Development of a Dye-Binding Method for Nanoplastics Detection in Snow Samples Using Capillary Electrophoresis with Laser-Induced Fluorescence Detection

Abstract

Advanced analysis of nanoplastics in snow presents unique challenges due to the intricacies involved in extracting and detecting small particles within a snow matrix. In developing a new method, samples were collected in aluminum cans during and after snowfall. The melting process was controlled within a closed metal chamber to prevent contamination. The melted snow samples were treated with a consortium of fluorescent organic dyes that bind to the target nanoplastics. Centrifugation effectively sedimented most dye-bound nanoplastics from the heterogeneous mixture. Subsequent electrokinetic injection of free dyes from the supernatant facilitated their efficient separation by capillary electrophoresis (CE), which enabled precise quantification through laser-induced fluorescence (LIF) detection. This novel dyes-CE-LIF method favorably compares with established instrumental methods for quantitatively analyzing melted snow samples containing nanoplastics amid microplastics and other colloidal particles. This novel application of the method enables rapid detection, within 15 min, at a lower limit as small as 0.1 μg of nanoplastics in melted snow samples. It requires minimal sample handling and enhances analytical throughput. Principal component analysis (PCA) revealed grouping trends among the melted snow samples and other water samples based on dye binding percentages. Variability due to differences in nanoplastic types, surface chemistry, particle size distribution, dye-nanoplastic interactions, and impurity levels were investigated across different sample sources.

Introduction

Air pollution has long been a focal point of atmospheric science research. Recently, the presence of microplastics, particularly fragments, in the atmosphere has been documented. The open combustion of plastics is proposed as a significant source of atmospheric nanoplastics and plays a pivotal role in the global plastic cycle. Nanoplastics, due to their ability to be transported over long distances, are globally widespread and may be more concentrated than microplastics. Significant research has identified nanoplastics as an emerging class of pollutants. Various types of nanoplastics, such as polyethylene (PE), poly­(ethylene terephthalate) (PET), polypropylene (PP), polystyrene (PS) and poly­(vinyl chloride) (PVC), have been commonly detected in the air, primarily due to their extensive commercial use. , Compared to aquatic environments, atmospheric nanoplastics are less susceptible to matrix interferences in field samples. These nanoparticles serve as carriers and vectors for disseminating organic environmental pollutants globally, traveling significant distances in the atmosphere. Airborne nanoplastics pose health risks to animals and humans through inhalation, reaching alveolar surfaces, depositing in the lungs, crossing the pulmonary epithelial barrier, and entering the bloodstream. − This pathway can lead to persistent inflammation, pulmonary fibrosis, lung carcinogenesis, major adverse cardiac events, or even neurological disorders. A growing consensus acknowledges that nanoplastics adversely affect the environment and the food chain, posing significant threats to human health and food safety. , Currently, the understanding of nanoplastics occurrence, distribution, abundance and fate in the environment is limited, primarily due to insufficient techniques for detecting and identifying airborne nanoplastics in the atmosphere.

Generally, collecting atmospheric plastic nanoparticles is challenging due to their small sizes and altitude constraints. These atmospheric nanoplastics can be deposited by rains on plants and pollinators. In remote regions like the Swiss Alps and Mount Everest, nanoplastics have been detected within snow layers, attributed primarily to atmospheric deposition. Estimates suggest around 43 trillion nanoplastic particles land in Switzerland annually. This finding suggests widespread environmental dissemination via atmospheric transport phenomena. Studies have identified nanoplastics in snow across various geographical locations, indicating their global prevalence. Nanoplastics are known to settle from the atmosphere through snow precipitation processes, impacting soil and water quality by depositing in various ecosystems. Millions of metric tons of nanoplastics are transported annually within the marine atmosphere and deposited into the oceans. , These nanoplastics can re-enter the atmosphere through the aerosolization of sea spray. Exploring innovative methodologies for the collection and analysis of nanoplastics is essential. Understanding their distribution aids in modeling pollution pathways and predicting environmental impacts. Hazardous nanoplastics can subsequently be inhaled by humans and wildlife, posing potential health risks. Research extensively assesses their airborne deposition rates, geo-archives, environmental impact, bioaccumulation, , human biomonitoring, , and links to cardiovascular diseases across the globe. By integrating exposure data, risk assessments, and analytical results, recent findings underscore the necessity for global policy action. These findings emphasize the importance of conducting region-specific research to advance air quality and public health initiatives.

Environmental studies emphasize improving analytical techniques to monitor atmospheric pollution by nanoplastics. Developing reliable methods to detect and measure airborne nanoplastics is crucial to providing a scientific foundation for regulatory frameworks. Understanding the presence and effects of atmospheric nanoplastics can inform policies for plastic use and waste management. However, sampling atmospheric nanoplastics presents significant challenges. Diverse sampling approaches for airborne nanoplastics each have unique advantages and limitations. A recent study has provided the inaugural evidence of nanoplastics contamination in the lungs of birds, underscoring the potential of larger-bodied birds to serve as bioindicators. Atmospheric deposition, such as rainfall and snowfall, can be readily collected for subsequent nanoplastics analysis. Recent studies indicate that microplastic concentrations are higher in snowfall than in rainfall. Additionally, poly­(vinyl chloride) (PVC) and polyacrylamide (PAM) are predominant in rainfall, whereas polyethylene (PE), polypropylene (PP) and PVC are more prevalent in snowfall.

Nanoplastics become increasingly difficult to quantify using standard microplastics detection methods. One exceptional method is nanoparticle tracking analysis, recently employed to assess the effects of human exposure to nanoplastics. Pretreatment methodologies for nanoplastics have enabled the practical application of surface-enhanced Raman scattering (SERS) analysis. Thermal desorption–proton transfer reaction–mass spectrometry has been used for the chemical characterization of nanoplastics, with a detection limit of less than 1 ng for polystyrene (PS) in Alpine snow. Tire wear, polystyrene, and polyethylene particles are dominant polymer types found at high-altitude glaciers. Pyrolysis–gas chromatography–mass spectrometry facilitates both qualitative and quantitative analyses of nanoplastics in PM_2.5 but is currently unsuitable for polyethylene and poly­(vinyl chloride) due to interferences from nonspecific pyrolysis products. All these traditional methods have been challenging for real-time or in situ detection of nanoplastics in aqueous and atmospheric environments. , An emerging sensor technology leverages photoelectrochemical-electrochemical dual-mode mechanisms, with capabilities to detect nanoplastics in concentrations ranging from 0.38 ng/mL to 800 μg/mL It highlights innovative detection methodologies could be incorporated or expanded with conductivity assessments. A novel methodology has been reported using cryogenic milling, UV weathering, and incineration on virgin plastic materials to synthesize environmentally relevant reference nanoplastics. This approach aims to represent the entire life cycle of these materials for comprehensive risk assessment studies. A versatile 3D-printed hollow-laser desorption/ionization mass spectrometry (HoLDI-MS) platform has recently been developed for the efficient detection of nanoplastics in environmental matrices. The concentration of poly­(ethylene glycol) (PEG) in urban snow was measured at 0.06 mg/L.

The adhesion of organic dyes to hydrophobic nanoplastic surfaces in aqueous solutions is well-known among chemists and chemical engineers. Recent reviews indicate that adsorption on plastic surfaces is highly dependent on the properties of the organic dyes. Other scientific studies have identified the potential of using dye adsorption for plastic identification in environmental samples. Congo red and methylene blue are two widely utilized dyes, particularly relevant in water and wastewater chemistry research. Current research suggests the principles of fluorescence, localized surface plasmon resonance, and speckle pattern analysis are effective in detecting and quantifying nanoplastics. Many organic dyes are fluorescent due to their chemically heterogeneous aromatic structures. Some contain hydrophilic moieties stabilized by electrostatic and electrosteric mechanisms. Their photo- and thermal stability enables photon emission when excited by laser light. Four fluorescent organic dyescyanine-3 phosphoramidite, rhodamine 6G, fluorescein sodium salt, and Vat Red 15adsorbed onto nanoplastics, resulting in peak shifts in the fluorescence spectrum. These shifts provide sufficient specificity for nanoplastic identification. Even at low concentrations (<10 ppm), these dyes can be readily detected by laser-induced fluorescence (LIF) following capillary electrophoresis (CE), allowing for high-sensitivity quantitative chemical analysis. This study aims to explore the potential of using dyes, CE, and LIF to detect different nanoplastics in snow samples collected postdeposition. This work represents the first attempt with a CE-LIF system to assess the exposure identity of airborne nanoplastics. Our objective is to enhance the understanding of atmospheric pollution by nanoplastics in three cities with diverse populations, lifestyles, and industries.

Materials and Methods

Organic fluorescent dyes, including rhodamine 6G (MW: 479.01 g/mol), coumarin 334 (MW: 283.32 g/mol) and fluorescein sodium salt (MW: 376.27 g/mol) were purchased from Millipore Sigma (Oakville, Ontario, Canada) in the analytical standard grade. Sodium phosphate dibasic (Na₂HPO₄) was obtained from BioShop (Burlington, Ontario, CA).

To analyze nanoplastics, snow samples were collected from areas with minimal contamination using tin containers to prevent the introduction of micro- or nanoplastic particles. In Ottawa, with a population of approximately 1.07 million, Sample 1 was obtained from Southbank Road by placing a tin can on a solid support to directly collect freshly falling snow, minimizing contamination from soil. Similarly, Sample 2 was collected at Pimlico Crescent in Ottawa South. Additional samples were gathered from Kemptville (population around 4500), North Grenville (population around 20,000), and Kingston (population around 175,000). A clean stainless-steel scoop was used to avoid introducing external contaminants. To avoid air contamination during transportation to our research lab, tin container caps were kept airtight. All snow samples were stored at 4 °C to allow for controlled melting over time. This temperature did not induce microbial activity or cause chemical degradation of the nanoplastics during the initial 4 weeks of CE-LIF analysis. Prior to analysis, the snow in each container was allowed to melt gradually on the lab bench at room temperature (24 °C). A repeat analysis of the melted snow samples after extending the storage up to 8 weeks yielded similar results.

A 20 mM background electrolyte (BGE) solution was prepared by dissolving 0.2839 g of Na₂HPO₄ in 100 mL of deionized water within a Pyrex beaker. Subsequently, 0.0250 g each of rhodamine 6G, coumarin 334, and fluorescein disodium were added to the BGE solution. The mixture was dissolved with the aid of a magnetic stirring bar placed on a magnetic stirrer to achieve a concentration of 250 μg/mL for each dye. A 1.0 mL aliquot of the resulting dye consortium solution was combined with 1.0 mL of deionized water in a glass vial to create a blank. This blank was analyzed via capillary electrophoresis (CE) with laser-induced fluorescence (LIF) detection to obtain reference peak heights for all three dyes. The capillary measured 34 cm from the inlet to the point of LIF detection, 62 cm to the point of UV detection and 68 cm to the outlet. The LIF detection system comprised a 450 nm diode laser for excitation, an interference filter with maximum optical transmission at 520 ± 5 nm, and a Hamamatsu H7827 series photosensor module (Iwata City, Japan).

The blank analysis was performed weekly to assess its chemical stability over 3 weeks. In a similar manner, a 1.0 mL aliquot of the 3-dye consortium solution was mixed with 1.0 mL of melted snow to create an unknown sample. Before CE-LIF analysis, all 2.0 mL blank/snow samples were transferred into centrifuge tubes and subjected to centrifugation at 10,000g-force for 5 min to effectively sediment most dye-bound nanoplastics from the heterogeneous mixture. Binding percentages were calculated as 100% × (1 – final peak height/initial peak height). For quality control and assurance, all blank and unknown samples were analyzed in triplicate to ensure the repeatability of % binding results.

Conductivity was measured (in triplicate or more times) using 0.1 mL of melted snow or water sample with a Horiba Twin Cond model B-173 m. The Hawaiian volcanic water was purchased from a local supermarket in Ottawa (Ontario, Canada). Dopamine hydrochloride (5.0 mg) was added to all samples, including the controls and standards. Sonication in a water bath by 35 W of 42 kHz ultrasonic waves for 180 s ensured a homogeneous solution for each, thereby initiating uniform polymerization. Capillary electrophoresis was employed to collect peak-height data over time for the kinetic profile analysis of polydopamine formation.

Results and Discussion

CE-LIF Analysis of Fluorescent Organic Dyes

A consortium of fluorescent organic dyes was meticulously selected based on their aqueous solubility, laser excitability, fluorescence quantum yield, and atmospheric rarity. Rhodamine 6G, coumarin 334 and fluorescein disodium can all be excited by the 450-/480 nm diode laser and their emission wavelengths cover the maximum optical transmission of the interference filter at 520 ± 5 nm. Prior to each CE-LIF analysis, HPLC-grade methanol was used to flush the capillary, effectively removing any old dye residues adhering to the inner capillary wall and preventing cross-contamination of subsequent sample analysis. Given the dye concentration was crucial for obtaining a measurable dye binding percentage, the dye consortium blank was set at 125 μg/mL to produce three strong peaks for rhodamine 6G (R6G), coumarin 334 (C334) and fluorescein (Fluor), as depicted by the electropherogram in Figure , at 2.7 min, 3.4 and 9.0 min, respectively. Electromigration studies had been previously conducted on fluorescent dyes with varying charges, specifically focusing on R6G, C334 and fluorescein as cationic, neutral, and anionic analytes, respectively, with similar sizes. Electromigration of two other positively charged dyes (rhodamine B and rhodamine 123) were also observed, with a different migration time for each dye.

1.

Two consecutive analyses of dye consortium blank by CE-LIF after flushing capillary with HPLC-grade methanol to remove dye residues adhering to inner capillary wall. Peaks for rhodamine 6G (R6G), coumarin 334 (C334) and fluorescein (Fluor) appear at 2.7, 3.4 and 9.0 min, respectively, under an operating voltage of 18 kV. Capillary length from inlet to outlet: 68 cm; to point of UV detection: 62 cm. The two panels illustrate the reproducibility of CE migration times and LIF signal intensities.

Assessing the stability duration of the three dyes within the consortium at room temperature was essential. Over an 18-day period, we continuously monitored the consortium using capillary electrophoresis with laser-induced fluorescence (CE-LIF). Figure illustrates that the signal peaks for each dye remained predominantly stable, indicating that daily preparation of the consortium is unnecessary for detecting nanoplastics in snow samples. Once the initial dye peak signals were determined via CE-LIF analysis, subsequent peaks for unknown snow samples could be compared to evaluate percentage binding results. A decrease in CE-LIF peak height indicates a reduction in dye concentration due to binding with nanoplastics from the melted snow. Minor daily fluctuations in the initial dye signals were observed due to technical variation in CE-LIF conditions and were found to have a negligible impact on binding percentages, thus preserving the integrity of the dye binding profile for each snow sample.

2.

Continuous analysis of the 3-dyes consortium using CE-LIF over 18 days of storage at room temperature.

The linear dynamic range for a working dye consortium concentration of 125 μg/mL extends from 0% to 100%. When analyzing snow samples with low nanoplastic concentrations, it becomes crucial to reduce the dye concentration to obtain measurable dye binding percentages that adequately distribute across the dynamic range. Achieving this requires enhanced detection sensitivity, which can be facilitated by employing laser-induced fluorescence (LIF) or, alternatively, mass spectrometry. As future guidance, improving detection accuracy necessitates minimizing baseline noise, achievable by optimizing the design of a light-proof LIF detection chamber or performing CE-LIF operations in a darkened environment.

Evaluation of Dye Binding to Nanoplastics in Melted Snow

Research indicates that nanoplastics, much like microplastics, can be washed from the atmosphere through precipitation processes, such as snowfall. This phenomenon is especially significant in remote regions, where atmospheric deposition serves as a primary transportation pathway. The identification of nanoplastics in snow from various locations suggests widespread geographical deposition. The presence of nanoplastics in snow raises potential environmental concerns, particularly regarding contamination in pristine or remote ecosystems, and further informs our understanding of atmospheric plastic pollution cycles. As snow melts, these nanoplastic particles enter surface water systems, potentially impacting aquatic life and ecosystems. Understanding typical transport mechanisms and sinks is critical in determining overall environmental impacts. This highlights the necessity for enhanced research into detection methods and the environmental consequences of nanoplastic deposition through atmospheric processes.

Toward the validation of a robust dyes-binding method for the detection of nanoplastics in snow, several snow samples were collected in Ottawa, Kemptville, North Grenville, and Kingston for analysis. Any decreases in fluorescence signal intensity would infer the well-known ionic, hydrophobic and π-π binding interactions. The results presented in Table reveal variations in the binding of rhodamine 6G, coumarin 334, and fluorescein onto nanoplastics (and other adsorbents) in those different snow samples after thawing. Notably, Ottawa Southbank snow 1 exhibited a significantly higher rhodamine 6G binding percentage (22%) than those (2%–10%) determined for the other snow samples. Some negative binding percentages were observed with coumarin 334 and fluorescein. This may suggest a lower concentration of nanoplastics or adsorbents in these snow samples compared to the deionized water blank. These negative values could indicate a systematic error associated with the deionized water blank. These binding tests will be repeated to double check the negative percentage results when fresh snow samples can be collected this coming winter to establish a temporal trend from year to year.

1. Comparison of Rhodamine 6G, Coumarin 334 and Fluorescein Dye Bindings at pH 7.1 with Melted Snow Samples from Three Urban Cities and One Suburban Municipality .

	binding
	mean ± standard deviation (n = 3)
snow source	R6G	C334	fluorescein
Ottawa Southbank 1	22% ± 4%	9% ± 2%	9% ± 2%
Ottawa Southbank 2	3% ± 1%	–4% ± 1%	12% ± 3%
Kemptville	2% ± 1%	–1% ± 1%	–6% ± 2%
North Grenville	4% ± 1%	–3% ± 1%	15% ± 3%
Kingston	10% ± 2%	–9% ± 2%	–14% ± 3%

^aThese represent the mean results and standard deviations derived from triplicate CE-LIF analyses.

Understanding and applying robust acceptance criteria are critical for ensuring reliable data in nanoplastics research, fostering confidence in regulatory compliance and decision-making. The acceptance criteria for blanks in nanoplastics analysis typically include method blanks, reagent blanks, and field blanks. Method blanks assess contamination across all parts of the analytical process. Nanoplastics levels should be less than 5% of the sample concentration. Reagent blanks monitor contamination from reagents used in the analysis. They should not contain detectable levels of nanoplastics. Field blanks evaluate contamination during sample collection and handling. Levels should be below 5–10% of the sample concentrations. In this study, despite minimal sample preparation, binding percentages were calculated as 100% × (1 – final peak height/initial peak height). This calculation inherently accounts for contamination of each dye by nanoplastics, albeit unlikely, in the initial peak height. The difference between the initial and final peak heights accounts for contamination during the analytical process. Contamination during sample collection and handling was evaluated through post-mortem analysis of the dye consortium, conducted using a spare aluminum can that was not used to collect any snow sample.

Quality control (QC) involves using calibration standards, spike recovery, replicate samples, external QC standards, and internal standards. Calibration standards ensure the instrument’s response across the expected range of nanoplastics concentrations. A correlation coefficient (R ²) of 0.99 or better is typically required. Spike recovery evaluates method accuracy and matrix effects, with the recovery of spiked analytes generally between 70–130% of the expected value. Replicate samples should demonstrate precision within 20% relative standard deviation (RSD). External QC standards validate method and instrument performance, requiring agreement within ±10% of certified values. Internal standards adjust for variations in sample analysis, with recovery typically between 70%–130%. In our previous work, calibration standards attained a positive correlation (R ² = 0.9959) between the % binding of R6G and the mass of polystyrene nanoplastics ranging from 0.1 to 0.4 μg in 1.6 mL samples via CE-LIF analysis. Replicate samples demonstrated precision ranging from 7% RSD to 15% RSD. Further work will involve method development and validation in complex snow matrices, interlaboratory comparisons, and using advanced detection technologies like mass spectrometry for improved nanoplastics analysis.

Principal Component Analysis

Principal component analysis (PCA) identified three principal components, limited by the input variables: binding percentages of rhodamine 6G, coumarin 521, and fluorescein. A scree plot was generated to depict the variance explained by each principal component, with eigenvalues plotted in descending order. This plot helps determine the number of meaningful components in the data structure. The elbow point, where the curve begins to level off, indicates how many components to retain, with those left of the elbow explaining the most variance. As shown in Supporting Information Figure 1, the scree plot reveals that PC1 and PC2 account for the majority of the variance before the curve flattens. PC1 surpasses the broken-stick model expectation with an eigenvalue of 1.83, and PC2 exceeds the 0.83 expectation. Both the Kaiser rule (retaining PCs with eigenvalues >1) and Jolliffe’s modification (retaining if eigenvalues >0.7) are satisfied as secondary heuristics. Least, PC3 adds minimal residual variance.

The principal component loadings in Table indicate how significantly each melted snow or water sample contributes to the two primary variation patterns in the data set, specifically characterized by the binding interactions between fluorescent organic dyes and nanoplastics. Loadings are unitless when they are computed as correlations, which is typical for PCA on standardized variables/correlation matrix. |Loading| ≈ 0.40 is treated as a starting point for substantial deviations. A high positive loading in Principal Component 1 (PC1), as seen with Carleton University snow 1 (3.40555) and 2 (2.87585) samples, suggests strong dye consortium interactions in these samples. Conversely, negative PC1 loadings observed in several lab water, residential tap water, and snow samples imply distinct compositions and concentrations of nanoplastics compared to the samples with positive PC1 values. These distinctions may arise from water source differences, nanoplastic types, or nanoplastic size distributions. The PC2 loadings display both positive and negative values, indicating a secondary trend in the data. Notably, North Grenville snow (2.288), Ottawa Southbank snows 1 and 2, and residence fountain water samples from the Glengarry Building exhibit relatively high positive PC2 values, suggesting a strong association with this secondary variance pattern.

2. Principal Component Loadings for Water and Snow Analysis.

sample source and kind	PC1 loading	PC2 loading
Reverse osmosis (sugar shack) water	0.36552	–0.20973
Residence fountain 1 water	0.73818	–0.89707
Residence fountain 2 water	0.03473	–0.51507
Residence fountain 3 water	–0.28767	0.66159
Residence fountain 4 water	0.44535	–0.21089
Residence fountain 5 water	–0.13766	0.98077
Residence fountain 6 water	–0.14875	1.64045
Residence fountain 7 water	0.14791	0.1404
River water	0.10649	0.1418
Lab tap 1 water	0.33901	–1.68188
Residence tap 1 water	–0.42321	–0.13986
Lab freezer frost	–0.37877	–0.14291
Deionized water	–0.90063	–0.58697
Background electrolyte solution (10 mM Na2HPO4)	–0.9453	–0.08828
Lab tap 2 water	–0.05236	–0.66809
Lab tap 3 water	–0.08918	–0.3318
Residence tap 2 water	–1.12811	0.58031
Residence tap 3 water	–0.86123	0.06638
Residence tap 4 water	–0.78019	0.22269
Lab tap 4 water	0.02725	–0.17362
Lab tap 5 water	0.00273	–0.33954
Lab tap 6 water	–0.03689	–0.49725
Reverse osmosis (Vermaire lab) water	0.55739	–0.88324
Reverse osmosis (Ormeci lab) water	0.48217	–0.54719
Carleton University snow 1	3.40555	0.85951
Carleton University snow 2	2.87585	0.57784
Ottawa Southbank snow 1	0.37392	1.27228
Ottawa Southbank snow 2	–1.14725	1.89804
Kemptville snow	–0.81153	–1.08871
North Grenville snow	–1.09699	2.38736
Kingston snow	–0.67632	–2.42731

To further analyze the binding test results for melted snow, PCA was utilized to create a biplot, depicted in Figure , that assessed the interactions with the dye consortium. Compared to water samples, most snow samples exhibit principal component 1 (PC1) values ranging from 1.000 to −1.000. PC1 accounts for 61.0% of the variance and displays a strong correlation with rhodamine 6G and coumarin 334. This indicates a relatively low presence of nanoplastics in the snow samples capable of interacting with these dyes. Conversely, their principal component 2 (PC2) values vary broadly from 2.500 to −2.500, explaining 34.0% of the variance and showing a stronger correlation with fluorescein. The snow samples, compared to water samples, demonstrate a wider distribution along PC2. A negative PC1 or PC2 value implies that certain nanoplastic propertiessuch as particle size, surface chemistry, or associated contaminantscontribute inversely to the principal component. Consequently, snow samples with lower PC2 values likely contain nanoplastics with differing chemical compositions and dye-binding affinities compared to those with higher PC2 values, revealing a more diverse interaction profile between the snow and the organic dyes.

3.

Principal component analysis of melted snow samples after binding with rhodamine 6G, coumarin 334 and fluorescein dyes. Biplot axes PC1 and PC2 account for 95.0% of the variance.

Two snow samples from Carleton University exhibit significantly elevated PC1 values, distinguishing them from other collected snow samples. This anomaly likely results from contamination with humic and fulvic acids originating from the ground soil. These humic substances interact with organic dye molecules primarily through the electrostatic attraction/repulsion, hydrogen bonding, and hydrophobic interactions. They contain negatively charged carboxyl (−COOH) and phenolic (−OH) functional groups that can bind with positively charged dyes, such as rhodamine 6G. Additionally, their hydrophobic aromatic backbones can associate with the nonpolar regions of dye molecules. Consequently, the two outlying samples were excluded from our study, underscoring the importance of sampling fallen snow only after it has accumulated at least a couple of inches off the ground. To further investigate this hypothesis, a binding test was conducted with 1.2 mg of humic acid using the dye consortium, resulting in unexpectedly high binding percentages of 98.5% R6G, 34.9% C334, and 39.1% fluorescein. Such interference could be validated by orthogonal techniques such as surface enhanced Raman spectroscopy (SERS). The normalized binding ratio was determined to be 1.000:0.354:0.397 (R6G: C334: fluorescein). This ratio contrasts with those obtained for commercially available polystyrene nanospheres, lab-synthesized polystyrene nanoparticles, and lab-synthesized PS–PET composite nanoparticles. We suggest that the normalized binding ratio is characteristic of specific types of nanoplastics or humic substances. Furthermore, water samples, including tap, fountain, river, and reverse osmosis water, devoid of soil or humic substances, served collectively as negative controls. Their PCA data points cluster more tightly near the origin in the biplot, showing relatively minor variations along the two principal component axes. Clearly, these water samples exhibit similar characteristics concerning their dye-binding interactions. A field blank sample may be used to assess contamination during storage and transportation. Isolation of nanoplastics from other particulates and membrane filtration are optional.

To elucidate the contribution of each dye (rhodamine 6G, coumarin 334, and fluorescein) to the principal components, regression analysis was performed using the principal component scores (PC1 and PC2) as dependent variables. The aim of linear regression is to determine how a variable Y (representing the percentage of dye binding) changes in relation to another variable X (representing the PC1 scores). This analysis allows for the examination of the association between each dye’s binding behavior and the overall trend-captured by the PC1 axis, which accounts for the largest variance across all samples. For rhodamine 6G, the regression equation is Y = 15.52 + 10.65X with an R ² value of 0.911, as depicted in Figure a. This indicates a strong positive linear relationship between the PC1 score and the R6G binding percentage. The steep slope suggests a 10.65% increase in R6G binding for every unit increase in PC1 score, while the high R ² value implies that 91.1% of the variability in R6G binding is explained by PC1 scores. The y-intercept of 15.52 represents the estimated binding percentage when the PC1 score is zero, serving as a baseline level of binding expected in samples strongly aligning with PC1. This relatively high intercept indicates R6G’s moderate inherent binding capacity across all samples, accentuating its dominant contribution to PC1 among the three dyes.

4.

Linear regression of (a) rhodamine 6G binding percentage, (b) coumarin 334 binding percentage, (c) fluorescein binding percentage, against PC1 scores.

Similarly, for coumarin 334, the regression equation is Y = 5.03 + 11.29X with an R ² value of 0.908, as seen in Figure b. Although the intercept is lower than R6G’s, the slightly higher slope indicates a significant increase in coumarin binding with changes in PC1 scores. The high R ² demonstrates a strong model fit, indicating systematic variation in coumarin binding with PC1 scores. In contrast, fluorescein follows a different pattern, with the regression equation Y = 1.06 + 0.51X and an R ² value of 0.007, depicted in Figure c. This suggests that only 0.7% of the variation in fluorescein binding is explained by PC1 scores. The y-intercept of 1.06 suggests minimal inherent binding in samples aligned with the central trend of PC1. Overall, these regression results highlight that R6G and coumarin are strongly correlated with PC1, indicating high positive loadings in the PCA model, thus playing a significant role in defining PC1’s direction. Conversely, fluorescein shows almost no correlation, contributing negligibly to PC1.

The same regression analysis was applied to the PC2 scores, revealing contrasting behaviors among the three dyes. While PC1 accounted for 61% of the variability associated with nanoplastics and potential contaminants present in water and snow, PC2, which explains 34% of the total variance, was explored to determine if it represents more subtle, dye-specific interactions or underlying environmental influences, such as trace-level dissolved organic matter (e.g., humic or fulvic acids), that might have been overlooked by the first principal component. For rhodamine 6G, the regression equation was Y = 15.52 + 0.80X with an R ² value of 0.005, as shown in Figure a. Similarly, coumarin 334 yielded a regression equation of Y = 5.03 + 1.04X with an R ² value of 0.008, as seen in Figure b. These extremely low R ² values suggest that the binding of rhodamine 6G and coumarin 334 contributes minimally to the variations in PC2 scores. This implies that the variance captured by PC2 is largely unrelated to the mechanisms governing these dyes’ interactions with nanoplastics and contaminants within the water/snow sample matrix. Interestingly, fluorescein displayed a markedly different trend. The PCA analysis for fluorescein binding shows a strong relationship between the PC2 scores and the binding percentage. The regression line for fluorescein was Y = 1.06 + 6.1X with an R ² value of 1.000, as illustrated in Figure c, indicating a perfect linear relationship between its binding percentage and the PC2 score. This means that increases in PC2 scores correlate consistently with increases in fluorescein binding percentages across analyzed samples. The high R ² value suggests that PC2 effectively captures the variance in fluorescein binding among the samples. Samples with more positive PC2 scores correspond to higher fluorescein binding, while more negative PC2 scores correspond to lower or no binding percentages. This difference supports the hypothesis that the three dyes interact differently with nanoplastics, consequently affecting the distribution of data points along the PC2 scale. To validate the robustness of the PCA model when differentiating dye-binding patterns, the Kaiser-Meyer-Olkin (KMO) test and Bartlett’s test of sphericity can readily be employed.

5.

Linear regression of (a) rhodamine 6G binding percentage, (b) coumarin 334 binding percentage, and (c) fluorescein binding percentage, against PC2 scores.

Analytical Interference by Humic Acid from Soil

Nanoplastics have the capacity to interact with environmental matrices, altering their fundamental characteristics and behaviors. Humic acid is a natural organic substance found in ground soil and water sources, originating from the decomposition of plant and microbial matter. The substance comprises complex aromatic and aliphatic molecular structures with various carboxyl, hydroxyl and phenolic functional groups. Humic acids typically range in molecular weight from about 10,000 to 100,000 Da. In aqueous environments, these acids generally form colloidal suspensions, with hydrodynamic diameters spanning a few nanometers to several hundred nanometers. The size of these colloids is influenced by ionic conditions, where increased ionic strength encourages aggregation, resulting in larger particle sizes. Electrolyte cations can decrease electrostatic repulsion and foster intermolecular interactions among humic acid molecules, thus promoting rapid aggregation. Given that humic acid is a complex mixture with a broad molecular weight distribution and nanoplastics surfaces are frequently functionalized, the presence of humic substances in environmental snow samples presents a significant analytical challenge. Humic acid can act as a competitive binding agent, interacting with both the dyes and nanoplastics, thus altering their affinities and reducing the reliability of our percentage binding measurements. In addition, humic acids can significantly influence the aggregation of nanoplastics. Furthermore, humic acid strongly absorbs UV–visible light and emits intrinsic fluorescence across a wide range of wavelengths from 200 to 300 nm, , with significant spectral overlap with the absorption spectra of rhodamine 6G, coumarin 334, and fluorescein. Consequently, the actual percentage binding results for rhodamine 6G, coumarin 334, and fluorescein with nanoplastics could be under- or overestimated if humic acid fluorescence excitation and emission contribute to the dye peak signal measurements by CE-LIF. This spectrally interferes with the fluorescence emission intensity measurements of dyes, impacting the accuracy and reliability of nanoplastics detection in our dye-binding approach via CE-LIF. Another critical issue is the potential for static and dynamic quenching effects caused by humic acid. Static quenching occurs when humic acid forms nonfluorescent complexes with the dyes, while dynamic quenching results from collisional interactions between humic acid molecules and dye molecules. Both mechanisms reduce the dye fluorescence intensity, leading to an overestimation of dye binding to nanoplastics.

Nanofiltration membranes can physically separate nanoplastics from humic substances due to size exclusion. High selectivity and effectiveness can be expected, but membranes are prone to fouling. , Humic substances cannot be effectively removed from aqueous systems using recycled bed sand from fluidized bed reactors. This method has demonstrated the capability to remove up to 80% of humic substances, with steady-state removal efficiencies ranging from 20 to 35%, depending on factors such as flow rate and particle size. To avoid contamination of snow samples (intended for analyzing nanoplastics) by humic substances in soil, an analytical chemist can adhere to the best practices and methodologies. Sampling sites should be chosen distant from exposed soil or vegetation that might contribute humic substances. Snow should be collected from areas not affected by meltwater runoff or where soil exposure is minimal. Snow should be collected from subsurface layers, discarding the uppermost layer to minimize contamination from airborne particulates of humic substances. Field blanks can be prepared by exposing clean containers to the sampling conditions without collecting any snow to assess potential contamination sources. Procedural blanks (containers filled with clean snow) within the analysis batch to identify laboratory-induced contamination. Ultimately, surface-enhanced Raman spectroscopy can be utilized to verify the absence of humic acid in the melted snow sample. The Raman spectrum of humic acids is marked by broad, less well-defined peaks, reflecting their complex and heterogeneous structure. Both nanoplastics and humic acids exhibit similar peaks in the aromatic and aliphatic regions. Spectral measurements, carefully corrected for baseline variations, can confirm the presence of nanoplastics. Typically, bands are detected around 1350 cm^–1 (D-band) and 1580 cm^–1 (G-band), similar to those in carbonaceous materials, attributable to aromatic and aliphatic CC stretching. Raman peaks from functional groups such as carboxylic (COOH), hydroxyl (OH), and ketone (CO) appear at various wavenumbers, adding complexity to the spectra. Avoiding spectral overlap and distinguishing signature peaks are crucial for the accurate trace detection and quantification of nanoplastics in environmental samples containing humic substances or soil. Advanced spectral analysis techniques, such as deconvolution and chemometrics, , can enhance the resolution and identification of nanoplastics in complex mixtures involving humic substances. A key assumption is that the concentration of humic substances is not so high as to cause a significant reduction in laser intensity during spectroscopy.

Conductivity

The contamination of fallen snow by road salt can potentially impact the electrokinetic injection of dyes within melted snow when conducting capillary electrophoretic analysis. As shown in Table , the electrical conductivity of the snow samples varied between 15 and 97 μS. Fortunately, this range is significantly lower than the conductivity observed for 10 mM Na₂HPO₄ and did not affect the electrophoretic migration times of any dyes. No conductivity correction or normalization was required. When a 1.0 mL aliquot of the 3-dye consortium solution in 20 mM Na₂HPO₄ was combined with 1.0 mL of melted snow to create an unknown sample for CE-LIF analysis, the minimal variations in the conductivity of melted snow were unlikely to significantly affect the electrokinetic injection of dyes.

3. Conductivity Measurements at Room Temperature (23 °C).

	conductivity (μS)
sample	mean ± standard deviation (n = 4)
Deionized water	4 ± 1
Distilled water	7 ± 1
Reverse osmosis water	80 ± 3
South Keys snow 1	64 ± 2
South Keys snow 2	37 ± 2
Kemptville snow	16 ± 2
North Grenville snow	94 ± 3
Kingston snow	58 ± 2
10 mM Na2HPO4 in water	1430 ± 14

Hydrophobicity of Organic Fluorescent Dyes

Liquid chromatography (LC) is an essential technique for determining the retention time of organic compounds, serving as an indicator of their hydrophobicity relative to other molecules. The retention time is influenced by various factors, including the composition and flow rate of the mobile phase, temperature, and the specific stationary phase employed. In reversed-phase liquid chromatography (RP-LC), which involves a nonpolar stationary phase and a polar mobile phase, more hydrophobic compounds tend to interact more strongly with the stationary phase. Consequently, these compounds typically have longer retention times as they spend more time interacting with the stationary phase rather than moving with the mobile phase. Although liquid chromatography holds potential for detecting nanoplastics, its use in the analytical separation of organic dyes is comparatively underexplored. The separation of pyranocoumarin was conducted using a symmetry C18 column with a mobile phase of methanol and water (70:30 v/v), via 276 nm UV or diode array detection. , For coumarins, a mobile phase consisting of methanol and 0.3% acetic acid in water was employed on a C18 column. Fluorescence detection was conducted with an emission wavelength range of 270–500 nm, and an excitation wavelength of 300 nm. In our laboratory, high-performance liquid chromatography (HPLC) was employed to evaluate the hydrophobic characteristics of three organic dyes. Two different mobile phase compositions were tested. The mixture of water/methanol/acetonitrile (28:12:28 v/v) achieved short retention times, whereas the combination of water/methanol/acetonitrile (28:21:28 v/v) provided good peak resolution, as outlined in Table . In both scenarios, retention times increased sequentially from fluorescein to coumarin 334 and then rhodamine 6G. This trend suggests that fluorescein has the lowest hydrophobicity, while rhodamine 6G demonstrates the highest. These results explain why fluorescein showed the lowest percentage of binding with nanoplastics, whereas rhodamine 6G exhibited the highest binding percentage. However, it must be noted that the hydrophobic properties of these dyes do not directly correlate with their binding percentages to various snow samples. This discrepancy might be due to the fact that hydrophobic interactions form only a part of the binding mechanism. Electrostatic attraction/repulsion and π-π interaction also play a role, which cannot be quantitatively assessed through RP-HPLC characterization. In RP-HPLC, the ionization state of an organic dye can lead to reduced retention times due to an increase in net polarity or dipole moment. This enhanced polarity diminishes the interaction with the nonpolar stationary phase, even if the dye’s inherent hydrophobicity remains unchanged. The ionization state is governed by the pK _a of each dye. Rhodamine 6G is typically characterized as a cationic dye upon protonation, with a pK _a of 3.8. At pH 7.0, the ratio [R6G]/[HR6G+] = 10^{pH–pK _a} = 10^3.2 = 1585, indicates that rhodamine 6G is only 0.06% protonated as the cationic HR6G ⁺ species. Thus, its long retention time in RP-HPLC is justified. Conversely, fluorescein, with a pK _a of 6.4, has a ratio [Fluor^–]/[HFluor] = 4.0 at pH 7.0, indicating 20% protonation as the neutral HFluor species, and 80% present as the anionic Fluor^– species, explaining its short retention behavior in RP-HPLC. The pK _a of coumarin 334 is not widely reported in the literature, as it remains primarily neutral under most pH conditions, resulting in a retention time between those observed for R6G and fluorescein in RP-HPLC.

4. HPLC Assessment of Dye Hydrophobicity Characteristics Based on Retention Time .

organic dye	retention time (min) mobile phase 28:12:28 v/v	retention time (min) mobile phase 28:21:28 v/v
Rhodamine 6G	6.12 ± 0.35	23.0
Coumarin 334	5.65 ± 0.21	21.5
Fluorescein	5.58 ± 0.25	20.0

^aColumn: Zorbax Eclipse XDB C8 (4.6 × 150 mm); column temperature: 23°C; mobile phase: water/methanol/acetonitrile (28:21:28 v/v or 28:12:28 v/v); flow rate: 1.0 mL/min; sample solvent: deionized water at pH 7; injection volume: 5 μL; fluorescence detector: excitation wavelength: 500 nm; emission wavelength: 550 nm; fluorescence detector gain: constant across runs.

Normalized % Binding Ratio

One of the major concerns is that quantification of nanoplastics using state-of-the-art methodologies was often reported without chemical confirmation of polymer types. The current methodology uses a variety of binding tests to assess the interaction of nanoplastics with three organic fluorescent dyes. By incorporating more diverse dyes into the mix for binding assays, a more comprehensive profile of how nanoplastics interact with different compounds can be achieved. Under controlled analysis conditions of background electrolyte concentration, ionic strength, pH, and temperature, each type of nanoplastic exhibits a unique binding profile in the present study. This binding profile is influenced by distinct mechanisms and can be standardized by chemical compositions characteristic to each nanoplastic type. For Carleton University snow sample 1, the normalized binding ratio of R6G: C334: Fluor was 2.52:1.00:1.48. This ratio parallels those observed for C334 and fluorescein but diverges from normalized ratios found in commercially available polystyrene nanospheres, lab-synthesized polystyrene nanoparticles, and PS–PET composite nanoparticles. In contrast, humic acid demonstrated a normalized binding ratio of 2.52:1.00:0.89. Natural organic matter (NOM) in soils typically possesses a net negative charge, comprising various functional groups, primarily carboxylic (−COOH) and phenolic (−OH) groups, which dissociate hydrogen ions (H⁺) in soil environments, imparting a net negative charge to the NOM. The natural zeta potential of organic soils ranges from −11.2 to −20.8 mV, depending on the organic content and degree of humification. The nearly identical ratios for R6G: C334 substantiate contamination of snow sample 1 with negatively charged humic acid, which exhibits strong binding affinity with R6G. However, significant variances in the ratios for C334: Fluor suggest the presence of another positively charged contaminant that binds strongly with fluorescein. Nanoplastics may consist of cationic polymers such as polyethylenimine (PEI), which are positively charged at neutral pH. Nanoplastics functionalized with amine (NH₂) groups exhibit positive charges, especially when environmental pH is below the pK _a of the amine, causing protonation of the amine groups. The incorporation of quaternary ammonium compounds can render the surface of nanoplastics permanently positively charged across a wide pH range.

The normalized percentage binding ratio serves as a molecular fingerprint for nanoplastics, offering insights into surface functionalization for the identification of nanoplastics in complex mixtures. Poly­(ethylene terephthalate) (PET) nanoparticles are among the most abundant airborne nanoplastics found in urban and suburban environments. Bis­(2-hydroxyethyl) terephthalate (BHET) is an organic compound and an intermediate in the production of PET. By modifying an emulsion polymerization method previously reported by Nuruzatulifah, several composite nanoplastics were synthesized from styrene and varying mole percentages of BHET. All synthesized composites were examined using transmission electron microscopy to determine their morphology and size distribution. Their binding ratios with the 3-dye consortium, normalized to C334, are plotted against their BHET mole percentages as illustrated in Figure . The trendlines indicate that the binding behavior is influenced by the mole percentage of BHET in the PS–PET composite nanoparticles. The trendline for C334 is horizontal, reflecting its role as the reference dye. The trendline for the positively charged R6G shows an upward slope due to the enhanced ionic attraction with increasing negative charges of BHET. Meanwhile, the trendline for fluorescein exhibits a downward slope, attributable to its ionic repulsion against BHET. Fluorescein is amphiphilic with both hydrophilic and hydrophobic characteristics; in its deprotonated form at neutral to basic pH, fluorescein is predominantly hydrophilic and negatively charged. The negative charge of PET likely repels fluorescein. These trends support our hypothesis that surface charge influences the binding of dyes with nanoplastics in this study. Additionally, PET’s hydrophobic aromatic ring structure may facilitate hydrophobic interactions with R6G and fluorescein, promoting binding.

6.

Trends of normalized binding ratios with increasing BHET mole percentages in synthesized PS–PET composites by CE-LIF analysis of a consortium of rhodamine 6G, coumarin 521, and fluorescein (125 μg/mL each). BGE: 10 mM Na₂HPO₄ at pH 7.1; injection: 10 s at 15 kV; CE voltage: 15 kV; CE current: 60 μA; diode laser λ_ex: 450 nm; optical interference filter λ_em: 520 nm.

New challenges remain when monitoring atmospheric nanoplastics within the broader exposome framework due to potential environmental and biological interactions. They can interact with bee pollen components, changing their surface properties and forming eco-coronas. Nanoplastics can adsorb various organic and inorganic materials from the environment, leading to the formation of eco-coronas. Chemical digestion using oxidizing agents can effectively break down organic matter when required. However, selecting the appropriate reagent is critical to minimizing the degradation of polymers. Nitric acid (HNO₃) digestion can remove organic matter in approximately 2 h, yielding polymer recovery rates between 64% and 110%. Magnesium, iron, and aluminum oxides exhibit strong adsorption capacities due to positive surface charges in acidic soils, which attract the anionic forms of fluorescein. , Soil is a complex and dynamic natural resource. Various naturally occurring clays such as kaolinite, montmorillonite, and illite cannot bind fluorescein due to their layered lattice structures and charge characteristics. The point of zero charge (PZC) is the pH at which a material’s surface carries no net electrical charge. Clay minerals like kaolinite, montmorillonite, and Illite typically have PZC values around 2–4, meaning their surfaces mostly remain negatively charged at natural soil pH values. All three clay minerals increase in anionic character as pH rises from 3 to 9, with montmorillonite possessing the highest negative charge. ,

Capillary Electrophoresis with UV and MS Detection

In developing this novel CE-LIF method, a consortium of fluorescent organic dyes was meticulously selected based on aqueous solubility, laser excitability, fluorescence quantum yield, and atmospheric rarity. Atmospheric rarity ensures that these dyes are absent from snow samples, thereby preventing interference. Laser excitability, specifically at a wavelength of 450 nm, ensures selectivity by avoiding absorption by other organic compounds in the snow. If LIF is considered an uncommon accessory in the laboratory, UV detection can be used as an alternative to address this instrumentation challenge. Figure presents a typical CE-UV analysis of the consortium blank, demonstrating the method’s feasibility. The three dyes yield UV absorption peaks at relatively long migration times due to the application of 15 kV, compared to 18 kV as shown in Figure , across a longer capillary. However, a drawback of this method is the inevitable appearance of interference peaks from other organic compounds present in the snow. In CE analysis, selecting 15 kV or 18 kV involves a trade-off between separation quality and analysis speed, influencing resolution, run time, and safety. Utilizing 15 kV enhances peak separation because of slower dye migration. This lower voltage improves resolution and precision in measuring the closely migrating R6G and C334 peaks but requires a longer analysis time for the last fluorescein peak, which may be inefficient in high-throughput scenarios with large sample batches. Additionally, reproducibility of migration time and peak height measurements may be compromised. However, this voltage reduces the risk of electrical arcing and minimizes baseline noise in electropherograms, leading to cleaner outputs. The increase from 15 kV to 18 kV was unlikely to affect analyte detection and accuracy.

7.

Typical analysis of dye consortium blank by CE with UV absorbance detection at 200 nm. Peaks for R6G, C334 and fluorescein appear at 5.8 min, 6.8 and 12.9 min, respectively, under an operating voltage of 15 kV. Capillary length from inlet to outlet: 68 cm; to point of UV detection: 62 cm.

Arguably, HPLC-UV and liquid chromatography with mass spectrometry (LC-MS) represent viable alternatives for the quantitative analysis of the 3-dye consortium. The UV and MS detection modes offer significant advantages, including simple instrumentation, enhanced selectivity, and increased sensitivity in quantifying all three dyes. Nonetheless, a major limitation of both HPLC and LC is their inability to determine the charge state of each dye, which constrains our understanding of ionic attraction and repulsion involved in the binding mechanism.

Confirmation of Nanoplastics by Dopamine Polymerization

The dye binding method utilized an indirect or subtractive approach for detecting nanoplastics in snow samples by analyzing the electropherogram of unbound fluorescent organic dyes from a three-dye consortium. To obtain a multidimensional characterization of the nanoplastics in snow, a direct approach involving laser-induced fluorescence imaging of dye-bound nanoplastics in the melted liquid of each snow sample could be used. However, the small size of nanoplastics renders this fluorescence imaging technique inadequate for accurate measurement due to insufficient spatial resolution. The fluorescence lifetime of dye moieties varies with different plastics, allowing for discrimination between types of nanoplastics. A rapid on-site detection protocol using Nile red staining in n-heptane has been established. Adsorption of fluorescent organic dyes on nanoplastics can cause shifts in their emission peaks, leading to either hypochromic (blue) or bathochromic (red) shifts due to changes in the dye’s local microenvironment. This interaction may also broaden or narrow the emission spectrum, depending on the extent of dye aggregation and binding nature. The quantum yield may decrease if fluorescence quenching occurs due to close proximity of dye molecules on the nanoplastic surface, facilitating nonradiative decay pathways. In certain instances, adsorption might enhance quantum yield if nanoplastics stabilize the dye’s excited state, thereby reducing nonradiative processes. Nanoplastics may also enhance photostability by isolating dye molecules from quenching oxygen or reactive species. Alterations in emission properties and quantum yield directly affect the sensitivity of fluorescence-based methods for environmental snow monitoring. The chemical nature and surface functionality of the nanoplastics crucially influence these changes. In our preliminary study, capillary electrophoresis with ultraviolet detection (CE-UV) served as a straightforward method for analyzing a sample (2.0 mL) that included dopamine hydrochloride (5.0 mg) and poly­(ethylene terephthalate) nanoplastic (5.0 mg) in 10 mM Na₂HPO₄ (pH 7.0). The sample was pretreated by sonication in a water bath (at a room temperature of 22.5 ± 0.5 °C) for 180 s and left standing on the lab bench under room light over 49 days. The electropherogram in Figure displays the UV absorption peaks of dopamine (at 7.81 min), poly­(ethylene terephthalate) (at 13.59 min), and dopamine-coated poly­(ethylene terephthalate) (at 13.71 min) under a CE operating voltage of 15 kV. Despite slight overlaps in the UV absorption peaks, peak deconvolution was unnecessary as only migration times were measured, not peak heights.

8.

CE analysis with 200 nm UV absorbance detection of dopamine (at 7.81 min), dopamine-coated poly­(ethylene terephthalate) (PD–PET) at 13.59 min, and polydopamine at 13.71 min in 10 mM Na₂HPO₄ (pH 7.0) under an operating voltage of 15 kV. Capillary length from inlet to outlet: 68 cm; to point of UV detection: 62 cm.

Polymerization of dopamine on nanoplastic surfaces offers a promising direct detection approach. Dopamine can undergo spontaneous oxidative polymerization to form polydopamine, which adheres readily to surfaces of various nanoplastics. This process is feasible in aqueous solutions at neutral pH, making it suitable for environmental snow and water samples. Fine-tuning polymerization conditions to ensure consistent coating thickness and uniformity is key to success. Once dopamine polymerizes to polydopamine on nanoplastic surfaces, it forms a coating providing abundant reactive sites for fluorescent probe tagging, enhancing visibility and detection. A polydopamine layer can improve the optical and electrochemical properties of nanoplastics, facilitating detection through various techniques. Polydopamine’s unique properties allow for enhanced interaction specificity. Specific binding agents can be integrated into the polydopamine coating, offering selectivity for certain nanoplastics based on their surface chemistry. Fluorescence microscopy becomes feasible by covalently attaching fluorescent dyes to the polydopamine coating. Raman spectroscopic signals can be enhanced via specific Raman-active label conjugation. Electrochemical sensors may be designed using polydopamine’s conductive properties for signal amplification with a redox couple. Polydopamine features unique properties such as wet adhesion, metal ion chelation, and hybrid electronic–ionic conductivity. Its molecular structure includes dihydroxyindole, indoledione, and dopamine units variably unsaturated and covalently linked by C–C bonds between benzene rings. A weak alkaline condition (pH > 8) is generally required for dopamine’s oxidative self-polymerization. However, generating protonated polydopamine (H-PDA) under acidic conditions via a hydrothermal method has expanded stable dopamine self-polymerization to pH 4 and pH 1. , H-PDA nanoparticles in aqueous suspension exhibit maximum UV absorbance at 200 nm. Synthetic polydopamine films have gained interest for diverse applications owing to their versatility as surface functionalization and adhesive coatings. These applications leverage the robust polymer films’ ability to deposit onto a wide range of surfaces as well as their inertness to harsh chemical environments. A PDA growth-based method for trace analysis of nanomaterials in aqueous suspension has been developed using colloidal silica and other inorganic nanoparticles as model targets. , PDA coatings were used in conjunction with polymer brush synthesis and click chemistry to create glycopolymer particles for biomedical applications. Sulfur-doped carbon dots were encapsulated with PDA on a glass fiber membrane to produce yellow fluorescence emission that enabled quantification of polyethylene microplastics with a lower detection limit of 4 mg. Incorporation of PDA enhanced the stability and uniformity of zeolitic imidazolate framework loading on the diatomite surface to adsorb 91.5% of polystyrene nanoplastics within 5 min. Modification of a carbon paste electrode with iron-incorporated PDA elicited strong π–π stacking interactions for the specific detection of polystyrene over other nanoplastics such as poly­(ethylene glycol) and poly­(vinyl alcohol).

The present dopamine-based detection method offers new possibilities and holds great potential for future applications. Previous studies indicate that the dopamine polymerization kinetics and hence the rates of polydopamine formation are notably influenced by pH levels. As the pH becomes more basic, the rate of polymerization tends to increase, potentially due to enhanced deprotonation and oxidation of dopamine molecules, which facilitates polymer formation. A common approach to study dopamine polymerization involves maintaining a slightly basic environment (optimally around pH 8.5). The deposition rate was found to be approximately 0.399 nm/min at pH 7.9, 0.466 nm/min at pH 8.2, and 0.486 nm/min at pH 8.5. Tris buffer solutions are commonly used to maintain consistent pH levels during experiments. In alkaline conditions, dopamine undergoes oxidative polymerization, where oxygen facilitates its oxidization to dopamine-quinone. Subsequently, the dopamine-quinone rearranges to yield leukodopaminechrome and polymerizes to form polydopamine. As the pH becomes more basic, the rate of polymerization tends to increase, potentially due to enhanced deprotonation and oxidation of dopamine molecules, which facilitates polymer formation. The pH influences also the mechanical and chemical properties of the resulting polydopamine coating or film. A stable pH of 8.5 ensures consistent morphology and functional group presentation on the film surface. Initiators or catalysts can be used to adjust the pH dependency, enabling polymerization at different pH levels by facilitating dopamine’s oxidative pathways. Further research will assess how these rates translate into practical applications in the detection of nanoplastics in snow.

Conclusions

Studying nanoplastics in the atmosphere is essential for understanding their role in the global plastic transport cycle, evaluating their health and environmental risks, and informing future research and policy decisions. This research developed a CE-LIF method for the analytical detection and characterization of nanoplastics based on their physicochemical properties and large specific surface area. The dyes rhodamine 6G, coumarin 334, and fluorescein were chosen for their distinct laser-induced fluorescence excitation/emission properties and good water solubility. Their ionic charges (positive, slightly positive, and negative, respectively) and aromatic structures facilitate binding with nanoplastics through electrostatic forces, hydrophobic and π–π interactions. Opportunities exist for researchers interested in generating more experimental data on the interactions between various dyes and different nanoplastics. The proposed dyes-CE-LIF method offers advantages, including minimal sample preparation, rapid analysis, high sensitivity, low detection limit, and reduced need for specialized operators, making it a promising tool for detecting and quantifying nanoplastics in melted snow samples. While the rapid binding kinetics of the dyes on nanoplastics expedite the process, the 15 min analysis time is determined by the need for complete CE separation of all three dye migration profiles. This green method significantly contributes to chemical and environmental toxicology studies, enabling chemists and scientists to use a common CE instrument for atmospheric snow research. Developing advanced sample preparation strategies will be crucial for achieving higher levels of nanoplastics identification and quantification. We envision a strategy that can preconcentrate nanoplastics without aggregation by sample filtration through a membrane to perform solid phase extraction. Reference to Canadian jurisdictional guidelines and consortiums dedicated to nanoplastic study, along with tools from initiatives like the NIST Microplastic and Nanoplastic Metrology Project, could yield valuable insights.

This work presents a new approach for detecting nanoplastics in snow samples using CE-LIF, coupled with organic dye binding. It provides an interesting proof-of-concept method involving rhodamine 6G, coumarin 334, and fluorescein, and apply this method to snow samples. The work does not provide any quantitative results on nanoplastics in the collected snow samples (e.g., μg/mL or particle counts). Future work would establish standard curves using commercially available nanoplastics to enable proper quantification. Our research focused on developing a dye-binding method using snow samples collected from representative sites in four cities to demonstrate the potential applicability of the new method for detecting nanoplastics. A larger sampling campaign will be necessary to verify the general conclusions after further work to validate the proposed method with certified or standard reference materials. Monitoring snowmelt quality is crucial, especially in areas where it significantly contributes to surface water sources. For example, studies have indicated that snowmelt runoff can influence river water quality, particularly during high discharge periods. In British Columbia, high correlations with sediment concentrations have been observed at stream discharges exceeding 1.0 m³/s during snowmelt. The method establishes a robust framework for future research in nanoplastics identification, environmental monitoring, and health impact evaluations. Advancements in linear algebra could facilitate the classification of different snow sources, allowing for the creation of a snow quality index (SQI) to evaluate atmospheric pollution by airborne nanoplastics. This research will assist decision-makers in crafting regulations to protect local populations from inhaling airborne nanoplastics, thus managing sustainable public health. In future work, the proposed method could be applied to other matrices, such as rain and lake water, or more complex environmental samples like soil leachates, to assess its potential limitations. To refine the current method, comprehensive characterization should integrate various methodologies for accurate nanoplastics identification and quantification. Coupled/hyphenated instrumental methods provide complementary analysis data, enhancing insights into nanoplastics within complex samples. Combining the dyes-CE-LIF method with techniques like centrifugal field-flow fractionation coupled with Raman microscopy and asymmetric flow field-flow fractionation coupled with pyrolysis GC–MS can enhance specific size fraction separation, polymer identification, and mass-based semiquantification. Future experiments will investigate other potential contaminants affecting dye-nanoplastic interactions. Understanding the mechanisms by which snowfall captures and deposits nanoplastics can aid in developing mitigation strategies and exploring their interactions with other airborne pollutants. Other related areas for further research include exploring how snow precipitation impact nanoplastic deposition rates and investigating the effect of nanoplastics on snow albedo and subsequent climate feedback mechanisms.

This work did not systematically test the dye binding behavior using well-characterized nanoplastics such as PE, PET, PS and PVC. Future work could use standard reference materials to validate the method’s specificity and to anchor PCA interpretations. The accuracy and reliability of this molecular fingerprinting approach for inferring polymer type in complex environmental samples could also be assessed. The novel approach of enhancing molecular fingerprinting using normalized % binding ratios could significantly bolster the capacity to identify and quantify nanoplastics in environmental samples. Advanced statistical and machine learning models could help interpret complex data from these assays, leading to better predictive models for nanoplastics behavior in the environment, improving our understanding of their ecological impacts and informing policy decisions. The development of dopamine polymerization for direct detection of nanoplastics could revolutionize the field by enhancing detection capabilities, allowing for more detailed studies on the role and fate of nanoplastics in environmental systems in various matrices such as snow, water, soil, or biological tissues. Future research could focus on optimizing the binding mechanisms and exploring other catecholamine-based polymers for diverse applications.

All data needed to evaluate the conclusions in the paper are presented in the paper. Additional data sets supporting this article will be made available upon request.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c13510.

• Scree plot generated by principal component analysis. The plot assists in identifying the most significant components for reducing dimensionality without losing critical information. The magnitude of eigenvalue effectively illustrates the portion of variance explained by each principal component, relative to the total variance across all components, In snow analysis, for detecting nanoplastics, PCA can help isolate noise from essential data signals by focusing on components that explain significant variance. Alternative dimensionality reduction techniques such as independent component analysis (ICA) could be considered for complex data sets (Figure SI1) (PDF)

The authors declare no competing financial interest.