Methodologies to characterize, identify and quantify nano- and sub-micron sized plastics in relevant media for human exposure: a critical review

a Micro- and nanoplastics (MNPs) in the environment are an emerging issue of global concern. They accumulate in natural ecosystems, and are ingested by organisms and transferred to humans potentially causing adverse toxicological e ﬀ ects. Knowledge on the magnitude of these e ﬀ ects is limited due to the lack of knowledge on realistic exposures especially for nano- and sub-micron size plastics. Their size and shape have a signi ﬁ cant in ﬂ uence on the encountered health e ﬀ ects as well as the presence of additives. Currently, there are no standardized protocols for their reliable characterization (size, shape), identi ﬁ cation and quantitation. There is a growing number of reported studies on occurrence of microplastics above 10 m m in size and of limited polymer types (mainly polystyrene, polyethylene terephthalate, polycarbonate and polyethylene). New analytical approaches are needed for a complete and reliable risk assessment of MNPs, especially of sizes below 1 m m, on human health. This review evaluates the progress made concerning the sub-micron (100 nm to 1 m m) and nanometer (<100 nm) size range of MNPs on: (i) human exposure to evaluate the intrinsic hazards, (ii) sampling and sample preparation methods and (iii) methods for characterization (size, shape), identi ﬁ cation and quantitation, with a focus on relevant media for human exposure. Methods that could be used for the extraction of submicron and nanoplastics from relevant matrices are recommended. Novel methods ( e.g. Raman imaging and single-particle inductively coupled plasma-mass spectrometry) and new combinations of analytical methods ( e.g. atomic force microscopy coupled to infrared/raman spectroscopy, ﬁ eld- ﬂ ow fractionation-multiangle light scattering o ﬄ ine coupled to pyrolysis-gas chromatography-mass spectrometry) are proposed and discussed.


Introduction
Plastics are produced in extreme quantities over the entire globe, and their production steeply increased over the past few decades to 368 million tons in 2019. This is expected to have tripled by the year 2050 (Statista 2019). 1 Our obsession with plastic can be attributed to its extremely low cost, versatility, inertness, and durability. Currently, plastics are used in all kinds of products such as packaging, clothing, electronics, industrial materials or office supplies. They degrade and transform via mechanical, chemical and biological processes accumulating and persisting in our environment and creating an emerging threat. 2,3 Not only are many different polymers used in consumer products, but most also contain co-polymers and additives to tailor the functionality to its intended use. Some additives, such as ame retardants, are toxic and plastics which contain them are certainly not suited to be recycled into things such as children's toys or food packaging. Therefore virgin plastic is oen preferred to recycled plastic by manufacturers. 4 Of all the plastic ever produced, roughly 12% gets incinerated for energy recovery while 60% is simply disposed of and allowed to accumulate in landlls and the natural environment. 5 Plastic debris fractionates into increasingly smaller particles, micro-and nanoplastics (MNPs). 6 MNPs have the tendency to accumulate in different matrices like soil, 7 freshwater, 8 sediment, 9 sh tissue 10 and air. 11 Nanoplastics (NPs) are more reactive and potentially more harmful to humans and ecosystem. 12 There are different denitions related to the size range of NPs (Allan et al. 2021), here we use the size denition by Hartmann et al. (2019) with for nanoplastics a size of 1 to <100 nm and for submicron-plastics a size of 100 to <1000 nm. Nanoplastics are polydisperse in physical properties and heterogeneous in composition as their occurrence and production are highly dependent on the degradation of microplastics. They present colloidal behavior which can induce aggregation, depending on the physical and chemical conditions of the medium such as the ionic strength, pH, temperature and UV light. It is believed that microplastics (MPs) can be formed in the environment by four main routes: photodegradation, thermooxidative degradation, hydrolytic degradation and biodegradation by microorganisms. 2 For example, photodegradation of larger plastic debris into MPs occurs by exposure to sunlight. 13 This triggers a free radical mechanism which is auto accelerated and decreases the average molecular weight of the polymer structure over time. Aer extensive degradation of the polymer, it becomes brittle enough to disintegrate into MPs and further on to NPs. 13,14 Knowledge on exposure levels, i.e. amounts of NPs present in air and media that are ingested via food and water, is still limited due to the lack of dedicated and standardized analytical methodologies. This complicates human risk assessment.
For humans, the major exposure routes are expected to be inhalation (e.g. indoor/outdoor dust, atmospheric fallout) and ingestion (food and beverages). The aim of this review is to give an overview of potential human health effects for sub-micron and nanoplastics, their occurrence in exposure media and a critical discussion on the advantages and limitations of reported sampling, sample treatment, characterization, identication and quantication methods of MNPs in relevant media. Recommendations on future developments of analytical methodologies and new combinations of analytical techniques that are required to make a step forward and cover the current knowledge gaps are being made.
2 Human exposure and health effects

Exposure routes
Humans can be exposed to MNPs via multiple routes; inhalation, ingestion and dermal contact. 15,16 Based on food consumption, a daily intake of 107-142 microplastic particles per person was estimated. 17 Fiber exposure during a meal through dust fallout was estimated to range from 38 to 187 particles per day. 18 Microplastics have been found in multiple food samples, e.g. in table salt (PE, PP within 171-515 mm), 19,20 beer, tap water (>100 mm bers, no identication). 21,22 PS particles in the nanometer range (122-295 nm) were recently characterized in spiked sh samples from a local supermarket in Beijing and were quantied at concentrations ranging 0.068-0.146 mg g À1 , 23 mainly accumulated in the gills, liver and guts of sh, which are not usually consumed by humans. On the other hand, MSc Carlo Roberto de Bruin (corresponding author) obtained his master degree (2021) in Analytical Chemistry at the University of Amsterdam. His research was divided over two projects. He worked on a literature review on the characterization of sub-micron and nanoplastics in complex media (e.g. food, air etc.) relevant for human exposure. Aerwards, he joined Genmab BV working on method development for the characterization of released N-glycans from mAbs by high-resolution mass spectrometry. Currently he is pursuing a PhD studying lipids and phytochemicals with cyclic ion mobility mass spectrometry in complex food matrices at the University of Wageningen.
Dr Eva de Rijke works as a labmanager in the soil and environmental chemistry laboratories of the Institute of Biodiversity and Ecosystem Dynamics (IBED) of the University of Amsterdam. She is an analytical chemist who did her PhD at the Vrije Universiteit Amsterdam. Aer that she worked over 10 years as a scientist/project leader in both government and industrial R&D in the area of food quality and safety involved in research and management of (european) consortia, before returning to environmental analytical chemistry research at the University of Amsterdam. She has broad (hands-on) knowledge of a wide range of analytical techniques to determine e.g. taste/odour compounds, environmental contaminants, additives and (veterinary) drugs. She was board member of the NVMS between 2012 and 2019 and is now board member of the Benelux Association of Stable Isotope Scientists (BASIS). mussels are consumed as a whole and MNPs accumulate in mussels as well. 24 These ndings are raising concerns on the bioaccumulation of MNPs in the food chain. Contaminated food and drinking water is one of the greatest concerns in the public media. 25,26 It has been reported that exposure routes from packaging could lead to contamination of food as well. 27,28 The air contamination with MNPs has multiple sources, such as bers from clothes and abrasion of materials (e.g. plastic sheets and tires) by wind. The particles are easily transported by the wind and are very persistent. Chen et al. found that airborne microplastics are mainly from synthetic textiles and the dominant shape in the atmosphere are bers. Fibers larger than 250 mm have been observed in human lungs and may cause chronic and acute inammation. 29 Multiple studies estimated the inhalation of microplastic particles per day. For instance, Prata reviewed the consequences of the inhalation of airborne microplastics and estimated that 26-130 particles could be inhaled per day for each individual. 30 This estimation was based on measured particles in the studies from Dris et al. 11,31 and the human tidal volume 6 L per min estimated by Guyton and Hall. 32 However, Vianello et al. estimated a much higher value of 272 inhaled particles per day for each individual. 33 Within all the estimations made in the mentioned studies, variating types of microplastics were identied such as PS, PP, PE, PET, polyester, nylon etc.
Unfortunately, such estimations were not yet made for submicron or nanosized plastics and they are required for the risk assessment of NP exposure through air. The results are highly dependent of sampling and sample treatment procedures and this information should be taken into account when discussing this type of information.
Microplastic exposure through dermal contact is less plausible due to the particle size, as it is not likely that they are able to cross the dermal barrier. In contrast, nanoplastics could potentially transverse the dermal barrier, although this was not proven yet. 34 However, due to the lack of knowledge on the properties and toxicity of these particles, this possibility should not be underestimated. Cosmetics containing nanoplastics, 35,36 dust particles in the air or polluted water may be potential exposure routes for nanoplastics across the dermal barrier. In addition, it has been shown that nanoplastics are capable of penetrating cell membranes, 37 which can cause changes of behavior from sh shown by a study of Mattsson et al. 38 Forte et al. studied polystyrene (PS) nanoparticles in adenocarcinoma gastric cells (AGS), and reported that smaller sized nanoplastics (44 nm) accumulated faster and more efficiently in the cytoplasm of AGS compared to larger ones (100 nm). 39 This Prof. Dr Annemarie van Wezel (1968, MSc Biology UU, PhD environmental chemistry and toxicology UU) has long experience as scientist in water quality, risk assessment and mitigation, environmental toxicology and chemistry, and environmental policy evaluation. She was granted as applicant and co-applicant many projects in the eld of chemicals of emerging concern and water quality, examples are the European projects FP7 Solutions, ITN ECORISK2050, ITN PERFORCE3, and Dutch NWO funded projects such as Shale gas & water, TRAMP (Technologies for Risk Assessment for Microplastics), EMERCHE (Effect-directed Monitoring tools to assess Ecological and human health Risks of CHemicals of Emerging concern in the water cycle), RUST (Re-USe of Treated effluent for agriculture), PsychoPharmac'eau (Psychopharmaceuticals Prevention & Pilots to Reduce Effects in the water cycle), NWO Large Scientic Infrastructure; Authoritative and Rapid Identication System for Essential biodiversity information (ARISE) and NWO Perspective AQUACONNECT. She is interested in the scienceto-policy interface, in scientic outreach and has ample experience in media appearances. She is a member of the Dutch Health Council and the Dutch Board on authorization of plant protection products and biocides CTGB. She holds the chair Environmental Ecology and is Scientic Director of IBED (Institute for Biodiversity and Ecosystem Dynamics) at the University of Amsterdam.
Dr Alina Astefanei obtained her PhD from the University of Barcelona on the characterization of carbon nanoparticles in environmental samples. She then joined the HIMS institute at the University of Amsterdam, where she was appointed assistant professor in 2019. Her work is now directed at methodological innovation to solve problems of high impact on society, such as environmental science and art conservation. Alina is developing tools for detailed characterization and quantitation of both large and small molecules, to understand how they interact with each other, and change over time in different conditions. Field-ow fractionation and mass spectrometry (so and ambient ionization) form the main technology platform. She also coordinates the joint analytical sciences master programme of the Amsterdam Universities.
indicates that the size determines the concentration in certain body tissues, and therefore, the smallest nanoplastics might cause the most damage to human health.
In summary, for all exposure routes there is limited knowledge on the exposure levels of sub-micron and nanosized plastics.

Health effects
The exposure to sub-micron and nanosized plastics can lead to potential adverse health effects within humans. 36 Effects such as oxidative stress, chronic inammation, cytotoxicity, endocrine disruption, immune disruption and neurotoxicity [40][41][42][43][44] have been reported. These effects may be more severe for nanoplastics, due to faster accumulation in the cytoplasm, more efficient translocation and agglomeration of particles. 45 Besides the potential health effects of MNP particles, the release of additives and associated chemicals/contaminants like POPs adsorbed to MNPs may also enhance their toxicity and can cause more signicant threats to organisms than MNPs themselves. 46 Additives may leach from the plastic into the surrounding environment and enter the human body, contributing to potential health effects. 40,47 Leaching will primarily occur at the surface of the plastic particles into the body uids or tissue. For example, plastic additives like phthalates, brominated ame retardants (BFRs) and bisphenol A (BPA) are most abundant and can be a concern to human health. 48 These additives can cause severe health effects such as endocrine disruption, 49 neurobehavioral effects 50 and carcinogenesis and mutagenesis. 51 The studies mentioned above conrm the need for further research on the physical and chemical weathering that can cause breakdown of MNPs and on the interactions of digestive uid and lipids with plastic matrices within organisms. This is needed for an accurate determination of leaching chemicals and their risk assessment.
Besides the potential hazards of additives from plastics, other compounds that may adsorb to or desorb from plastics may also cause potential health effects. The hydrophobicity of plastic in combination with the high surface area of a microand especially nanoparticle, makes them great adsorbents for compounds such as POPs including polychlorinated biphenyls (PCBs), polycyclic aromatic hydrocarbons (PAHs) or heavy metals. 52 Similarly as for MNPs, these kind of compounds, can also induce carcinogenesis and mutagenesis, 53 and therefore, the combination of these compounds may enhance the toxicity of the whole complex. In this context, Liu et al. studied the adsorption of PAHs on 70 nm PS spheres and concluded a higher adsorption of PAHs to nanoplastics when compared to microplastics (>1000 nm). 54 This can be explained by the higher surface to volume ratio of nanoplastics. This nding is relevant because the abundance/amount of nanoplastics will increase over time due to the continuous degradation of (micro)plastics, leading to nanoplastics which could potentially be more harmful to humans.
Regarding the adsorption of heavy metals to MNPs, Liao and Yang conducted a study on spherical PE, PP, PVC and PS microplastics (150 mm) that serve as vector for chromium (Cr) in an in vitro human digestive model. 55 Cr can be released in the gastric and intestines phase, but in accordance with the calculated daily intake of Cox et al., the released amounts would not pose any hazards for human health. 17 However, for nanoplastics this could be different due to a probable higher adsorption of metals per volume unit, due to the larger surface area. First studies are conducted already on the potential interactions between NPs and metals (e.g. PS and Ag), which showed an increase of the harmful cellular effects. 56 This can be of concern when/if metals are added to the plastic during the production process. Furthermore, the adsorption ability of different pollutants is also dependent on several factors such as salinity, temperature, pH, dissolved organic matter and the physicalchemical properties and aging of MNPs. 57

Toxicity studies
The toxicity and potential hazardous properties of NPs are assessed by toxicity studies. Nowadays, animal or in vitro studies are used to provide knowledge on the health effects and toxicity of MNPs. 58 For example, Sökmen et al. investigated the exposure of PS nanospheres (20 nm) to zebrash embryos, and it was shown that these particles can reach the brain and bioaccumulate there, leading to oxidative DNA damage in the brain. 59 However, using animals for this purpose is complex due to qualitative and quantitative differences. To explore qualitative and quantitative differences and interactions of toxic compounds within organisms, toxicity-based-toxicokinetic/ toxicodynamic (TBTK/TD) modelling can be used. 60 TBTK/TD modelling is a powerful mechanistic approach clarifying fate and behaviors of specic toxicants, facilitating to translate exposure to time course of toxic effects on related biomarkers, for example the inhibition of cytochrome P450. A TBTK/TD model was used to quantify organ-bioaccumulation and biomarker responses from PS microplastic particles in mice, that generally serve as mammalian terrestrial model organism. 60 This model offers a framework for microplastic exposure in mammals and offers an algorithm for the extrapolation from animals to humans for health risk assessment perspective, which also have the potential to be used for nanoplastics. Unfortunately, such a study has not been reported yet.
Currently, the studies published on in vivo nanoplastics exposure are increasing, for example the studies from Auguste et al., 61 Elizalde-Velázquez et al. 62 and Wang et al. 63 In addition, ex vivo studies are gaining interest as well. 64 However, more studies are needed to clarify the mechanisms for bioaccumulation of nanoplastics in mammals.
Animal testing is not promoted due to ethical issues, and therefore in vitro studies that can provide complementary valuable information are used. Instead of TBTK/TD modelling, physiologically-based toxicokinetic (PBTK) modelling can be used, which enables animal-free risk assessment. 65 Mammalian cell lines have proven to be excellent models for the determination of cytotoxicity of potential harmful compounds to human health. Gopinath et al. exposed human blood cells to different forms (virgin, isolated and coronated) of PS nanoplastics (100 nm) in concentrations ranging from 10 to 100 mg mL À1 . 36 Conformational changes in blood protein, cytotoxicity, genotoxicity and hemolysis were observed aer different exposure durations (4 h or 24 h). In addition, there was a signicant decrease in cell viability and also damage to the DNA structure. The disadvantage of in vitro studies is the lack of insight in the bioaccumulation process of MNPs, because this process may inuence the cytotoxicity. To tackle this limitation, a combination of in vivo and in vitro studies would be more appropriate to investigate toxicity and uptake and bioaccumulation processes of MNPs. 45 For example, the question still remains if it is possible for microplastics to degrade in the body of a human or animal into nanoplastics during its excretion process. Future toxicity studies should include different types of nano-sized plastics of various shapes and sizes rather than exclusively using commercially available PS nanospheres. In this context, Gray and Weinstein investigated the inuence of different sizes and shapes of microplastics (PS, PE, PP) and it turned out that the mortality of shrimps was highest when exposed to ber shaped PP microplastics instead of spheres and fragments. 66 Until now, doses employed for exposure studies 62 (50 mg mL À1 to 0.025 mg mL À1 , both in vitro and in vivo) seem unrealistic for environmental exposure of NPs with sizes between 20 and 100 nm. In addition, variation in results may be explained by differences in chemical nature of MNPs such as; size, shape, surface chemistry, other physicochemical properties and different exposure routes. The production of commercially available nanoplastics with variating shapes and sizes needs to be expanded to support the development of toxicological and analytical studies. In Fig. 1, an overview is shown which summarizes the progress in risk assessment of MNPs.

Sampling and sample preparation
Contamination is the main issue in any sampling procedure for NP studies, because plastic equipment is widely used and therefore a signicant risk for contamination is expected to be widespread. It is thus of great importance to identify potential sources that can contaminate the samples and prevent this as best as possible. Obviously, the role of negative blanks and positive controls is pivotal for data interpretation. Tools and setups should preferably not contain any plastics but rather non-polymer materials to avoid systematic contamination. Samples should be handled under a laminar ow hood and shielded against airborne contamination.
Standardized sampling procedures will make comparisons easier for all kinds of samples containing NPs. Hermsen et al. 67 provided a standardized protocol for the detection of ingested microplastics in biota comprising specic requirements for each step in a method, from sampling to detection. This protocol also has potential for the extraction of nanoplastics from biota. It is recommended to follow these requirements before setting up a method, especially for contamination control and the use of positive (spiked samples) and negative controls (blanks). Reports involving extraction of sub-micron and nanoplastics from real samples, such as sh products from markets or environmental air samples, are currently still scarce. 68 This section provides an overview of extraction methods that may potentially be used for the sub-micron and nano-sized plastics that are present in matrices relevant for the exposure routes ingestion, inhalation and dermal contact. Preconcentration and ltering methods that are needed for adequate collecting of sub-micron and nano sized plastics are summarized in Section 4.

Food and beverages
Any type of food and drink samples are relevant for MNP exposure through ingestion. Foods and drinks need to be treated under certain conditions and during sampling contamination with plastics must be avoided. Samples can be taken anywhere from markets, stores and the environment itself. Samples from living organisms have to be frozen at À21 C according to the International Council for the Exploration of the Sea, 69 or could also be preserved in xatives like formaldehyde or ethanol. 67 This is because a living organism will start decomposing aer 30 minutes, 70 hereaer the sample is not representative anymore.
There are some interesting studies that performed multiple digestion methods for microplastics extraction in sh. Dehaut et al. created a benchmark protocol for the extraction of microplastics (1-1000 mm) in sh species using 10% KOH solution. 71 The incubation was performed at 60 C for 24 hours. The most common polymers found were PE, polyester and rayon. This treatment was not efficient for gill samples. Karami et al. 72 reported an incubation time of 72 hours at a lower temperature (40 C) for the successful treatment of gill samples. It seems that treatments have to be altered for specic parts of the tissue which is not desired for a standardized protocol. Rist et al. 73 tested multiple treatments on exposed Daphnia pulex to MNPs such as alkaline digestion with NaOH, 30% H 2 O 2 treatment, acid digestion (nitric acid, HNO 3 ), 25% tetramethyl ammonium hydroxide (TMAH) and an enzymatic digestion with Proteinase K. Although, Daphnia pulex is not indicated as eatable food, the ndings in terms of sample treatment were interesting. Consequences of the treatments with NaOH, H 2 O 2 and HNO 3 were strong agglomeration of particles and loss of particle uorescence. The use of TMAH resulted in an incomplete dissolution of the tissues and Proteinase K only gave minor agglomeration of the particles, however, the particle uorescence signal was completely maintained. The protocol employed for enzymatic treatment (3 hours) was less time consuming compared to alkaline and acid digestion (few days). 74 Alkaline digestion with KOH was not tested in this protocol, however, alkaline digestion with NaOH resulted in a signicant loss of uorescence and more agglomeration of the particles. Therefore, it appears that enzymatic digestion is more suitable for the analysis of MNPs with uorescence detection. When using thermal fragmentation and spectroscopy techniques, digestion with KOH appears to be more suitable. 71 Although, enzymatic treatment was not tested in this protocol as it was assumed to be difficult to implement and present digestion efficacy issues.
It is clear that for a reliable an accurate result, the sample treatment used should not alter the MNPs present in the samples. For example, with the use of optical microscopy and dynamic light scattering it has been shown that aggressive methods such as acid, alkaline or H 2 O 2 treatment can cause aggregation of the particles. 68 The aggregation could be caused by the signicant change of the ionic strength. Furthermore, these treatments could also have negative effects on the uorescence signal of labeled MNPs (e.g. in toxicology experiments). Enzymatic digestion is milder than acid digestion, alkaline and H 2 O 2 treatment, and therefore, it is likely to be more suitable as treatment protocol before uorescence or light scattering analysis, as it has been demonstrated to cause no or less aggregation of the particles in food matrices. 68 In the study of Correia and Loeschner, the authors have successfully used an enzymatic digestion with Proteinase K for the characterization of spiked PS nanoplastics (600-60 nm) with asymmetrical ow-FFF-multi-angle light scattering (AF4-MALS) (method further elaborated in Section 5.1). The treatment of samples that are indicated as drinkable products such as drinking water is more straightforward as digestion procedures are not required. Murray andÖrmeci tested multiple treatments for nanoplastics (<400 nm) from water, where bench-scale ltration, centrifugation, and ballasted occulation were successfully used. 75 All samples relevant for the ingestion route will need preconcentration and ltering steps for the collection of sub-micron and nanosized plastics (Section 4).
The above-mentioned studies that are efficient for the extraction of microplastics from complex samples, can be used as a starting point in future studies on the extraction and analysis of sub-micron and nano-sized plastics.

Airborne samples
Air samples can be collected through a stand-alone pump, 76 vacuum cleaner, 77 lters installed indoors or outdoors, or with innovative technologies such as a breathing thermal manikin. 33 The limitation of each technique is related to the mesh sizes of lters used, which limits and impacts the collection of plastic samples especially difficult for the nanosized range ones. Therefore, additional preconcentration and ltering steps can be necessary (Section 4).
Compared to other sample types, air samples need to be treated extra carefully as they are mainly consisting of bers. For this purpose, the ltration system employed needs to be thoroughly cleaned between samples and the used lters need to be exposed to very high temperatures in order to remove the bers and other contaminants. 78 Airborne contamination by synthetic bers originating from atmospheric fall out, clothing or gear is probably the most difficult to avoid. To tackle this, blank samples and recovery studies using the proposed analytical method should be performed at all times.
There are some interesting studies that performed microplastic extraction from air samples and might be used in the future for sub-micron and nanosized plastics. For instance, the most common treatment reported involves density separation with ZnCl or NaI. 76,79 Prata et al. 78 used a different approach that involved an initial step to remove the organic matter by using 15% H 2 O 2 during an 8 day treatment prior to ltration over a washed glass ber lter and transferring to a NaI solution for density separation. Two procedural blanks were added and subjected to same treatment as the samples. The blanks contained 27 ber particles which were likely released from the cotton lab coat and paper towels. This method delivered 94.4% recovery of PS spiked in common textile bers and was applied to real indoor and outdoor samples. The method highlights the need for organic matter removal, providing a satisfactory recovery value. Nevertheless, the methods reported is very time consuming as the studies take several days. [8][9][10][11][12][13][14] The studies mentioned above provide an insight in treatment procedures for air samples containing microplastics, which need to be adjusted for adequate collecting and treatment of sub-micron and nanosized plastics.

Personal care products
Samples that are representative for the exposure of NPs through dermal contact are mostly personal care products such as facial scrubs. Although there are limited studies on this matrix, in our belief such a matrix is relevant because nanoplastics could potentially transverse the dermal barrier. 34 Hernandez et al. reported a study on the extraction and analysis of nanoplastics in facial scrubs. 35 In this study, a simple extraction was performed by adding 10 mL of reverse osmosis water, which reduced the viscosity of the samples. This was followed by multiple consecutive ltering steps to isolate the NPs. The analysis was performed by dynamic light scattering. Again, such a study is limited to the mesh size of the lter (100 nm) and therefore nanoparticles around and below 100 nm cannot be isolated. This technique was also used by Gopinath et al., which automatically was restricted by the same size limit. 36 Therefore, the isolation of NPs within this matrix also needs improvements in the preconcentration and ltering steps to isolate nanoplastics below 100 nm. Other potential sources of NP exposure by dermal contact could be sand or snow, which are matrices that have shown to contain NPs. 80,81 Although it has not been proven that such matrices can transfer NPs across the dermal barrier and they are less likely to cause human exposure compared to personal care products.

Preconcentration and filtering
For all the relevant matrices, preconcentration and ltering steps are needed aer extraction of sub-micron and nanosized plastics and to improve the limit of detection and limit of quantication (LOD and LOQ) of existing methods. The ltering techniques used within microplastic research are not applicable because of the high mesh size of the lters. Multiple studies showed that collecting nanoparticles can be rather tricky. 82 In fact, all type of samples should undergo preconcentration and ltering steps due to the extremely low amounts of sub-micron and nano sized plastic particles. Multiple techniques can be used for this purpose, such as membrane ltration, ultraltration, ultracentrifugation, continuous ow centrifugation and cloud point extraction [83][84][85][86] which can aid in the collection and enrichment of nanoparticles. 87,88 Other standard techniques such as freeze-drying and evaporation of the solvent can also be used for this purpose dependent on the type of sample.
However, techniques like membrane ltration and ultraltration are limited to the sizes of their inner channels, which still complicates the collection of NPs with sizes below 100 nm. Liu et al., reported the use of surface tension gradients 89 but this technique is limited by the amount of sample and it is also not able to distinguish the nanoplastics (<100 nm) from the submicron plastics (100-1000 nm). For this purpose, it is more likely that non-destructive separation techniques like FFF (Section 5.1) are more suitable to deliver fractions of particle size distributions (PSD).

Characterization and identification of MNPs
In this section the methods reported for the characterization and identication of submicron and nanosized plastics are reviewed and discussed. Fig. 2 shows an overview of the different strategies that can be used for the relevant matrices from sample treatment to analysis. Table 1 includes the relevant reports on the treatment and analysis of the sub-micron and nanosized plastics in such samples. In addition, it also contains reports that could be used in the future.

Size and shape characterization
MNPs can be very different in physical and chemical properties such as hydrodynamic radius, zeta potential, geometry and surface characteristics. These parameters have an inuence on their identication and quantitation, and therefore, detailed characterization is crucial. In this section multiple techniques that are able to characterize the size and shape of MNPs are discussed.
5.1.1 Microscopic techniques. These techniques provide information on the morphology of a sample including the geometry and surface characteristics. Optical microscopy is a technique used for single particle analysis of microplastic particles, where the stereomicroscope is oen reported in literature. However, it was reported that this technique is limited due to the difficulty of distinguishing microplastics from other small organic/inorganic debris particles which may lead to false positives and false negatives. 105 In addition, a stereomicroscope is not capable of visualizing nanoplastics due to restricted diffraction limits. On the other hand, microscopy is oen used in combination with uorescence for the tracking and translocation of MNPs. Forte et al. performed an in vitro study where human gastric cells were exposed to unmodied PS nanoparticles. 39 In this study, uorescence microscopy was used to track and localize dyed PS nanoparticles and the observed intensities were compared with the maximal intensity to calculate exact concentrations. Rist et al. performed an in vivo study where mussels were exposed to PS beads of 2 mm and 100 nm and uorescence was used for the same principle (see Table 1). 24 The quantication of MNPs with use of uorescence is discussed in Section 6.
Besides optical microscopy, electron microscopy (EM) is a very powerful technique for detailed information of MNPs. It can observe very small differences between the wavelengths of high energy electrons which illustrates its resolution, which makes it possible to image nanosized particles. 106 EM can be divided into the TEM and SEM techniques. Both techniques have high resolution and are mostly coupled to energy dispersive X-ray spectrometer (EDS) allowing visualization of the sample whilst simultaneously gaining qualitative information on the elemental composition. SEM-EDS is powerful combination for the characterization of MNPs, although there are some limitations. The technique is expensive and very time consuming with many sample preparations steps, hence limiting the number of samples that may be analyzed in a given timeframe. In addition, SEM cannot provide colored images which means that the colors of particles cannot be used as identiers. EDS can detect trace amounts of specic elements (including Na, Al, Ca etc.), and it may therefore determine the presence of additives by the chemical signature of these elements. The major limitation of EDS spectra is its inability to differentiate between elemental signatures originating from the polymer and elemental signatures originating from additives. 107 In the discussed study of Correia and Loeschner, SEM was also used for the conformation of the particle size and morphology of the spherical PS nanoplastics. 68 These techniques are very useful to visualize the presence of nanosized particles.
El Hadri et al. studied the degradation of primary plastics (PS, PE) into MNPs where plastics were mechanically degraded using a planetary ball mill. 108 In addition, environmentally degraded plastics collected from the beach were also tested. The   samples were fully characterized by TEM, DLS and AF4-MALS. It was shown that representative environmental samples can be obtained through mechanical degradation, however this should be tested on multiple and various environmental samples as well. The degradation process of plastics may differ in all kinds of environmental matrices. In the study reported by Gigault et al. and discussed above, TEM analysis was used to determine the particle size and shape of all the types of nanoplastics PE nanoparticles (<100 nm) and PS nanoparticles ($500 nm). The shapes observed were very heterogeneous, which encourages to perform studies on particles of different shapes and chemistries, and not exclusively on spherical PS particles as it is currently done in most of reported studies. In the research of Caputo et al., it was also stated that multiple complementary techniques are needed to provide accurate size and shape characterization of MNPs, which is also clear from the comparison of all the studies above. Until now, there is no technique that is capable of the complete characterization of both micro-and nanoplastics. 109 Therefore, combinations of the mentioned techniques are needed for the characterization of MNPs and to bridge the gap between submicron and nanosized plastics.

Light scattering techniques.
Light scattering is a detection method used in many studies for the characterization of MNPs. For example, dynamic light scattering (DLS) can deliver a broad PSD in the range of 1 nm to 3 mm. 110 Despite the broad range of particle sizes that can be measured, a mixture of particle sizes may cause problems, as the technique can only measure average hydrodynamic sizes. As a consequence, the measured radii can be skewed towards higher sizes. This is because larger particles will scatter with more intensity than smaller ones, and therefore, the signals of large particles will hinder the signals of the small particles which will be overlooked. Another problem might be caused by contamination of dust bers or formed aggregates from the sample matrix. This could be a difficult issue for the relevant food, beverages, inhaled particles and personal care products, and therefore, strict measures should be taken for sample preparation when using this detection technique (Section 3). Another approach involves the use of (static) light scattering is multi-angle light scattering (MALS). This technique measures the scattered light from the sample by different angles and can determine the molecular weight and the size distribution (radii of gyration or root mean square radii) of molecules in solution. MALS is commonly used as online detector for size-based separation techniques such as size exclusion chromatography (SEC) or asymmetrical ow-eld ow fractionation (AF4). The implementation of AF4-MALS for nanoplastics research is discussed in the next section.
Nanoparticle tracking analysis (NTA) is a light scattering technique complementary to DLS. Both techniques calculate the hydrodynamic size of particles based on the measured Brownian motion. NTA uses a microscope and a high-sensitivity video-camera which makes it possible to visualize (video image) and record every particle. Therefore, it can determine the hydrodynamic size of each individual particle instead of average size data as generated by DLS. 111 On the other hand, very PS, 100-1000 nm  . FFF is one of the emerging techniques for the separation and size characterization of nanoplastics. The power of FFF is the broad range of particles that can be covered (1-1000 nm) and because it involves minimal to no shear stress and it is non-destructive. 118 Because of the minimal to no shear stress involved, the agglomeration behavior of nanoparticles can be studied using this technique. 119 The most common variant is asymmetrical ow-FFF (AF4), which is typically coupled to multiple detectors like UV-vis, refractive index, uorescence, MALS and DLS. 120 These detection techniques in combination with AF4 provide information on concentration, number of particles, particle size distributions and molar masses for the characterization of MNPs.
Monikh et al. reported the use of AF4-MALS to successfully fractionate and characterize PS nanoparticles (60,200, 300, 600 nm) spiked in eggshells at 100 mg L À1 (see Table 1). 90 The developed method had a sufficient recovery (>60%) for nanoplastics and could be able to deliver suitable fractions for further identication. However, the results might be different for various types and non-spherical nanoplastics that are weathered in the environment. Correia and Loeschner used AF4-MALS for the analysis of sh tissue samples which were spiked with 100 nm PS particles (at a nal concentration of 5.2 mg mL À1 ) (see Table 1). 68 As control, the authors have analyzed non-spiked sh samples. The authors reported the overlayed fractograms obtained by analyzing dye red aqueous uorescent spherical polystyrene nanoparticles (FIPSNP) in ultrapure water and sh. It was observed that the particles extracted from the sh sample show minor deviation with the peak obtained from the PS standards. The results show the capability of AF4 to separate PS nanoparticles from such a complex matrix. Besides the study involving PS, the authors reported that aer optimizing the carrier liquid composition for the AF4 experiments, it was also possible to analyze PE nanoparticles. For the PS nanoparticles, 0.47 mM NaHCO 3 (pH 7.7-7.9) was used as the carrier liquid, while FL-70 concentrate was used for PE nanoparticles. However, it was not possible to detect the PE nanoparticles when spiked (10 mg mL À1 ) to sh samples. The authors attribute this to an elevated light scattering background signal from the organic sh residues in the AF4 running conditions. This could mean that a method developed for a certain type of polymer based nanoplastic may not be applicable to other types and this should be systematically investigated. This can make it complicated to standardize these protocols for multiple types of plastics, unless suitable studies are performed with a wide variety of MNPs of different chemistries. This technique is less likely to be useful when microplastics of sizes above 1 mm are present in the sample as the elution mode will be changed from normal to steric and the separation is jeopardized. In this case, the large particles (of micrometer range) undergo stronger forces from the laminar ow 87 and will elute faster than the smaller particles. To prevent this, a ltration step at the inversion point is needed to exclude the larger particles which can be studied by complementary techniques. Methods such as ultra-ltration, ultracentrifugation/centrifugation, can be used. 87 For example, Correia and Loeschner used centrifugation before AF4-MALS analysis of nanoplastics in sh. 68 Gigault et al. reported the use of AF4-MALS for the characterization of nanoplastics (PS particles, 1 nm to 800 nm) in sh samples. 121 It was found that the selectivity increased signicantly when the size range was divided in subpopulations. Therefore, the elution prole was tuned into four different subfractions (1-100 nm, 100-200 nm, 200-450 nm, 450-800 nm). Additionally, in these subfraction methods, it was also found that constant cross-ow rates (0.1 and 0.3 mL min À1 ) enhanced the fractionation power compared to a programmed cross ow rate. The developed method and additional four subfraction methods combined, may be used to study all the submicron populations in sh samples. This study demonstrates the advantages of AF4 coupled to MALS, and the developed methods were also used in a more recent follow up study where the degradation of microplastics to nanoplastics was studied. 108 5.2 Chemical identication 5.2.1 Spectroscopic techniques. FTIR/m-FTIR, Raman/m-Raman spectroscopy are useful techniques for the identication of microplastics. However, the spatial resolution of FTIR is not sufficient to identify particles below 50 mm (ref. 122) and difficulties may arise from environmental matrix effects, unless proper sample preparation is used. 123,124 Liu et al. reported the use of m-FTIR to reveal the presence of many kinds of microplastics above 10 mm in air samples. 125 Unfortunately, even when combined with microscopic techniques, a spatial resolution below 10 mm cannot be reached. 126,127 Therefore, the iden-tication of nanoplastics by FTIR is not currently possible. Compared to FTIR, Raman has a better spatial resolution due to the shorter laser wavelengths that can be utilized, therefore particles down to 10 mm can be analyzed. Additionally, in the case of m-Raman particles down to 1 mm can be analyzed. 124 In addition, Raman measurements have less interference from water and are not dependent on sample thickness. 128 On the other hand, sample clean-up is essential for Raman measurements to increase the signal to noise ratio and eliminate potential uorescence interferences from sample tissue or other compounds in the sample. UV degradation can also alter the Raman spectra, for example the intensity loss of the specic C-Cl bond in poly vinyl chloride (PVC). 129 Alternatively, m-Raman can be coupled to an Atomic Force Microscopy Based Tip-Enhanced Raman Spectroscopy (AFM-TERS) system which can deliver a spatial resolution of 10 nm. 130 This may have potential for the identication of nanoplastics, but has not been yet reported. Recently, Sobhani et al. 91 successfully analyzed NPs down to 100 nm by Raman imaging (see Table 1), where imaging particles can be visualized and identied. The produced method was also tested on real paint-polishing dust samples. These results are encouraging, but there are several limitations. The main limitation arises when the nanoplastic size is smaller than that of the laser spot. From the Raman image, the size of the imaged nanoplastic is actually determined and limited by the collected Raman signal, the stage-stepping resolution/pixel size and the laser spot size, rather than by the nanoplastic size itself. Therefore, the image resolution needs to be increased which is limited by the diffraction limit of the laser spot. This method was optimized in two recent follow-up studies. Firstly, the resolution was increased by decreasing the mapping pixel size in order to produce a high-resolution image. 131 This made it possible to categorize imaged NPs by size groups via their Raman intensity. Secondly, multiple algorithms such as logic-OR, logic-AND and logic-SUBSTRACT were added and combined to prevent false positives and increase the mapping certainty for NP imaging. 132 Until now, this approach is the most promising for adequate identication and visualization of sub-micron and nanosized plastics.
The coupling between atomic force microscopy (AFM) and IR spectroscopy allows to characterize nanoparticles and may have potential for nanoplastics. 133 However, this coupling is not easy because AFM has a limited sample size which can cause problems for large microplastics that may be present in the sample. Additionally, it can only detect at the surface area of the sample, which means that a sufficient sample preparation is needed to discover smaller sized particles that could be present beyond the surface. AFM and its hyphenation to FTIR or Raman should be further explored for the characterization of MNPs. For instance, AFM-IR has been used for the identication of various types of nanoparticles (polylactic acid, silver & gold) already and has potential for quantication purposes. 134,135 Merzel et al. showed the applicability of AFM-IR recently with the characterization of PS nanoplastics (beads of 1000 nm) in mussel siphons. 136 However, improvements can be made by optimizing the sample preparation (see Section 3) to investigate nanoplastics beyond the surface.
In terms of imaging, hyperspectral imaging can turn a dark-eld optical microscope into a powerful chemical characterization tool. 137 This technique has been used for the identication of various nanoparticles around sizes of 5-100 nm. It has the major advantage of imaging particles in unxed wet samples, which means no sample treatment is needed. However, the major limitation of this technique is the interpretation of complex spectra, therefore, instrumental advances are required such as deconvolution soware. Recently, this technique has also been used for successful identication of sub-micron sized plastics (PS, 400-1000 nm) in Caenorhabditis elegans. 138 5.2.2 Mass spectrometry (MS). A different approach with respect to microscopy and spectroscopy are mass spectrometrybased methods. MS is a powerful technique for the identication of MNPs based on their m/z ratio. Techniques such as pyrolysis gas chromatography-MS (Pyr-GC-MS), 139,140 thermal gravimetry/desorption gas chromatography MS (TED-GC-MS) or thermal desorption-proton transfer reaction-MS (TD-PTR-MS) 81,141,142 and matrix-assisted laser desorption/ionization-MS (MALDI-MS) can be used. These techniques have the advantage that samples can be analyzed in bulk, which is a solution for the lack of sensitivity posed in single-particle analysis. In addition, MS is not limited by the low particle sizes of NPs. However, the main disadvantage of mass spectrometry techniques for NP analysis is the fact that information on particle sizes cannot be obtained. Additionally, the concentration of NPs in environmental samples needs to be sufficient as every MS approach has a certain detection limit (down to ppm or ppt). 143 To overcome the limited knowledge on particle sizes, size-based separation techniques can be used prior the MS analysis to obtain a complete picture. In this context, chromatographic techniques such as size exclusion chromatography (SEC) 144 and hydrodynamic chromatography (HDC) 145 have been reported for the separation of engineered nanoparticles. Hence, they may be also suitable for the separation of nanoplastics. For instance, Pirok et al. combined HDC and SEC in a comprehensive 2D-LC system, where the combined two-dimensional distribution of particle sizes and molecular sizes of PS and polyacrylate particles was obtained successfully. 146 Such methods could contribute to the characterization of MNPs and may be combined with MS based approaches as they are nondestructive. The limitation of detection limits is not that easy to overcome, besides making the analysis as sensitive as possible by sufficient sample preparation and preconcentration.
As an example, Lin et al. recently reported a method where thermal fragmentation in combination with MALDI-MS was used (see Table 1). 23 Thermal fragmentation decomposes the sample and subsequently the MNPs are identied by ngerprint peaks in both low and high mass regions of the MALDI-MS spectra. Environmental samples are composed of heterogenous MNPs with different molecular weights which causes many variations in peak intensities. Therefore, the low MS responses of the ngerprint peaks need greater intensity considering the low concentration of MNPs in environmental samples. This was done by a thermal fragmentation step which enhanced the intensities of ngerprint peaks and made quan-tication possible.
The combination of pyrolysis-GC with mass spectrometry is promising and still in a developmental phase. 147 Other MS based approaches such as inductively coupled plasma-mass spectrometry operated in single-particle mode (sp-ICP-MS) 80,96 are interesting due to their identication and quantication (number of particles) qualities. In the approach of Jiménez-Lamana et al., 96 conjugated nanoplastics with Au-nanoparticles were used, which provides a very sensitive analysis. This technique was also widely discussed in the review of Velimirovic et al. 143 The labelling of NPs with metal probes has also been studied by Marigliano et al., again Au-nanoparticles were used and showed most efficiency for NP identication and quantication. 102 Imaging with TOF-secondary ion mass spectrometry could also be used for the chemical identication of NPs, as the spatial resolution (>100 nm) is suitable. 148 Although it has limitations in long analysis times and only a small area can be covered by each analysis, while sufficient analysis of multiple spots are needed for representative data. This technique has not been used for NPs yet. 13 It is clear that MS based approaches are a very powerful tool that can be used for NP analysis, but most likely not as a stand-alone technique. On-and offline combinations remain needed to solve this broad range of research questions along NP analysis.
Additionally, some MS based methods also reported quan-tication of nanoplastics, which are discussed in Section 6.

Quantification of MNPs
The quantication of MNPs in different sample types is not an easy task, as shown by the lack of quantication methods present in literature. The adequate quantitation of particle number, mass, volume and concentration of MNPs is still lacking. Fluorescence is a technique that is capable of the identication and quantication of MNPs based on staining. This technique was mostly reported in toxicology studies where organisms are exposed to labelled PS nanoparticles with a known concentration to investigate the translocation of the particles in organisms and to hypothesize cytotoxicity and health effects, for example in the study of Pitt et al. 149 In a recent study of Molenaar et al., uorescence video microscopy was used in combination with Nile red staining and single particle tracking (SPT). 97 The developed method was able to detect 45 nm sized nanoparticles and concentrations of 2 Â 10 6 nanoparticles per mL were reported. Despite of the successful quantication, only spherical PS particles were used and the method is not tested on environmental samples yet, which could cause difficulties due to matrix effects.
Analytical techniques in combination with mass spectrometry (GC-MS) are used for the identication of MNPs, but are also be used for quantication. Dümichen et al. reported multiple studies on the identication and quantication of microplastics with thermal fragmentation. 141,142 In 2015, the rst paper published using TED-GC-MS demonstrated its capability of identifying and quantifying PE standards, but it has not been tested on real samples. 142 In 2019, an optimized method was published with an increased sample throughput and reproducible automated fractioned collection of decomposition products. 141 Quantication was achieved by the linear regression curves of PS, PP and PE standards which showed excellent linearities and an internal standard solution was used to compensate for instrumental errors. This method reached lower limits of quantication by a factor 10 (LOQ 0.395 mg) compared to the method from 2015. 142 The MS was used in a full scan mode and the method is expected to reach even lower limits in the single ion monitoring (SIM) mode. However, this method is not validated and tested on real samples yet, but it does shows promising results for routine analysis. Another successful quantication method with Pyr-GC-MS was developed by Sullivan et al. 95 Micro-and nanoplastics (PP, PS, PVC, 100 mm to 100 nm) were quantied below 50 mg L À1 (LOD) within water samples. Additionally, Materić et al. used TD-PTR-MS for the quantication of PET (<200 nm, 4.6-23.6 ng mL À1 ) nanoplastics in snow samples 81 (see Table 1).
The limitation of a stand-alone pyrolyis or thermal desorption-GC-MS method is the lack of information on particle sizes within environmental samples. This could be tackled, by using AF4-MALS in combination with offline Pyr(or TD)-GC-MS, as it can rst separate MNP fractions according to their size and then chemical identication and quantitation can be further performed. This can be a very powerful combination for the detailed characterization, identication and quantication of MNPs, Fig. 3 illustrates this combination of techniques.
As regards AF4-MALS, Battistini et al. successfully validated an AF4-MALS method for the identication and quantication of nanosized PS particles (20-200 nm) at a LOD of 15-33 mg mL À1 . 94 The same accounts for the protocol of Bocca et al. that was also able to quantify PS NPs (20-200 nm) at a LOD of 50 mg mL À1 . 101 However, both methods are limited to nanoplastics that present UV absorbance.

Conclusions
It is clear that NPs are a serious environmental issue and a potential risk to human and ecosystem health. Multiple studies investigated the exposure of NPs to animals and negative health effects were found. Additionally, negative health effects were also found within in vitro studies using human cell lines. However, the exposure dose of NPs used in current studies are generally unrealistic as environmental samples are likely to contain low or trace amounts. The observed health effects in toxicity studies are mostly negative. Nevertheless, these studies show the intrinsic hazardous properties of NPs and the need of decreasing plastic debris around the world. The production of commercially available nanoplastic standards with variating shapes needs to be expanded to support the development of analytical studies. Contrary to MPs, NPs are hardly measured in real environmental matrices that are relevant for human health such as drinking water, sh, air and personal care products. More studies are thus urgently required to study sub-micron and nanosized plastics in relevant matrices that are correlated with the main exposure routes: ingestion, inhalation and dermal contact. Recently, some great advances were made in the successful identication of NPs by mostly imaging techniques such as (Raman imaging and dark-eld/ hyperspectral microscopy). Advances in MS based methods are promising in terms of identication and quantication of NPs, but combinations with other techniques are necessary to characterize them as well (see Future perspectives). More research and improvements are especially necessary in the sampling and sample treatment of NPs in all relevant matrices. This is extremely important to support the progress in analytical techniques for characterization, identication and quantication of sub-micron and nanosized plastics. Further development, harmonization and in time also standardization of quantitation protocols is needed to deliver realistic exposure doses which will lead to accurate risk assessment of NPs on human health.

Future perspectives
The characterization and quantitation of sub-micron-and nanosized plastics remains a challenge. Therefore, new combinations of multiple analytical techniques are needed, to make progress and provide more useful data. Combining suitable techniques seems the only way to fully characterize, identify and quantify NPs, as no stand-alone technique is capable of doing all. The latest developments also showed that different combinations can be successful and make signicant progress (see Table 2 for an overview). For example, AFM-m-FTIR or AFMm-Raman spectroscopy are combinations of techniques that could be promising for NP identication. Currently, FTIR and Raman spectroscopy can only identify microplastics, because they are limited due to their spatial resolution. Coupling AFM to FTIR or Raman may help to overcome this limitation and identify nanoplastics successfully. This combination is expected to identify nanoplastics down to sizes of 10 nm as the spatial resolution will be lowered signicantly. On the other hand, recent advances in Raman imaging and dark-eld/hyperspectral microscopy are already very promising as they have shown condent identication of NPs. It is expected to see more studies in the future that utilizes these techniques. For the characterization of nanosized-plastics (10-1000 nm) AF4-MALS has great potential and new studies keep appearing. 150 It is capable to deliver the molecular weight, particle numbers, concentration by optional UV detection and particle size distribution. It has the advantage that this technique is nondestructive and it can collect fractions, which can be used for additional analysis with for example Pyr-GC-MS based approaches for identication and quantitation. AF4-MALS can also be combined with spectroscopic and microscopic techniques such as confocal Raman and SEM. This was shown recently by Valsesia et al., where characterization (SEM/AF4-MALS), identication (Raman) and quantication (particle counting soware and UV absorbance) was achieved. 151 Clustered particles on a chip were used to make the NPs detectable with confocal Raman and this study was successfully applied on C. Robusta. This approach has great potential, but it has to be mentioned that it was still only applied on PSNPs. Recently there was also an extensive combination made of DLS, TEM-XPS, FTIR, AFM-IR and Pyr-GC-MS to characterize and identify NPs (PO, PS, PVC and PA, 58-255 nm) in tap water. 104 Even plastic additives such as P(E-MMA), MBS and PBMA were found which is a relevant capability of this method regarding the enhanced toxicity of NP-additive complexes. The indicated example studies provided different combinations of techniques that show the most potential to analyze NPs to date.
Other MS based approaches such as sp-ICP-MS can also provide very sensitive and specic identication and quantication (number of particles) as well. 80,96,102 Capillary electrophoresis (CE) is a separation technique which separates analytes based on their charge to hydrodynamic radius ratio and could be suitable for the analysis of MNPs. CE has been reported for the analysis of different nanoparticles such as gold and fullerene nanoparticles. [152][153][154] In addition, CE-MS was also used for the characterization of nanomaterial in protein corona, where even PS microplastics were found as contamination. 155 Unfortunately, there is no CE study available yet that is focused on the analysis of MNPs to the best of our knowledge. There are still options and combinations of techniques that can be explored and may contribute to the analysis of sub-micron-or nanosized plastics. Besides making new combinations, specic sample preparations can simplify the analysis of NPs. For example, Li et al. used a unique extraction protocol with alkaline digestion and cellulose precipitation to characterize, identify and quantify NPs (PS and PMMA, 50-500 nm) in cucumber plants with a combination of Pyr-GC-MS, SEM and ICP-MS. 103 Such approaches are desired to move forward in the eld. In addition, newly developed or adapted data processing soware is also necessary to strengthen the data analysis, which potential was shown in some recent studies. For example, Primpke et al. provided a new soware tool (siMPle) for the systematic identication of microplastics within spectroscopic analysis. 156 The future developments need to improve the capabilities of the current methodologies for the adequate analysis of these extremely challenging nanoplastic particles. This review focused explicitly on external exposure matrices that can contain NPs. This is because, these sources are more suitable for routine monitoring compared to internal exposure matrices (e.g. NP concentration in blood). However, the described strategies and techniques could also be applied for internal exposure matrices. For example, the research of Gray et al. describes the extraction and analysis of Ag and Au nanoparticles in biological tissues. 157 A similar strategy with Proteinase K digestion and sp-ICP-MS analysis was used that also has potential for NPs as indicated in Sections 3.1 and 5.2. Internal exposure matrices will become more interesting within the future, as the concentration of NPs in the environment and exposure to organisms is still likely to increase in the upcoming years.

Conflicts of interest
There are no conicts to declare.