Identification of microplastics using 4‐dimethylamino‐4′‐nitrostilbene solvatochromic fluorescence

Abstract In this work, we introduce the use of 4‐dimethylamino‐4′‐nitrostilbene (DANS) fluorescent dye for applications in the detection and analysis of microplastics, an impendent source of pollution made of synthetic organic polymers with a size varying from less than 5 mm to nanometer scale. The use of this dye revealed itself as a versatile, fast and sensitive tool for readily discriminate microplastics in water environment. The experimental evidences herein presented demonstrate that DANS efficiently absorbs into a variety of polymers constituting microplastics, and its solvatochromic properties lead to a positive shift of the fluorescence emission spectrum according to the polarity of the polymers. Therefore, under UV illumination, microplastics glow a specific emission spectrum from blue to red that allows for a straightforward polymer identification. In addition, we show that DANS staining gives access to different detection and analysis strategies based on fluorescence microscopy, from simple epifluorescence fragments visualization, to confocal microscopy and phasor approach for plastic components quantification.

residues is observed, so that the small size MPs enter the food chain as internalized by living organisms, which are a primary source of food for humans (Desforges, Galbraith, & Ross, 2015;Silva-Cavalcanti, Silva, de França, de Araújo, & Gusmão, 2017). In fact, the presence of MPs has been detected in mineral water, sea salt, sugar, honey, and several kinds of marine organisms, such as fish, jellyfish, and algae (Rainieri & Barranco, 2019). Although MPs contaminations have also been found in terrestrial ecosystems, the marine organisms are currently the most investigates species to elucidate the MPs effects on living organisms' biological functions, since they are reported to be strongly exposed to plastic debris, and could represent the main route for plastic particles introduction in the human diet (Wang, Gao, Jin, Li, & Na, 2019). In addition of their direct effects, MPs are supposed to be potential vehicles of toxic chemicals and pathogenic microorganisms as well, that can interact and bind on the MPs polymeric surface, then be delivered into the superior organisms, resulting in chemical and biological detrimental effects (De-la-Torre, 2020). Therefore, MPs are of special concern as they have been shown to induce different adverse effects, which still need to be further unveiled (Hwang et al., 2020;Ryan, 2015;Sharma & Chatterjee, 2017).
In the last few years, large efforts of the scientific community have been focused in the characterization of their main sources and pathways, with the aim to reduce their impact on the ecosystems, then on the human health (Rochman, 2018). Identification of MPs and their origin may help planning littering prevention solution of most found items into the environment (Sun, Dai, Wang, van Loosdrecht, & Ni, 2019).
In this context, a growing literature exists aimed to develop analytical methods to identify and characterize MPs (Hidalgo-Ruz, Gutow, Thompson, & Thiel, 2012;. Indeed, there is an urgent need to increase our specific knowledge about these heterogeneous and complex mixtures of polymers and additives. The understanding of MPs impact can be reached through the knowledge of their accumulation in specific areas, along with the characterization of the plastics micro-/nanoparticles in terms of size, shape, chemical composition, fragmentation, and degradation pathways. Although several analytical strategies have been applied in MPs investigation (Mariano, Tacconi, Fidaleo, Rossi, & Dini, 2021), no method has been adopted as a standard analytical approach up to date (Prata, Alves, da Costa, Duarte, & Rocha-Santos, 2020). MPs identification usually consists in a preliminary visual examination of the sample often supported by the bright-field optical microscopy (Lusher, Welden, Sobral, & Cole, 2017). This first step is largely dependent on operator and is quite time consuming. The identification procedure generally goes on with the analysis of their chemical composition by means of mass spectrometry, Fourier transform infrared (FTIR) and Raman spectroscopies, often aided by thermoanalytical methods (Käppler et al., 2016). In particular, FTIR and Raman spectroscopies implemented in imaging mode have revealed high potentiality allowing for noninvasive identification and characterization of MPs by mapping chemical properties of the sample with microscale spatial resolution (Harrison, Ojeda, & Romero-Gonz alez, 2012;Tagg, Sapp, Harrison, & Ojeda, 2015). Although these methods provide a large amount of information, including possible presence of pollutants and modifications due to environmental conditions, they may be time consuming especially for the analysis a relevant number of particles (Anger et al., 2018). Furthermore, due to spatial resolution limits, they may be less effective for the identification of MPs with size in the submicron scale leading to inaccurate observations such, for example, the underestimation of MPs contamination levels in the samples. Analytical low time-consuming methods with minimal specimen preparation are largely desired which may allow to map and quantitatively characterize this heterogeneous system, simultaneously giving the possibility of monitoring biochemical processes or MPs interactions with living systems. An ideal method would enable to characterize and distinguish MPs, to readily quantify their morphological properties, to identify the polymeric constituents, and to monitor degradation effects. Ideally, in situ 4D (x,y,z,t) measurements in different matrices should be accessible of samples coming from seawater, soils, and biological tissues.
In this context, fluorescence microscopy has emerged as an ideal tool for these applications providing fast high-resolution analysis of MPs through the use of fluorescent dyes, small organic molecules whose spectral properties critically change depending on the interaction with the environment (Cole, 2016). When interacting with MPs, some fluorescent dyes allow the plastic fragments visualization with high signal to noise ratio. In specific cases, such an interaction can be also used to gain relevant information concerning changes in the dye absorption and fluorescence spectra, quantum yield, and fluorescence lifetime, that can be related to the MPs composition (Prata, Reis, et al., 2019;Sancataldo, Avellone, & Vetri, 2020). In this framework, a large number of studies have focused on the use of the Nile Red (9-diethylamino-5H-benzo [a] phenoxazine-5-one) fluorescence, as a cheap and easy tool for the identification of MPs in environmental samples, also accessing automated analysis (Erni-Cassola, Gibson, Thompson, & Christie-Oleza, 2017;Hengstmann & Fischer, 2019;Prata, Reis, et al., 2019;Tamminga, 2017).
Nile red stains the MPs resulting in fluorescent plastic fragments, whose emission varies from yellow to red according to the polarity of the constituent polymers (Maes, Jessop, Wellner, Haupt, & Mayes, 2017;Shim, Song, Hong, & Jang, 2016). The classification of MPs according to their polarity represents a useful property to be assessed as related to important features of the materials.
Moreover, analytical fluorescence-based experiments at increasing level of complexity can be also designed to investigate MPs physicochemical modifications by mimicking different environmental degradation processes, for example, sunlight exposure, O 2 /watermediated oxidation, thermally induced modifications. Thus, it is noteworthy that the use of fluorescent dyes may have feasible applications in the MPs analysis, as the particles can be easily detected, counted, and classified by means of a wide-field fluorescence microscope, a common instrumentation widely present in many laboratories. This approach can be carried out for MPs detection in water and sediments samples, as well as in biological matrices by means of confocal microscopy. In particular, concerning the biological samples, the fluorescence microscopy has stood out and affirmed as one of the principal technique for noninvasive analyses of tissues and cells, with consolidated protocols that can be directly translate for the MPs detection in living organisms, toward a more comprehensive knowledge of the MPs biochemical fate and their effects on human health.
Since fluorescence microscopy and related quantitative techniques in MPs research are evident promising analytical tools, the identification of other MPs interacting fluorescent dyes and their spectroscopic properties is a key point.
In this work, we show how the dye 4-dimethylamino-4 0nitrostilbene (DANS) allows discriminating different MPs, with high selectivity, according to their polarity. DANS is a small fluorescent molecule whose emission spectrum is known to dramatically redshift as the polarity of its microenvironment increases (Lakowicz, 2006). Its positive solvatochromic behavior is known since the 1970 and it is attributed to charge transfer mechanisms which stabilize the molecule in the excited state (Reichardt, 1994). Physicochemical properties leading to changes in DANS fluorescence are then analogous to the ones regulating Nile Red.
However, the spectral shift of DANS emission spans over a larger wavelength range, respect to Nile Red, providing larger sensitivity (Lakowicz, 2006). Proof of concept of the analytical applicability of DANS fluorescence to the identification of MPs from common use plastic items is given in this work, using five different model plastics: polypropylene (PP), low-and high-density polyethylene (LDPE and HDPE, respectively) polystyrene (PS) and polyethylene terephtalate (PET). These plastics are among the most commonly produced ($80% of worldwide production) and their presence in form of MPs is highly abundant both in marine and in freshwater ecosystems (Shahul Hamid et al., 2018).
Our results clearly indicate that DANS solvatochromic behavior confers to plastics different fluorescence emissions, spanning from blue to red, over the entire visible range, once the sample is illuminated with UV-light. We show that a qualitative screening of MPs is made easier as the MPs assume distinguishable colors, which underlies significant differences both in terms of shape and position of the DANS fluorescence emission band, making the use of the dye well suited for the application of spectral phasor analysis. The coupling of DANS staining with spectral phasor analysis provides a simple graphical method for the rapid identification and counting of MPs in the samples. We present different analytical strategies based on DANS fluorescence microscopy, from simple epifluorescence fragments visualization, to confocal microscopy and phasor approach for plastic components. Samples were incubated for 1 h at 60 C under vigorous agitation.

| Attenuated total reflection-FTIR
An FTIR spectrometer (Bruker ATR FTIR, model ALPHA) in attenuated total reflection (ATR) mode, equipped with a diamond measurement interface and controlled by OPUS software, was used to collect IR spectra. Spectra have been acquired in the range 4,000-400 cm À1 with a resolution of 2 cm À1 . Each measurement is the result of the average of 64 scans. The ATR diamond crystal was cleaned with 70% ethanol/water and a background measurement was performed between each sample. Sample was compressed against the diamond to ensure good contact between sample and ATR crystal.

| Steady-state fluorescence
Steady-state fluorescence measurements were carried out on a JASCO FP-8500 spectrofluorimeter at room temperature. All samples dispersed in water were placed in a 1 cm path plastic disposable cuvette (BRAND) and spectra were measured under excitation at λ exc 405 nm using 0.5 nm wavelength step intervals, with excitation bandwidth of 5 nm and emission bandwidth of 10 nm using scan-speed of 100 nm/min and integration time of 1 s. To avoid MPs sedimentation, a crosshead magnetic stirring bar (BRAND) was used for dispersing the microsized sample in the water solution (400 rpm). RGB pictures of stained sample, under UV excitation (LED 395 nm, Eletorot), were simply recorded by a phone camera (Huawei P20). The autofluorescence of each sample was checked and it is null or neglectable in all presented experimental conditions.

| Fluorescence confocal microscopy
Confocal images were acquired with an Olympus FluoView1200 confocal laser scanning microscope (Olympus, Japan) using a 10 Â 0.3 NA objective. Aliquots of DANS-stained MPs dispersed in water were deposited on a cover glass (BRAND #1) using a disposable pipette.
Measurement was acquired using laser excitation at 405 nm. Emitted fluorescence was acquired in photon-counting mode. Spectral detection has been performed using a bandwidth of 5 nm and a step size of 3 nm in the range 420-740 nm. The scan area was 256 Â 256 pixels and the scan speed was 12 μs per pixel.

| Phasor analysis
The spectral phasor is a global algorithm for spectral imaging allowing the analysis of the heterogeneous emission of fluorescence molecule in images. It enables to map and characterize spectral features of multiple dyes or to detect spectral shape variations. The detailed description of the method and its applications is reported in Fereidouni, Bader, and Gerritsen (2012) ;Fereidouni, Bader, Colonna, and Gerritsen (2014).
Briefly, in this work, reported spectral phasors are obtained from the first harmonic of the Fourier transformation of the spectrum measured at each point (pixel) of a fluorescence microscopy image. The real and the imaginary components of the transform, calculated according to the following definitions, are the X and Y component of the phasor which is plotted in a polar plot: where I (λ) is the measured intensity at each step of the spectrum, L is the amplitude of the spectral range and n = 1. In the presented work, step is 2.5 nm and L is 320 nm. Each pixel in the spectral image contains the emission spectrum at that pixel. Therefore, an emission curve is associated with every pixel. From the definition, X and Y coordi-

| RESULTS
MPs used in this study were obtained by uniformly filing commercial plastic objects identified by the standard numbered symbols according the International Resin Identification Coding Systems (Worrell & Reuter, 2014). These plastics were used as models to proof that the analytical use of DANS fluorescence properties allows fast and reliable identification via the fluorescence color they assume, after staining, due to the spectral properties of this dye.
In Figure 2a, we report representative ATR-FTIR spectra of PP, LDPE, HDPE, PS, and PET microsized plastics obtained by the fragmentation process. These particles appear to the naked eye as homogeneous ground powders as shown in Figure 2b. Infrared spectroscopy is, nowadays, recognized as an efficient method for identifying plastic polymers as their absorption peaks are well known and easily distinguishable. Infrared analysis provides fingerprint spectra of the specific polymer of plastic material, as well as it may reveal the possible presence of other components (contaminants, additives, or polymer blends) (Jung et al., 2018). This method is currently applied in the identification of MPs in different environments (Tagg et al., 2015). Normalized ATR spectra acquired for each sample are reported in absorbance mode in the range 3,500-400 cm À1 and present expected absorption peaks previously reported in literature (Jung et al., 2018). Characteristic absorption bands (cm À1 ) used to identify each polymer is reported in Table 1. These bulk measurements confirm MPs composition and as no spare peaks are detected; it is possible to exclude significant contamination of the plastics within the experimental error.
The same samples where stained with DANS fluorescent dye as described in the experimental section and dispersed in water.
As reported in the introduction DANS is an archetype of solvatochromic molecule whose spectrum shifts to longer F I G U R E 1 The spectral phasor plot for fluorescence emission spectra with different maximum emission wavelengths and spectral widths. (a) The phasor moves clockwise within the semicircle as the emission maximum moves from blue to red. (b) As the width of the spectrum increases the phasor moves toward the center wavelengths and critically changes its shape as the polarity of its environment increases because of the stabilization of the dipolar excited state (Lakowicz, 2006). This feature already revealed its potential in In Figure 3a, the molecular structure of DANS is reported. In   and readily accessible to many scientists. In addition, the analysis of the resulting data can be simplified using the spectral phasor approach (Fereidouni et al., 2012(Fereidouni et al., , 2014Golfetto, Hinde, & Gratton, 2014). As described in the experimental section this method allows to transform the fluorescence emission spectra measured at each pixel of the image to a point in a polar plot (phasor) whose coordinates are the real and imaginary parts the of the first harmonic of the Fourier transform of the fluorescence emission spectra. This method was originally developed for spectral demixing of multiple fluorescence signals (Fereidouni et al., 2012(Fereidouni et al., , 2014Golfetto et al., 2014;Malacrida, Gratton, & Jameson, 2015)  As can be seen in Figure 5a, the plastic fragments are easily detectable and present uneven size and shape and quite heterogeneous intensity in each image. In these, measurement is not possible To proof the possibility of distinguishing and quantifying different MPs in the same sample we report in Figure 6a the results of analogous measurements and analysis on a sample obtained by mixed MPs.
A mosaic collection of images is shown to enhance the field of view.
Using the very same selections as the ones reported in the phasor plot in Figure  We here present results using a less known small fluorescent molecule with solvatochromic properties. DANS is a fluorescent dye whose fluorescence emission spectrum modifications are due to analogous mechanisms to the one reported for Nile Red. For both dyes, the identification is mainly based on the spectral red shift which occurs with increasing polarity of the dye environment allowing classification of MPs in large chemical groups based on fluorescent shift (Reichardt, 1994

| CONCLUSIONS
In conclusion, we herein presented the possibility to use the fluorescent dye DANS for the sampling, the detection and the analysis of MPs. Different levels of fluorescence-based screening have been shown, from the simple identification of the MPs by the naked eye to the more complex phasor approach, which allows for a more refined analysis able to better distinguish the fluorescent signals even in case of similar polarity. In particular, the near-UV illumination of DANS-stained MPs results in glowing plastics fragments that clearly show different colors spanning from blue to red as a consequence of the different DANS emission spectra, which change according to MPs polarity. Differences in spectral features can be further exploited using fluorescence microscopy, which may allow to readily distinguish, count, and measure MPs by detecting the fluorescence signals from the sample in different spectral ranges. In addition, we also proposed the analysis of the spectral images obtained by fluorescence confocal microscopy to improve the performance of our approach on the analytical point of view. Spectral phasor was elected as preferential analysis method as it allows separating pixels characterized by similar fluorescence spectra, with the aim to overcome limitations due to partial spectra overlap. Moreover, it provides a simple and straightforward strategy to monitor and quantify both shape modifications and spectral shifts occurring in different spatial region of the sample.
As a future perspective, DANS staining, combined with the application of the spectral phasor method, can be considered as an emerging analytical strategy for chemical detection of the polymeric components constituting MPs detected in food and environmental samples. As well as this method can be used as a way to monitor environmentally relevant physicochemical processes occurring in MPs, such as polymer chemical degradation, or chemical and biological surface contamination, that are expected to induce spectral changes.
Additional future developments could advance this approach with neural system Ricci et al., 2020;Sancataldo, Gavryusev, et al., 2019). Indeed, depending on the experimental setup, especially on the numerical aperture of lenses, plastics fragments of size ranging from millimeter size down to light diffraction limit (≈200 nm) can be analyzed (Diaspro, 2016;Pawley, 2002). As a consequence, the presented approach can address the increasing demand of a reliable analytical strategy for nanoplastics analysis as