with real environmental water samples as the diluent for every stage of the dilution procedure to the required NPL solution concentrations (i.e., 2, 4, 6, 8, 10, and 12 μg mL⁻¹). Following the optimized conditions, sample mixtures were prepared and filtered. SERS detection was performed after the filter paper dried.

Results and discussion

Characterization of AgNPs/PSNSs/salts and filtration efficiency of the filter paper

In the first part of the study, we set the initial concentrations of the aggregating agents and 800nm PSNSs at 0.15 mol L⁻¹ and 100 μg mL⁻¹, respectively. The prepared samples were filtered to evaluate the suitability of ordinary filter paper for filtering NPL aggregates suspended in the solution before SERS detection.

Salt-induced aggregation of the AgNP colloid occurs when the presence of salts in the surrounding environment causes the AgNPs to form aggregates. The interaction between the salt ions and the surface charge of the AgNPs primarily drives this phenomenon. Meanwhile, during the synthesis of AgNPs, the weakly bound citrate ions imparted a negative surface charge onto the AgNPs, preventing their spontaneous agglomeration in the solution.⁶⁴ As shown in Fig. 1a, zeta (ζ) potential analysis revealed a strong negative potential higher than −30 mV,⁴⁶ providing certain colloidal stability for both PSNSs (−36.1 mV) and AgNPs (−40.7 mV) as well as their mixture (PS/AgNPs, −37.1 mV). Interestingly, all the ζ potentials of the SERS samples were less negative than PS/AgNPs, which is the control sample, indicative of weak electrostatic interactions among NPs in the mixture, making them unstable and easy to form the aggregation, as also confirmed by the set of aggregate size for the SERS sample that is prominently larger than the control sample. The TEM images in Fig. 1b–d support our findings. In the absence of salt, the dry state shows individual PSNSs (spherical, ∼800 nm diameter) and AgNPs (quasi-spherical/polygonal, ∼67 nm diameter) with no apparent aggregation. This suggests high colloidal stability in the liquid phase, preventing interaction between the AgNPs and the PSNSs, as evidenced by their physical separation. However, upon adding MgCl₂ salt, the TEM image (Fig. 1e) reveals a dramatic change. The AgNPs become tightly attached to the PSNS surfaces, forming large PS/AgNP clusters. These clusters provide ideal SERS measurement areas due to the enhanced electromagnetic field.

Fig. 1 (a) Hydrodynamic diameter and zeta potential of the prepared samples. TEM images of (b) PSNSs, (c) AgNPs, (d) PS/AgNPs, and (e) PS/AgNPs/MgCl₂ aggregate.

Because the pore size of the filter paper was 2.7 μm, usually, the nano-sized AgNP and PSNS suspension was easily passed through it (Fig. S1a and b†), and only the AgNP/PSNS aggregates were successfully retained on the surface of the filter paper (Fig. S1c†), indicating that the salts significantly improved the filtration process due to the clusters formed by the AgNPs and PSNSs, which are big enough to get stuck within the rough, porous, fibrous, and vast area of the filter paper. However, in the absence of salts, no aggregation and no slight discoloration of the surface of the filter paper was observed, as the PSNSs and AgNPs tend to pass through the pores of the filter paper.

Fig. S2a† shows the UV-visible spectra of the AgNP colloid with and without a salt solution. As revealed in the spectrum, the maximum extinction of the PS/AgNPs without salt was found to be around 450 nm. Whereas the addition of a salt solution to the AgNP colloid leads to a marked decrease in intensity and peak broadening within the original plasmon resonance band, accompanied by the appearance of a redshifted LSPR peak (broad absorptions) towards the near-infrared region to a broader region. Notably, the aggregation of AgNPs in the presence of aggregating agents led to a shift in the color of the solution towards a more opaque or turbid appearance compared to PSNSs and AgNPs alone (Fig. S2b†).

Optimization of the SERS conditions

As an ultra-surface sensitive technique, SERS can only provide effective spectra when the targeted molecules are sufficiently close to the surface of the SERS substrate.⁶⁵ As the TEM image in Fig. 1b revealed, the PSNSs were strongly aggregated or clustered with the AgNPs in the presence of salt, suggesting a significantly higher SERS activity due to abundant plasmonic hot spots. As shown in Fig. 2a, adding KI as an aggregating agent has dramatically improved the SERS sensitivity of PS/AgNPs/KI aggregates compared to the others. In contrast, a weak Raman signal of the PSNSs at 1002 cm⁻¹ was discernible in the absence of aggregating agents, even at a high concentration (100 μg mL⁻¹). This may be because the directly synthesized AgNPs are wrapped by citrate, providing the repulsion force to the PSNS particles, and decreasing contact between them, thereby weakening the SERS activity. The characteristic band of PSNSs was observed strongly at 1002 cm⁻¹, corresponding to the benzene ring vibration mode in the structure of a PS polymer,^53,66 which is selected as the reference peak for our discussion.

Fig. 2 (a) Raman spectra collected from the filter paper with 800nm PSNSs at 100 μg mL⁻¹ prepared under different conditions. (b) SERS spectra of PSNSs with a diameter of 800 nm (100 μg mL⁻¹) in the presence of different aggregating agents (initial concentration fixed at 0.15 mol L⁻¹) and (c) the corresponding SERS intensity at 1002 cm⁻¹.

It was suggested that the salt anions do not directly cause the AgNPs to aggregate but rather repulse the other anions present in the solution (citrate and nitrate), as they compete for metallic surface adsorption with stronger affinities.⁶⁷Fig. 2b and c show the SERS spectra and SERS intensity of the PSNSs in the presence of different aggregating agents. It can be seen from Fig. 2b that not all aggregating agents were suitable for the PS/AgNP agglomeration process, such as sulfate salts (SO₄²⁻). It has a much weaker binding affinity to Ag atoms compared to citrate. Thus, it would not displace the adsorbed citrate from the AgNPs, causing the SO₄²⁻ band at 1004 cm⁻¹ to interfere with the spectrum of the PSNSs at 1002 cm⁻¹.⁶⁸ Therefore, sulfate salts (Na₂SO₄ and Al₂(SO₄)₃) were not considered for use in further studies.

Moreover, using NaCl, CaCl₂, MgCl₂, KCl, KBr, and NaF results in a significantly different SERS intensity. Cl⁻ has appropriate adsorption ability compared to F⁻ and Br⁻ and also has a stronger chemical affinity to AgNPs than citrate ions (also called the “chloride ion effect”),⁶⁹ and thus, can be adsorbed well on the AgNP surface and provides good SERS intensity. However, the SERS signal intensity with I⁻ was significantly lowest owing to its low electronegativity compared to the other halides. Furthermore, as supported by previous studies,⁴⁷ the co-adsorbed PSNSs via the electrostatic interaction with I⁻ might have made a certain contribution to the SERS signal but was negligible, as demonstrated by the almost disappeared SERS signal of PSNSs for KI salt. It could be attributed to the lack of coverage on the AgNP surface by I⁻ or the steric effect brought about by the difference in the spatial structure of the polyatomic ligands such as NO₃⁻ and CO₃²⁻ for the NaNO₃ and Na₂CO₃ salts, leaving the gaps suitable for this species to co-adsorb and hinder the direct contact of the target PSNS particles on the SERS-active AgNP surface.⁴⁷ Such phenomena indicate that the presence of aggregating agents affects not only the aggregation state of the AgNPs but also how PS particles adsorb onto the AgNP surface. In the case of a cation, the aggregation depends on the valence of the counter ion. The higher the valence of the counter ion, the lower the critical coagulation concentration (CCC) and the charge neutralization capacity, that is why the SERS intensity of the NPLs was stronger for tri (Al³⁺) and di (Ca²⁺ and Mg²⁺) valent cations than monovalent (Na⁺ and K⁺) cations.⁷⁰ Therefore, the most preferable SERS signal enhancement was achieved with a trivalent cation. However, as discussed above, due to the uncertainty of this salt, the chloride salt with a divalent cation with good SERS intensity, i.e., MgCl₂, shown in Fig. 2c, was selected for further experiments instead.

Since it is proved that the intensity of the SERS signal is affected by the plasmonic coupling of the AgNPs, more hot spots can be expected with an increase in salt concentration. Hence, the effect of MgCl₂ concentration on SERS performance was studied. Initially, filter paper embedded with aggregated AgNPs prepared using a series of MgCl₂ concentrations (i.e., 0.05, 0.10, 0.15, 0.20, 0.25, and 0.50 mol L⁻¹) was evaluated by collecting the SERS signals of PSNSs (100 μg mL⁻¹), respectively. The SERS performance was evaluated by the peak intensity of the PSNS signal at 1002 cm⁻¹. Generally, as the concentration of the aggregating agents increases, the SERS intensities for PSNSs increase and reach their eventual maxima before experiencing a decline. Still, interestingly, the maximum enhancement was observed at 0.15 mol L⁻¹. As presented in Fig. 3a, the SERS intensity of the PSNS spectra at 1002 cm⁻¹ increased significantly with an increase in MgCl₂ concentration (0.05–0.15 mol L⁻¹) but gradually decreased as the concentration continued to increase beyond 0.15 mol L⁻¹ as also presented in the bar graph in Fig. 3b. The sudden decrease might be because of the over-aggregation of AgNPs, which would have affected the homogeneity of the aggregates and probably decreased the number of SERS-active hotspots. These results were consistent with our observations and literature reports.^52,53,66 Thus, the best SERS performance was obtained when the final concentration of the MgCl₂ solution was 0.15 mol L⁻¹.

Fig. 3 (a) SERS spectra of 800nm PSNSs (100 μg mL⁻¹) with different MgCl₂ concentrations and (b) the corresponding SERS intensity at 1002 cm⁻¹. (c) SERS spectra of the 800nm PSNSs (100 μg mL⁻¹) with the different volume ratios of PSNSs and AgNPs (∼110 ppm) and (d) the corresponding SERS intensity at 1002 cm⁻¹.

As the feasibility of NPL analysis by SERS was verified, the SERS activity of PSNSs as a function of the amount of AgNPs was investigated to achieve the highest possible SERS enhancement further. As discussed, the optimal SERS signal can be achieved when the PSNSs and AgNPs are maintained at the appropriate ratio. In this study, the SERS spectra of the Ag/PSNSs were collected for various volume ratios (i.e., 1 : 2, 1 : 1, 2 : 1, 4 : 1, and 6 : 1). As shown in Fig. 3c and d, there was a sharp increase in the SERS intensity when the volume ratio of the NPs to the PSNS colloid decreased from 6 : 1 to 1 : 2. It can be observed that the SERS signal is highest when the volume ratio of AgNPs : PSNSs is 1 : 2, demonstrating that even a small amount of AgNPs can sufficiently enhance PSNS signal, while the ratios of 4 : 1 and 6 : 1 cause weak SERS signals due to dense AgNP aggregation. The intensity of the SERS signal increased with the amount of PSNS suspension, which is good for conserving the AgNP colloid. Therefore, the selected volume ratio between the AgNPs and the PSNSs was fixed at 1 : 2 for the following SERS analysis.

Quantitative analysis of PSNSs in water

To verify the quantitative ability and sensitivity of the proposed SERS method, standard calibration curves for PSNSs of various sizes (i.e., 100, 300, 460, 600, and 800 nm) were developed using the spectra collected from the optimized SERS conditions. The LOD of the PSNSs in this work was statistically approximated from the slope value of the standard curve, represented by the following equation:
where the standard deviation (σ) represents the noise. The error bars are the σ of the three measurements. As shown in Fig. 4a, the SERS intensity at 1002 cm⁻¹ of 100nm PSNSs is increased significantly with an increase in concentration and reaching the saturation at the highest concentration (10.0 μg mL⁻¹). Incidentally, at the lowest concentration (0.50 μg mL⁻¹), it still yielded a good signal-to-noise ratio, which can be observed at 1002 cm⁻¹, providing a limit of detection (LOD) down to 0.31 μg mL⁻¹ (see Table S1†). A good linear relationship (R², correlation coefficient = 0.9966) between the SERS intensity at 1002 cm⁻¹ and the concentration of 100nm PSNSs is shown in Fig. 4b. The fitted curve satisfied the linear relationship of y = 30.884x + 81.49, which could be applied to the quantitative analysis of the NPLs in the water samples. Subsequently, quantitative analyses of PSNSs at 300, 460, 600, and 800 nm were performed using this method and exhibited a similar trend with the 100 nm (see Fig. S3a–e†). The 300, 460, 600, and 800nm PSNSs exhibited a good linear relationship in the range of 0.5–8.0, 1.0–8.0, 2.0–10.0, and 2.0–10.0 μg mL⁻¹, respectively. However, the SERS sensitivity tends to be dropped after increasing the PSNS particle size, i.e., 300, 460, 600, and 800 nm (Table S1†), which provides the LOD of about 0.53, 0.76, 1.04, and 1.42 μg mL⁻¹, respectively. This phenomenon could be due to the specific surface area of PSNSs that tends to decrease with an increase in the plastic particle diameter, so there is a slight decrease of surface contact between the PSNSs and the AgNPs, which is a lower SERS hot spot exposure to the NPLs during the analysis. This phenomenon proves that this method adheres to the size-dependent effect of the detection.

Fig. 4 (a) SERS spectra of 100nm PSNSs at various concentrations. (b) A calibration curve of the SERS intensity at 1002 cm⁻¹vs. the concentration of 100nm PSNSs.

The analytical performance of the proposed green SERS method was further evaluated by comparison with other SERS-based methods,^{48–56,61,62,65,66,70,71} as shown in Table 1. The LOD achieved by our green SERS method using untreated filter paper is highly competitive when compared to existing PSNS detection techniques. Notably, our method achieves LODs as low as 0.31–1.42 μg mL⁻¹ for various PSNSs sizes (100–800 nm). Notably, AgNPs can be oxidized. A fresh calibration curve for each batch of samples analyzed using SERS is recommended. This surpasses the sensitivity of many previously reported SERS methods, which often exhibit LODs in the range of 5–50 μg mL⁻¹ for similar PSNS sizes. While some studies have achieved even lower LODs (e.g., 0.05 μg mL⁻¹ with GO/MWCNT-ANS membranes),⁷⁰ these methods often involve more complex procedures or materials. Our approach offers a significant advantage in terms of simplicity and environmental friendliness, as it utilizes readily available filter paper and eliminates the need for hazardous pre-treatments. Overall, this comparison highlights the effectiveness of our green SERS method for sensitive PSNS detection while maintaining practicality and environmental responsibility. The proposed green method was also compared with current conventional methods for NPL analysis. Compared with Py-GC/MS and other mass spectrometry methods, the proposed green method is easy to implement and can keep the size and morphology information of NPLs. Compared with infrared and normal Raman spectroscopy, this green SERS-based method is superior in sensitivity and spatial resolution.

Table 1 The comparison of SERS methods for nanoplastic analysis

SERS substrate	Plastic type and size	Detection environment	LOD (μg mL^−1)	Ref.
Klarite (Au)	PS: 360 nm	Pure water	26.3	55
	PMMA: 500 nm
Ag-coated AuNSs inserted into AAO nanopores	PS: 400 nm	Pure water and spiked samples of tap water, river water, and seawater	50	50
AuNPs assembled on glass slide	PS: 33 and 161 nm	Pure water	10	65
	PET: 62 nm		15
AuNRs and AgNWs on regenerated cellulose films	PS: 84 and 630 nm	Pure water	500	48
			100
AAO/MoS2/AgNPs	PS: 100–300 nm	Pure water	—	51
Self-assembled SiO2 sputtered with Ag films	PS: 100–1000 nm	Pure water	5	61
AgNPs sol (liquid environment)	PS and PE: 100 nm	Pure water and seawater spiked samples	40	52
	PP: 500 nm
AgNPs sol (liquid environment)	PS: 50–500 nm	Pure water and lake water spiked samples	6.25	66
AgNPs sol	PS: 50 and 1 μm	Pure water and river water spiked samples	100	56
Au sputtered glass slide	PS: 100, 300, 600, and 800 nm	Pure water and spiked samples of tap water, mineral water, and river water	0.10–0.26	62
AuNPs	PS: 20 and 200 nm	Pure water	10	49
AuNRs	PS: 350 nm	Pure water	6.5	53
Au nanopore at the tip of a glass nanopipette	PS: 20 nm	Pure water	500	54
	PMMA: 10 nm
3D plasmonic Au nanopocket structure	PS: 1000 nm and PE: 1000 nm	Tap, river, and sea water	2.5	71
GO/MWCNT–AgNS membranes	PS: 50, 100, 200, 500, and 1000 nm	River water, aquaponic, and pond aquaculture water	0.05	70
AgNPs aggregated by salt	PS: 100, 300, 460, 600, and 800 nm	Pure water and spiked samples of bottled drinking water and tap water	0.31–1.42	This work

---

Effect of interference on SERS measurements

Different types of interference were spiked onto the deionized water further to evaluate the viability of our proposed SERS method. In this part, the concentration of the PSNSs was set as 10 μg mL⁻¹ for each particle size, and the concentration of each interference was fixed at 0.1 wt%. As shown in Fig. 5a, all sizes of the PSNSs were successfully detected even in the presence of interference, as also demonstrated from the SERS spectrum in Fig. S4a–e.† Notably, compared with the SERS intensities obtained using the reference (DI water), the intensity of the PSNS prominent peak at 1002 cm⁻¹ significantly increased in the presence of ionic compounds, indicating the efficiency of the surface enhancement effect. Since the thickness of the electrical double layer around the AgNPs decreases as the ionic strength increases, the presence of additional inorganic salts (i.e., NaCl, KCl, MgCl₂, CaCl₂, NaNO₃, and Na₂SO₄) enhances the close encounter of NPLs and allowed further aggregation of NPLs.

Fig. 5 (a) SERS intensity of various sizes of PSNSs spiked with different interferences. Plots of the SERS intensity at 1002 cm⁻¹vs. the concentration of 800nm PSNSs in (b) bottled drinking water and (c) tap water.

Likewise, glycine is an amino acid that can interact with AgNPs through its amino and carboxyl groups.^72,73 It was speculated that glycine competes with citrate for partial adsorption on AgNPs at this concentration. On the other hand, the SERS intensity with reducing sugars (i.e., sucrose, glucose, and fructose) as the interference showed almost no change. It has been known that sugars⁷⁴ act as a steric stabilizer and reducing agent by adsorbing (the hydroxyl groups) onto the surface of the AgNPs and creating a negatively charged layer, which repels other negatively charged AgNPs and prevents their aggregation.

In agreement with previous studies,^75–77 the non-ionic surfactant (Triton X-100) and cationic surfactant (dodecyl trimethyl ammonium bromide, DTAB) showed the highest ability to stabilize and prevent the aggregation of AgNPs, compared to an anionic surfactant (sodium lauryl sulfate, SLS). These long-chain surfactant polymers are generally coated on the surface of AgNPs against aggregation and, thus, affect the contact of the analyte with the AgNPs, resulting in a reduced or restricted SERS sensitivity. No aggregations were observed for the analysis for both samples with Triton-X and DTAB (Fig. S5†).

SERS detection of PSNSs in spiked real-world samples

As proof of practical application, real-world water samples (i.e., bottled drinking water, tap water, and river water) were chosen to study the SERS activity of 800nm PSNSs in this study. However, there was no observable aggregate formation in the river water. Thus, no significant SERS signals were obtained for all PSNS concentrations. This could be attributed to river water contaminants such as dissolved organic matter and common cations, which act as stabilizers and prevent them from forming the aggregation.⁵⁶ As shown in Fig. 5b and c, the plot of SERS intensity as a function of PSNS concentration in bottled drinking water and tap water is nearly the same. They both exhibited good linearity in the range of 2 to 10 μg mL⁻¹ (R² > 0.98), which is insignificantly different from the data obtained from the DI water of 800nm PSNSs, as discussed in the previous experiment. Moreover, the LOD for the bottled drinking water (1.47 μg mL⁻¹), and tap water (1.55 μg mL⁻¹) are slightly different from that of the DI water (1.42 μg mL⁻¹), indicating that this method is not sensitive enough to the interference that might be presented in both of these real-world media, which makes this method available to detect NPLs in every size in a suitable medium.

Conclusions

In conclusion, this work has established a novel and green approach for detecting PSNSs in water using SERS with untreated filter paper. This method offers several advantages over existing techniques. First, it eliminates the need for complex and potentially hazardous pre-treatments, aligning with green chemistry principles. Second, it achieves a LOD of 0.31 μg mL⁻¹ for PSNSs of various sizes, demonstrating high sensitivity. Third, the use of untreated filter paper and a simple syringe filtration setup makes the method cost-effective and field deployable. Finally, successful detection of PSNSs in real-world samples of drinking and tap water highlights its applicability for environmental monitoring. Overall, this green SERS method provides a promising approach for sensitive and practical detection of PSNPs in water, potentially aiding in environmental and human health protection efforts.

Author contributions

Boonphop Chaisirikhwun: supervision, investigation, and writing – review and editing. Mary Jane Balani: investigation and writing – original draft. Sanong Ekgasit: writing – review and editing, resources, and funding acquisition. Yunfei Xie: writing – review and editing. Yukihiro Ozaki: writing – review and editing. Prompong Pienpinijtham: conceptualization, methodology, resources, supervision, funding acquisition, project administration, and writing – review and editing.

Data availability

Supporting information is available in the ESI.†

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This research project is supported by the Second Century Fund (C2F), Chulalongkorn University, and financially supported by the Thailand Science Research and Innovation Fund Chulalongkorn University and NANOTEC (a member of NSTDA) under the Research Network of NANOTEC (RNN) program. Balani, M. D. acknowledges the scholarship program for ASEAN or Non-ASEAN countries supported by Chulalongkorn University, Thailand.

References

1. K. Zhang, A. H. Hamidian, A. Tubić, Y. Zhang, J. K. H. Fang, C. Wu and P. K. S. Lam, Environ. Pollut., 2021, 274, 116554.
2. L. Rani, G. Kaur, P. Sood, J. Kaushal and A. L. Srivastav, in Role of Green Chemistry in Ecosystem Restoration to Achieve Environmental Sustainability, ed. A. L. Srivastav, A. S. Grewal, M. Tiwari and T. D. Pham, Elsevier, 2024, pp. 191–200.
3. P. J. Anderson, S. Warrack, V. Langen, J. K. Challis, M. L. Hanson and M. D. Rennie, Environ. Pollut., 2017, 225, 223–231.
4. R. C. Thompson, Y. Olsen, R. P. Mitchell, A. Davis, S. J. Rowland, A. W. John, D. McGonigle and A. E. Russell, Science, 2004, 304, 838–838.
5. J. P. da Costa, P. S. Santos, A. C. Duarte and T. Rocha-Santos, Sci. Total Environ., 2016, 566, 15–26.
6. M. Enfrin, L. F. Dumée and J. Lee, Water Res., 2019, 161, 621–638.
7. Y. K. Song, S. H. Hong, M. Jang, G. M. Han, S. W. Jung and W. J. Shim, Environ. Sci. Technol., 2017, 51, 4368–4376.
8. I. Chubarenko, I. Efimova, M. Bagaeva, A. Bagaev and I. Isachenko, Mar. Pollut. Bull., 2020, 150, 110726.
9. A. L. Andrady, Mar. Pollut. Bull., 2011, 62, 1596–1605.
10. S. Allen, D. Allen, V. R. Phoenix, G. Le Roux, P. Durántez Jiménez, A. Simonneau, S. Binet and D. Galop, Nat. Geosci., 2019, 12, 339–344.
11. I. E. Napper, B. F. Davies, H. Clifford, S. Elvin, H. J. Koldewey, P. A. Mayewski, K. R. Miner, M. Potocki, A. C. Elmore and A. P. Gajurel, One Earth, 2020, 3, 621–630.
12. H. Dong, L. Wang, X. Wang, L. Xu, M. Chen, P. Gong and C. Wang, Environ. Sci. Technol., 2021, 55, 12951–12960.
13. I. Peeken, S. Primpke, B. Beyer, J. Gütermann, C. Katlein, T. Krumpen, M. Bergmann, L. Hehemann and G. Gerdts, Nat. Commun., 2018, 9, 1505.
14. M. Bergmann, S. Mützel, S. Primpke, M. B. Tekman, J. Trachsel and G. Gerdts, Sci. Adv., 2019, 5, eaax1157.
15. A. R. Aves, L. E. Revell, S. Gaw, H. Ruffell, A. Schuddeboom, N. E. Wotherspoon, M. LaRue and A. J. McDonald, Cryosphere, 2022, 16, 2127–2145.
16. L. Xie, K. Gong, Y. Liu and L. Zhang, Environ. Sci. Technol., 2023, 57, 25–43.
17. B. Toussaint, B. Raffael, A. Angers-Loustau, D. Gilliland, V. Kestens, M. Petrillo, I. M. Rio-Echevarria and G. Van den Eede, Food Addit. Contam.: Part A, 2019, 36, 639–673.
18. M. Carbery, W. O'Connor and T. Palanisami, Environ. Int., 2018, 115, 400–409.
19. L. Peng, D. Fu, H. Qi, C. Q. Lan, H. Yu and C. Ge, Sci. Total Environ., 2020, 698, 134254.
20. F. Ribeiro, J. W. O'Brien, T. Galloway and K. V. Thomas, TrAC, Trends Anal. Chem., 2019, 111, 139–147.
21. H. A. Leslie, M. J. Van Velzen, S. H. Brandsma, A. D. Vethaak, J. J. Garcia-Vallejo and M. H. Lamoree, Environ. Int., 2022, 163, 107199.
22. S. Sarkar, H. Diab and J. Thompson, Int. J. Environ. Res. Public Health, 2023, 20, 1–28.
23. S. Abbasi, N. Jaafarzadeh, A. Zahedi, M. Ravanbakhsh, S. Abbaszadeh and A. Turner, J. Environ. Sci., 2023, 126, 95–102.
24. T. Yang, J. Luo and B. Nowack, Environ. Sci. Technol., 2021, 55, 15873–15881.
25. W. Fu, J. Min, W. Jiang, Y. Li and W. Zhang, Sci. Total Environ., 2020, 721, 137561.
26. E. Gregoris, G. Gallo, B. Rosso, R. Piazza, F. Corami and A. Gambaro, J. Anal. Appl. Pyrolysis, 2023, 170, 105903.
27. H. Mizuguchi, H. Takeda, K. Kinoshita, M. Takeuchi, T. Takayanagi, N. Teramae, W. Pipkin, K. Matsui, A. Watanabe and C. Watanabe, J. Anal. Appl. Pyrolysis, 2023, 171, 105946.
28. S. Adhikari, V. Kelkar, R. Kumar and R. U. Halden, Polym. Int., 2022, 71, 543–551.
29. F. Corami, B. Rosso, B. Bravo, A. Gambaro and C. Barbante, Chemosphere, 2020, 238, 124564.
30. C. Lefebvre, C. Saraux, O. Heitz, A. Nowaczyk and D. Bonnet, Mar. Pollut. Bull., 2019, 142, 510–519.
31. Y. K. Song, S. H. Hong, S. Eo and W. J. Shim, Mar. Pollut. Bull., 2021, 173, 113101.
32. N. Jin, Y. Song, R. Ma, J. Li, G. Li and D. Zhang, Anal. Chim. Acta, 2022, 1197, 339519.
33. J. Weisser, T. Pohl, M. Heinzinger, N. P. Ivleva, T. Hofmann and K. Glas, TrAC, Trends Anal. Chem., 2022, 148, 116535.
34. P. Ribeiro-Claro, M. M. Nolasco and C. Araújo, in Comprehensive Analytical Chemistry, ed. T. A. P. Rocha-Santos and A. C. Duarte, Elsevier, 2017, vol. 75, pp. 119–151.
35. M. Sunil, N. Mithun, M. Charles, S. Chidangil, S. Kumar and J. Lukose, J. Environ. Manage., 2024, 354, 120351.
36. M. Sunil, N. Pallikkavaliyaveetil, N. Mithun, A. Gopinath, S. Chidangil, S. Kumar and J. Lukose, J. Water Process Eng., 2024, 60, 105150.
37. N. P. Ivleva, Chem. Rev., 2021, 121, 11886–11936.
38. M. Simon, N. van Alst and J. Vollertsen, Water Res., 2018, 142, 1–9.
39. L. Cabernard, L. Roscher, C. Lorenz, G. Gerdts and S. Primpke, Environ. Sci. Technol., 2018, 52, 13279–13288.
40. S. M. Mintenig, I. Int-Veen, M. G. Löder, S. Primpke and G. Gerdts, Water Res., 2017, 108, 365–372.
41. P. M. Anger, E. von der Esch, T. Baumann, M. Elsner, R. Niessner and N. P. Ivleva, TrAC, Trends Anal. Chem., 2018, 109, 214–226.
42. N. P. Ivleva, A. C. Wiesheu and R. Niessner, Angew. Chem., Int. Ed., 2017, 56, 1720–1739.
43. M. M.-L. Leung, Y.-W. Ho, C.-H. Lee, Y. Wang, M. Hu, K. W. H. Kwok, S.-L. Chua and J. K.-H. Fang, Environ. Pollut., 2021, 289, 117648.
44. K. Dodo, K. Fujita and M. Sodeoka, J. Am. Chem. Soc., 2022, 144, 19651–19667.
45. Z. Sobhani, M. Al Amin, R. Naidu, M. Megharaj and C. Fang, Anal. Chim. Acta, 2019, 1077, 191–199.
46. B. H. Dong and J. P. Hinestroza, ACS Appl. Mater. Interfaces, 2009, 1, 797–803.
47. G. P. Glaspell, C. Zuo and P. W. Jagodzinski, J. Cluster Sci., 2005, 16, 39–51.
48. Y. Jeon, D. Kim, G. Kwon, K. Lee, C.-S. Oh, U.-J. Kim and J. You, Carbohydr. Polym., 2021, 272, 118470.
49. S. Kihara, A. Chan, E. In, N. Taleb, C. Tollemache, S. Yick and D. J. McGillivray, RSC Adv., 2022, 12, 20519–20522.
50. Q. T. Lê, N. H. Ly, M.-K. Kim, S. H. Lim, S. J. Son, K.-D. Zoh and S.-W. Joo, J. Hazard. Mater., 2021, 402, 123499.
51. J. Li, H. Liu, S. Chen, X. Liang, Y. Gao, X. Zhao, Z. Li, C. Zhang, F. Lei and J. Yu, J. Phys. Chem. Lett., 2022, 13, 5815–5823.
52. L. Lv, L. He, S. Jiang, J. Chen, C. Zhou, J. Qu, Y. Lu, P. Hong, S. Sun and C. Li, Sci. Total Environ., 2020, 728, 138449.
53. L. Mikac, I. Rigó, L. Himics, A. Tolić, M. Ivanda and M. Veres, Appl. Surf. Sci., 2023, 608, 155239.
54. X.-L. Nie, H.-L. Liu, Z.-Q. Pan, S. A. Ahmed, Q. Shen, J.-M. Yang, J.-B. Pan, J. Pang, C.-Y. Li, X.-H. Xia and K. Wang, Chem. Commun., 2019, 55, 6397–6400.
55. G. Xu, H. Cheng, R. Jones, Y. Feng, K. Gong, K. Li, X. Fang, M. A. Tahir, V. K. Valev and L. Zhang, Environ. Sci. Technol., 2020, 54, 15594–15603.
56. X.-X. Zhou, R. Liu, L.-T. Hao and J.-F. Liu, Talanta, 2021, 221, 121552.
57. T. Itoh, M. Procházka, Z. C. Dong, W. Ji, Y. S. Yamamoto, Y. Zhang and Y. Ozaki, Chem. Rev., 2023, 123, 1552–1634.
58. N. K. Mogha and D. Shin, TrAC, Trends Anal. Chem., 2023, 158, 116885.
59. H. K. Lee, Y. H. Lee, C. S. L. Koh, G. C. Phan-Quang, X. Han, C. L. Lay, H. Y. F. Sim, Y.-C. Kao, Q. An and X. Y. Ling, Chem. Soc. Rev., 2019, 48, 731–756.
60. C. L. Haynes, A. D. McFarland and R. P. Van Duyne, Anal. Chem., 2005, 338.
61. L. Chang, S. Jiang, J. Luo, J. Zhang, X. Liu, C.-Y. Lee and W. Zhang, Environ. Sci.: Nano, 2022, 9, 542–553.
62. B. Chaisrikhwun, S. Ekgasit and P. Pienpinijtham, J. Hazard. Mater., 2023, 442, 130046.
63. T. Sukmanee, K. Wongravee, S. Ekgasit, C. Thammacharoen and P. Pienpinijtham, Anal. Sci., 2017, 33, 89–94.
64. H. Kamiya, K. Gotoh, M. Shimada, T. Uchikoshi, Y. Otani, M. Fuji, S. Matsusaka, T. Matsuyama, J. Tatami and K. Higashitani, in Nanoparticle Technology Handbook, Elsevier, 2008, pp. 113–176.
65. J. Caldwell, P. Taladriz-Blanco, B. Rothen-Rutishauser and A. Petri-Fink, Nanomaterials, 2021, 11, 1149.
66. R. Hu, K. Zhang, W. Wang, L. Wei and Y. Lai, J. Hazard. Mater., 2022, 429, 128388.
67. W.-L.-J. Hasi, X. Lin, X.-T. Lou, S. Lin, F. Yang, D.-Y. Lin and Z.-W. Lu, Appl. Phys. A, 2015, 118, 799–807.
68. J. Qiu, X. Li and X. Qi, IEEE Photonics J., 2019, 11, 1–12.
69. A. Fargašová, R. Prucek, V. Ranc, A. Panáček, L. Kvítek and R. Zbořil, RSC Adv., 2015, 5, 9737–9744.
70. Y. Jiang, X. Wang, G. Zhao, Y. Shi, Y. Wu, H. Yang and F. Zhao, Water Res., 2024, 255, 121444.
71. J. Y. Kim, E. H. Koh, J. Y. Yang, C. Mun, S. Lee, H. Lee, J. Kim, S. G. Park, M. Kang and D. H. Kim, Adv. Funct. Mater., 2024, 34, 2307584.
72. N. Agasti, V. K. Singh and N. K. Kaushik, Mater. Res. Bull., 2015, 64, 17–21.
73. M. J. Vesga, D. McKechnie, S. Laing, H. Kearns, K. Faulds, K. Johnston and J. Sefcik, Colloids Surf., A, 2021, 621, 126523.
74. E. Filippo, A. Serra, A. Buccolieri and D. Manno, J. Non-Cryst. Solids, 2010, 356, 344–350.
75. F. Begum, S. A. Jahan, M. Y. A. Mollah, M. M. Rahman and M. A. B. H. Susan, J. Sci. Res., 2017, 9, 431–447.
76. S. Skoglund, T. A. Lowe, J. Hedberg, E. Blomberg, I. O. Wallinder, S. Wold and M. Lundin, Langmuir, 2013, 29, 8882–8891.
77. L. Kvítek, A. Panáček, J. Soukupová, M. Kolář, R. Večeřová, R. Prucek, M. Holecová and R. Zbořil, J. Phys. Chem. C, 2008, 112, 5825–5834.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4an00702f

---