Toward a Framework for Environmental Fate and Exposure Assessment of Polymers

Development of risk‐assessment methodologies for polymers is an emerging regulatory priority to prevent negative environmental impacts; however, the diversity and complexity of polymers require adaptation of existing environmental risk‐assessment approaches. The present review discusses the challenges and opportunities for the fate and exposure assessment of polymers in the context of regulatory environmental risk assessment of chemicals. The review discusses the applicability and adequacy for polymers of existing fate parameters used for nonpolymeric compounds and proposes additional parameters that could inform the fate of polymers. The significance of these parameters in various stages of an exposure‐assessment framework is highlighted, with classification of polymers as solid or dissolved being key for identification of those parameters most relevant to environmental fate. Considerations to address the key limitations and knowledge gaps are then identified and discussed, specifically the complexity of polymer identification, with the need for characterization of the most significant parameters for polymer grouping and prioritization; the complexity of polymer degradation in the environment, with the need to incorporate the fate and hazards of degradation products into risk assessment; the requirement for development and standardization of analytical methods for characterization of polymer fate properties and degradation products; and the need to develop exposure modeling approaches for polymers. Environ Toxicol Chem 2022;41:515–540. © 2021 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.


INTRODUCTION
The prevalence and persistence of polymers in the environment have resulted in heightened concern in public, scientific, and regulatory communities. Polymers have previously been subjected to reduced regulatory requirements compared to low-molecular weight (LMW) chemicals, for example, under the European Union regulation Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH;European Commission, 2006); and there increasingly have been calls for regulation and efforts to develop risk-assessment approaches for polymers (European Centre for Ecotoxicology and Toxicology of Chemicals [ECETOC], 2019). In particular, the potential risks of plastics and microplastics have been the focus of a vast amount of research because of their widespread release into, and persistence in, the environment Derraik, 2002;Ivleva et al., 2017;Koelmans et al., 2017;Thompson et al., 2009), with a number of riskassessment strategies being suggested for microplastics (Gouin et al., 2019;Hüffer et al., 2017;Syberg et al., 2015). However, microplastics represent a single group of polymeric material, and in contrast, the environmental impacts of other groups including water-soluble polymers have been given considerably less attention (see Arp & Knutsen, 2020;Xiong, Loss, et al., 2018). Water-soluble materials were excluded from the definition of microplastics in the recent European Chemicals Agency (ECHA) report for restriction of intentionally added microplastics (ECHA, 2019), which could lead to the potential environmental impacts of water-soluble polymers being overlooked. This is despite the fact that water-soluble polymers have many applications, including in agriculture, wastewater treatment, consumer products, and detergents (Arp & Knutsen, 2020); and it is inevitable that they will be These fate parameters include basic physicochemical properties such as water solubility, partition coefficients, bioconcentration and bioaccumulation factors (BCFs and BAFs, respectively), and biotic and abiotic degradation rates, with standard Organisation for Economic Co-operation and Development (OECD) test methods for their measurement.
Because experimental fate and property data are sometimes only available for a small proportion of chemical substances in use, structure-activity relationships and quantitative structure-activity relationships (QSARs) are often utilized where the data are insufficient or unavailable. Models of QSARs such as those in EPI Suite have been established for prediction of physicochemical properties (e.g., water solubility, vapor pressure, Henry's law constant, and octanol-water partition coefficient [K OW ]) and environmental fate parameters (e.g., degradation half-lives and sorption coefficients) of chemicals (US Environmental Protection Agency [USEPA], 2012).
Both experimental and predicted property and fate parameters can ultimately be used as input parameters in exposure models. A multitude of exposure models exist for chemical compounds including very simple lower-tier models through complex higher-tier models. Examples include the OECD Tool, a consensus model for ranking overall persistence and long-range transport potential of organic chemicals (Wegmann et al., 2009); the European Union system for the evaluation of substances (Vermeire et al., 1997), which may be used to quantify exposure and risk of chemicals (e.g., under REACH); and the Forum for the Co-Ordination of Pesticide Fate Models and Their Use (FOCUS) models for estimating concentrations of plant protection products (FOCUS Working Group on Surface Water Scenarios, 2001). Lower-tier models are often very simplistic and provide "worst-case" concentrations in the environment, often ignoring dissipation processes. Higher-tier exposure models typically may rely on a large number of input parameters including partition coefficients and degradation half-lives in different media and aim to characterize transport and transformation of a chemical before its ultimate degradation, uptake, or sequestration (Di Guardo et al., 2018).
These different methods for measuring or estimating the properties and fate of molecules and for modeling exposure concentrations may, however, not be appropriate for polymeric substances. In the following sections, we therefore discuss why polymers are different and assess the validity of these existing methods for exposure assessment of polymers, before proposing strategies that could be used for polymer exposure assessment.

WHAT ARE POLYMERS, AND WHY DO THEY REQUIRE A DIFFERENT APPROACH?
monomer unit, and a distribution of molecular weights (MWs) with less than a simple weight majority of molecules of the same MW, where differences in MW are primarily due to differences in the number of monomer units (OECD, 1991). Polymer MW is therefore typically defined in terms of number and weight average molecular weight (MW N and MW W , respectively) and molecular weight distribution (MWD). Polymers have widespread usage and are released to the environment in both solid form (e.g., plastics [Kawecki & Nowack, 2019]) and dissolved form (e.g., from water treatment and agriculture [Arp & Knutsen, 2020]).
There are a number of unique characteristics of polymers that require additional consideration in exposure assessment compared to LMW chemicals. Polymers often comprise multiple components (including residual monomer, oligomers, polymer chains of varying MW, and chemical additives) and are poorly defined compared to most simple LMW chemical compounds. For example, for polymers (alcohol ethoxylates, alcohol ethoxysulfates, and polycarboxylates) incorporated in the Human & Environmental Risk Assessment on Ingredients of European Household Cleaning Products (HERA), in addition to MW distribution for each MW N , polymers of a wide range of MW N were in use, with different fate properties (such as degradation and sorption) requiring separate incorporation or consideration in risk assessment (HERA, 2004(HERA, , 2009(HERA, , 2014a(HERA, , 2014b. Identification of polymers is complex; names and Chemical Abstracts Service (CAS) numbers (which are based on incorporated monomers) are insufficient to differentiate polymers because the same name and CAS number may apply to two polymers with vastly different properties. In addition, compared with LMW chemicals, polymers may form a more complex mixture of products when they transform in the environment, including cross-linked polymer chains, micro-and nanoscale particles, oligomers, and LMW chemical compounds (see Lambert et al., 2013aLambert et al., , 2013bSaad et al., 2010;Ter Halle et al., 2016;Weinstein et al., 2016). It has been highlighted that there has been inconsistency in the size classes used for plastic debris; for the purposes of the present review, the terms "macropolymer," "mesopolymer," "micropolymer," and "nanopolymer" will refer to polymeric substances with size ranges of ≥10 mm, 1 to <10 mm, 1 to <1000 μm, and 1 to <1000 nm, respectively, according to the recommendations given by Hartmann et al. (2019) for plastic debris.
It is likely that for lower-tier, worst-case ERA scenarios, existing exposure-assessment methods will be generally sufficient for polymers, with only information on usage/production volumes and emissions estimates being necessary, although the availability of these data for many current-use polymers may limit characterization of exposure (Duis et al., 2021). However, for more complex, higher-tier environmental exposureassessment studies which incorporate data on fate behavior, additional considerations are likely to be necessary for polymers. Only a limited number of environmental exposure and risk assessments have been performed for polymers to date, including for polyethoxylated surfactants, polycarboxylates, and polyquaterniums (see Cumming, 2008;DeLeo et al., 2020;HERA, 2004HERA, , 2009HERA, , 2014aHERA, , 2014b, with detailed information on polymer characteristics often being limited (Duis et al., 2021). For example, the assessment of polyquaterniums conducted by Cumming (2008) was limited by insufficient information to estimate the mixture of polyquaterniums present or their range of charge densities and MWs.
A primary concern for higher-tier environmental exposure assessment is the establishment of key parameters to measure the behavior and fate processes of polymers in the environment. In the present review, a detailed analysis of the relevance and applicability of fate parameters to polymers has been performed, exemplifying the need for additional considerations in higher-tier exposure assessment of polymers and application of fate parameters in exposure modeling. The applicability of established fate parameters for LMW chemical compounds to polymers is first discussed and summarized in Table 1. We suggest that both homo-and copolymers can be grouped either as solid polymers (including bulk macroscopic solids and particles) or as dissolved polymers (defined in the present review to cover polymers which are dissolved in solution, such as water-soluble polymers in an aqueous environment and polymers which exist in the liquid state [which may be water-insoluble]), based on the applicability of both established fate parameters and suggested polymer-specific parameters. This grouping underpins the following discussion.

Basic physicochemical properties
Boiling points are typically not relevant for most polymers because, by definition, polymers exist as macromolecules with HMWs and typically decompose before boiling (see Schupp et al., 2018). Similarly, vapor pressure (P) will generally remain low for dissolved polymers because of their HMW. While vapor pressure can be measured for some liquid polymers, it is likely that it is LMW and oligomeric components that contribute most to P (Schupp et al., 2018); P may therefore be a relevant parameter for some LMW polymers and substances containing high levels of oligomers or residual monomer (Risk & Policy Analysts, GnoSys, & Milieu, 2012).
Conversely, melting points (T m ) are applicable to both LMW and HMW polymers. In the context of polymers, the T m refers to the transition between crystalline and amorphous states and applies only to semicrystalline polymers (Alsleben & Schick, 1994). The physical properties of the polymer matrix in a solid polymer may play an important role in environmental fate and effects. For example, LMW constituents may leach more readily from a flexible polymer compared to a rigid one (Hoekstra et al., 2015), and amorphous polymers or polymer regions may undergo preferential (bio)degradation before those that are crystalline-structured (Fukushima et al., 2013;Khatiwala et al., 2008).
Reactive functional groups (RFGs) also influence environmental fate, and in contrast to LMW chemical compounds, the functional group equivalent weight is important for polymers because it describes the relative proportion of RFGs within the polymer (ECETOC, 2019). Anionic and cationic polymers are analogous to acidic and basic polymers, respectively (see Guiney et al., 1998;Hennecke et al., 2018;Ostolska & Wiśniewska, 2014), and measurement of their dissociation constants can enable prediction of their charge or charge distribution at environmental pH (see Schupp et al., 2018). Ionic polymers have multiple applications, including in household products (Pecquet et al., 2019) and wastewater treatment (e.g., Shen et al., 2013); and there has been concern over the ecological hazard potential of cationic polymers (see Cumming et al., 2008;Costa et al., 2014;Goodrich et al., 1991;USEPA, 1997). Charge also influences environmental fate processes such as sorption (Blachier et al., 2009;Galvão et al., 2007). Surface tension (γ) is relevant for dissolved and colloidal polymers with surfactant properties, with surfactant behavior being recognized as significant for environmental fate and effects (see Jardak et al., 2016).

Partition coefficients
Parameters such as the soil-water and soil organic carbon-water partition coefficients (K d and K OC , respectively [Kookana et al., 2014]) are used to assess the partitioning of chemicals between soil/sediment/sludge and water (Amiard & Amiard-Triquet, 2015) and are useful in predicting the concentrations of a chemical in these environmental compartments. Although terrestrial environments and soils are important receiving compartments of both solid and dissolved polymers (because of application of, e.g., sludge, mulch, agrochemicals, and soil conditioners [Arp & Knutsen, 2020;Felsot et al., 2011;Horton et al., 2017]), the use of K d and K OC in the context of bulk solid polymers is not appropriate. As has been highlighted in the literature, colloidal dispersions do not reach thermodynamic equilibrium. Instead, processes such as sorption to soils are kinetically controlled and dependent on time, concentration, and system conditions (Kookana et al., 2014;Praetorius et al., 2014). It therefore follows that application of K OC and K d , as well as other commonly used equilibrium-based partition coefficients such as K OW , to partitioning of nano-sized polymer particles, as well as micro-scale particles and larger solids which can undergo sedimentation, is not appropriate and may lead to erroneous results (Praetorius et al., 2014). Such equilibrium-based partitioning parameters should only be applied to polymer molecules, not bulk solids (see Min et al., 2020).
These parameters may therefore be applied to dissolved polymers because these will follow equilibrium partitioning behavior. Equilibrium partition coefficients have been previously applied to polymer macromolecules (Gorbunov & Skvortsov, 1995;Lazzara & Deen, 2004;Tong & Anderson, 1996;White & Deen, 2000), usually in the context of partitioning between a gel and solution but also in an environmental context, albeit rarely (Cumming et al., 2011a(Cumming et al., , 2011b. However, use of K OW to indicate potential for bioaccumulation may be insufficient for HMW polymers because of uptake by nonpartitioning processes (see section Bioconcentration and Bioaccumulation). Given that polymer molecules in solution can also exist in the nano-size range (Armstrong et al., 2004;Arp & Knutsen, 2020;Xiong, Miller, et al., 2018), it may also be relevant to test and verify the applicability domain of equilibrium-based parameters to such polymers. Furthermore, as has been discussed by Cumming (2008) and Duis et al. (2021) in the context of the environmental fate of polyquaterniums, polyethylene glycols, and acrylic acid polymers, conformation of polymer chains is likely to play a role in sorption and desorption, which will affect partitioning to soils and sediment in the environment for dissolved polymers.

Bioconcentration and bioaccumulation
Often, BCFs and BAFs are used in fate and hazard assessments of chemicals (see Berrojalbiz et al., 2009;Castro et al., 2019;J.-P. Wu et al., 2011) to characterize uptake and accumulation into organisms in the environment. While the BCF accounts for uptake of a chemical substance only via dermal and respiratory absorption, the BAF accounts for additional uptake via ingestion (Arnot & Gobas, 2006;Mackay et al., 2013). Because the concept of BCF assumes passive diffusion, it is known to be inapplicable to nanoparticles (Kookana et al., 2014;Kühnel & Nickel, 2014) because equilibrium partitioning does not apply and active processes such as endocytosis play a significant role in nanoparticle uptake because of their size (Fröhlich, 2012;Kookana et al., 2014;Utembe et al., 2018). This is also the case for larger solids such as microplastics (von Moos et al., 2012). The role of active processes also means that BCF and BAF may be dependent on exposure concentration, and thus differences between substances cannot only be attributed to differences in bioaccumulation (Utembe et al., 2018). It has been highlighted that parameters such as uptake and internalization rates and attachment efficiencies (α) should be identified and developed for nanoparticle bioaccumulation (Kühnel & Nickel, 2014;Praetorius et al., 2014). Test methods based on concentrations and rate constants may need to be modified and should be interpreted such that they reflect uptake/depuration rates rather than BCFs (Kookana et al., 2014); uptake and depuration rate constants, as well as assimilation efficiency, have been applied to nanoparticles previously (Dai et al., 2015;Zhao & Wang, 2010).
Knowledge from medicinal chemistry shows that endocytosis also plays a role in cell uptake of polymer molecules (Apostolovic et al., 2011) because of their large size, suggesting that BCF and BAF are also likely to be insufficient to describe uptake of dissolved polymers. Polymer and particle properties that influence cell membrane interactions and uptake have been identified from medicinal applications of polymers and nanoparticles in drug delivery and include size, shape, composition, hydrophobicity, surface charge, and distribution of functional groups (Fröhlich, 2012;Liechty et al., 2010). These properties may therefore be important for characterization of biological fate processes of both solid and dissolved polymers.

Abiotic and biotic degradation
Degradation rates have been often assessed for polymers (see Auta et al., 2018;Gómez & Michel, 2013;Hennecke et al., 2018;Lambert et al., 2013a), and the degradation parameters half-life and degradation rate constant (t 1/2 and k deg , respectively) remain applicable; however, the increased complexity of polymer degradation mechanisms and products should also be considered. While degradation products of LMW chemicals are routinely incorporated into ERAs, the number and variety of products formed from polymer degradation may be far greater, potentially including HMW molecules, micro-and nano-scale particles, and oligomers and LMW chemical compounds (see Lambert et al., 2013aLambert et al., , 2013bSaad et al., 2010;Ter Halle et al., 2016;Weinstein et al., 2016). The complexity of the product mixture from degradation of a solid polymer along with the implications for polymer properties and key fate parameters are illustrated in Figure 1.
Degradation mechanisms and t 1/2 and k deg values depend on both polymer properties (including the presence of certain RFGs, hydrophobicity, MW, glass transition temperature [T g ], and fragment size, among others Ter Halle et al., 2017]) and environmental factors (including light and oxygen availability, temperature, pH, salinity, and biofilm formation [Da Lambert et al., 2013a;Morohoshi et al., 2018]). Polymer transformation products are likely to have different fate and degradation characteristics compared to one another and to the parent material, which will itself be altered, presenting challenges for characterizing potential risk. Standard test methods will require modification and additional analytical techniques to characterize these products and corresponding degradation pathways.
Polymer particles may be formed from breakdown of a solid polymer; in addition, while water-soluble polymers are most likely to degrade into oligomers and chemical compounds rather than particles, there has been speculation over the potential for soluble polymers to form insoluble material in the environment (Arp & Knutsen, 2020); and it should be noted that polymer solubility does not preclude nonbiodegradability and environmental persistence (Arp & Knutsen, 2020;Hennecke et al., 2018;Swift, 1998). Particles formed from polymer degradation can further fragment or aggregate (Liu et al., 2019); importantly, these secondary particles formed by polymer fragmentation are likely to differ from primary emitted particles such as primary microplastics. They will be more irregular in shape (see Frydkjaer et al., 2017), and both primary and secondary particles which have been exposed to the environment may have altered density (Chubarenko et al., 2016;Morét-Ferguson et al., 2010) and surface properties (Waldman & Rillig, 2020), with different RFGs, charge (S q ), and topography (Fotopoulou & Karapanagioti, 2012). These changes will influence fate; for example, the surfaces of ultraviolet lightdegraded polystyrene nanoparticles have been shown to be more oxygen-rich, potentially influencing aggregation behavior, compared to nondegraded particles (Liu et al., 2019).
Ultimately, chemical products will form from polymer degradation; several LMW chemical products have been identified from plastic degradation (reviewed by Bond et al., 2018) and other solid polymers such as latex (Lambert et al., 2013b). Most prioritization methodologies classify polymers of high average MW (≥1000 Da) as low concern (PLC) because of the expectation that they may be less able to cross organism membranes (OECD, 2009). However, all polymers have the potential to degrade into LMW species following emission to the environment, with many PLC exclusion criteria acknowledging "substantial" (bio)degradation as indicating potential concern (ECETOC, 2019).

Additional parameters for polymer exposure assessment
In addition to the established parameters for LMW chemicals that have been discussed and that are summarized in Table 1, it is clear that there are a number of properties of FIGURE 1: Summary of degradation and fate processes, including changes in key fate parameters, for a solid polymer material in an aquatic environment. UV = ultraviolet; MW N = number average molecular weight; T g = glass transition temperature; S A = surface area; RFG = reactive functional group; FGEW = functional group equivalent weight; t 1/2 = degradation half-life; k deg = degradation rate constant; PSD = particle size distribution; S q = surface charge; α = attachment efficiency; MWD = molecular weight distribution; q = charge or charge distribution; T m = melting point; T b = boiling point. polymers that are not applicable to LMW chemicals but which may be instrumental in polymer exposure assessment. Suitable parameters and descriptors for such properties are suggested in the present review. A combination of established and novel parameters to describe polymer environmental fate is likely to be necessary and will again be facilitated by classification of polymers as solid or dissolved. The overall picture is complex, with different sets of parameters likely being key for LMW chemical compounds, solid polymers, and dissolved polymers. This has been summarized and illustrated in Figure 2.
An obvious distinction of polymers is their distributed MW (OECD, 1991), which can be measured in terms of MW N , MW W , and MWD. The presence of leachable LMW compounds or oligomers in a polymer is also important because these may be more bioavailable (see Bejgarn et al., 2015). The MW N , MW W , MWD, and LMW content of polymers can be characterized using size exclusion chromatography (OECD, 1996a(OECD, , 1996b).
An important property determining fate is solubility. Hildebrand and Hansen solubility parameters (δ; Miller-Chou & Koenig, 2003) have been used to predict polymer solubility in various solvents (Venkatram et al., 2019); however, there are a number of limitations of such methods, and they should be considered only predictive (Venkatram et al., 2019). Experimental determination of a polymer's concentration in solution is critical (Hartmann et al., 2019;OECD, 2000). Polymer solubility is also key for their classification within the framework of the present review, along with polymer solidity or hardness; solidity is also significant for the ECHA definition of microplastics as solid (ECHA, 2019) and may influence environmental fate (e.g., by influencing biofilm formation [Muthukumar et al., 2011]). Solid polymers also have several properties which are not shared with dissolved polymers but are likely to be key for environmental fate, including particle size distribution (PSD), shape, surface properties, and aggregation characteristics.
Particle size, for example, will influence environmental fate and may dominate over other parameters such as density. Density (ρ) can be assessed via a number of methods (OECD, 2012a) and can influence position in the water column and settling into sediment in an aqueous environment (Chubarenko et al., 2016). However, in a modeling study, Besseling et al. (2017) found that while retention of 1-200-μm plastic particles in a river stretch increased with polymer density, retention of 0.1-1-µm particles was almost density-independent, instead being driven by particle size. Similarly, some plastic types that are denser than seawater have been found in the form of micro-and nanoparticles on the sea surface, suggesting that smaller debris may have different floatation behavior despite density considerations (Ter Halle et al., 2017). This phenomenon highlights the complexity that can arise through the overlapping influence of multiple fate parameters.
Standard methods for measurement of PSD are based on sedimentation, centrifugation or Coulter counter, or microscopic techniques for fibers (OECD, 1981). While size is most commonly used to refer to solid particles, dissolved polymer molecules may exist in the nano-size range, and thus measurement of hydrodynamic radius may be important in characterizing their fate. As well as influencing transport and vertical distribution, particle size may influence polymer degradation rate, along with particle shape . Particle shape may also influence residence time in organisms (Frydkjaer et al., 2017), as well as surface area (S A ) and therefore degree of biofouling, which can in turn influence settling time, heteroaggregation, and degradation (Chubarenko et al., 2016;Michels et al., 2018;Morohoshi et al., 2018). Shape and S A are thus potentially important fate parameters for particles.
Other surface characteristics of particles such as S q may be important (see Fotopoulou & Karapanagioti, 2012). Surface charge of nano-scale polymer particles in colloidal suspensions can be assessed by measurement of the zeta potential (ζ), which influences stability and therefore aggregation behavior (Cai et al., 2018;Liu et al., 2019;Oriekhova & Stoll, 2018;Saavedra et al., 2019;J. Wu et al., 2019). Aggregate formation is also key and may influence vertical transport of polymer particles in the environment (Michels et al., 2018). As described previously, the use of partition coefficients is not relevant to describe partitioning of solid particles via aggregation and deposition. Instead, kinetic parameters such as attachment FIGURE 2: Summary of the applicability of various fate parameters and key properties to low-molecular weight (LMW) chemical compounds, bulk solid polymers (including particles), and dissolved polymers. Parameters that are typically used in environmental exposure assessment of LMW chemicals are further categorized in terms of basic physicochemical properties (purple), partition coefficients (red), bioconcentration and bioaccumulation (green), and biotic and abiotic degradation (light blue). Additional and polymer-specific parameters suggested in the present review, which may be useful in polymer exposure assessment, are also shown (dark blue). ζ = zeta potential; PSD = particle size distribution; MW W = weight average molecular weight; MWD = molecular weight distribution; S A = surface area; α = attachment efficiency; T g = glass transition temperature; FGEW = functional group equivalent weight; MW N = number average molecular weight; ρ = density; R h = hydrodynamic radius; η = viscosity; k dep = deposition rate constant; δ = Hildebrand and Hansen solubility parameters; S q = surface charge; t 1/2 = degradation half-life; k deg = degradation rate constant; q = charge or charge distribution; K OC = soil organic carbon-water partition coefficient; AE = assimilation efficiency; pKa = dissociation constant; K OW = octanol-water partition coefficient; K d = = soil-water partition coefficient; k u = uptake rate constant; k d = depuration rate constant; γ = surface tension; T m = melting point; P = vapor pressure; BAF = bioaccumulation factor; BCF = bioconcentration factor; T b = boiling point; MW = molecular weight. efficiency (α) can be used (Praetorius et al., 2014). Attachment efficiency has been determined experimentally for analysis of heteroaggregation between microplastics, nanoplastics, and clays .
The deposition rate constant may also be relevant (along with α) to assess settling times in an aquatic environment when equilibrium partitioning to sediment does not apply. Deposition of airborne polymeric particles in the micro and nano ranges (Bergmann et al., 2019;Kawecki & Nowack, 2019;Wright et al., 2020) and dissolved polymers present in aerosols, for example, in agricultural sprays (see Felsot et al., 2011;Lewis et al., 2016), may also be significant. The deposition rate constant has been used to describe deposition of engineered nanoparticles both to soil and water from the atmosphere and to sediment from an aqueous environment (Meesters et al., 2014).
There are other fate properties that may be key to polymer exposure assessment. For example, viscosity (η; OECD, 2012b), also used in environmental fate analyses of oil spills (Sebastião & Soares, 1995), may be important for liquid polymers. In addition to T m , T g is useful in polymer matrix characterization because it describes the transition from rigid and glassy to rubbery and has been found to influence sorption and desorption of organic contaminants (Teuten et al., 2009) as well as polymer degradation rate .
In addition, metrics for quantifying exposure are key; while mass concentration remains sufficient for dissolved polymers, for solid polymers and particles, number concentration and PSD are likely to also be significant (Kookana et al., 2014). This is illustrated by the fact that larger particles may dominate in terms of mass, but smaller particles may dominate in terms of number (Schwaferts et al., 2019;Ter Halle et al., 2016), meaning the metric measured may influence conclusions drawn about relative environmental impacts.

Analytical techniques for polymer characterization
It has been recognized that standard test methods may need to be adapted for application to polymers (ECETOC, 2020). While some methods do exist that are specifically tailored to polymers or solids, such as for assessment of solubility, MWD, and PSD (OECD, 1981(OECD, , 1996a(OECD, , 2000, an array of additional techniques may be required for full characterization of a polymer. The traditional methodologies used for chemical analysis, including chromatography and mass spectrometry, may need to be adapted or replaced to characterize parameters such as shape, aggregation behavior, and topography. In addition, the existence of a "methodological gap" in the nano-size range has been highlighted (Schwaferts et al., 2019), and it has been recognized multiple times in the literature that there is a lack of both standardization and adequate validation of some techniques for plastic particle analysis Hidalgo-Ruz et al., 2012;Ivleva et al., 2017;Pico et al., 2019). Knowledge from nanoparticle and microplastic analyses will be invaluable in further developing techniques for polymer analysis in exposure and risk assessment. Importantly, given the potentially massive range of products that may be formed from polymer degradation, use of a wide array of techniques will most likely be necessary for a single environmental degradation study if all products are to be characterized. Fully characterizing the rate, route, and products of polymer degradation may therefore be difficult to achieve in a time-and cost-effective manner, despite the importance of such studies for ERA.

Structure-activity relationships and exposure models for polymers
Given that most QSAR models have been developed specifically for LMW organic compounds, many will be insufficient for application to polymers (ECHA, 2016), and prediction of polymer environmental fate should also address additional influences as a result of polymer size, MW, and macromolecular properties. A lack of data on polymer environmental fate will also limit development of polymer QSARs. Although models such as the Ecological Structure-Activity Relationship model include recommendations for assessing the aquatic hazard of polymers (Mayo-Bean et al., 2017), they are limited by availability of data and have been developed only for specific polymer classes, meaning they are often not applicable to new polymers (Nolte, Peijnenburg, et al., 2017).
Given the added complexity of polymers compared to LMW compounds and the additional parameters influencing polymer fate, complex exposure models for polymer ERA may also require additional considerations. While many simple, lower-tier models are likely to be appropriate for polymers, higher-tier models which require fate parameters as inputs may need to be adapted to account for the polymer-specific processes that we have described. For example, models such as the FOCUS models for pesticides (FOCUS Working Group on Surface Water Scenarios, 2001) and the exposure to Pharmaceuticals in the Environment (ePiE) model developed for pharmaceuticals, incorporate partition coefficients and loss processes such as degradation (Oldenkamp et al., 2018). However, for a solid polymer particle, partition coefficients are not applicable, and degradation processes may not indicate a decrease in exposure because initial degradation may simply form a larger number of smaller particles. Parameters such as size, shape, density, and attachment efficiencies, among others, will dictate transport and fate of particles (Kooi et al., 2018) in place of partition coefficients. Similarly, given the general lack of fate analyses of dissolved polymers, assessment of the applicability of fate models for LMW chemicals may be necessary, given that parameters such as size, MW, and macromolecular properties such as chain conformation are likely to influence dissolved polymer fate.

TOWARD A FRAMEWORK FOR POLYMER EXPOSURE ASSESSMENT
To move toward a framework for polymer environmental exposure assessment, we have identified key fate parameters and descriptors that are likely to be most significant (Figure 3). These include key physicochemical properties required for identification and characterization of polymers, which can also facilitate polymer grouping and prioritization. Approaches to polymer grouping have been discussed in detail by ECETOC (2019); in the present review we highlight key parameters for polymer characterization for exposure assessment based on the discussion of fate parameters, including properties such as MW parameters, solubility, presence of functional groups, and transition temperatures.
We have also identified the most relevant parameters for higher-tier exposure modeling ( Figure 3) and recommend that classification of polymers in terms of whether they will be in dissolved or solid form is likely to be useful in ERA because this will define the relevance of all other fate parameters to the polymer in question. This is particularly relevant for in-depth exposure assessment, to focus assessment efforts and avoid incorrect application of parameters. While parameters such as k deg , t 1/2 , and many of the key physicochemical properties identified previously will be relevant to both groups, properties such as PSD, attachment efficiencies, and surface properties are unique to solid materials; and equilibrium partition coefficients are only applicable to dissolved polymers. It is important to note that development of analytical techniques is key moving forward, both for monitoring studies and in characterization of key parameters for polymers.
From this framework (Figure 3), key considerations to address knowledge gaps can be identified, including the most important parameters for polymer identification, grouping, prioritization, and fate analysis; complex degradation processes and by-products of polymers; available analytical techniques for polymer analysis; and fate and exposure modeling of polymers. These considerations are addressed in the context of the exposureassessment framework (Figure 3) in the following section.

CONSIDERATIONS AND KEY RESEARCH NEEDS FOR POLYMER EXPOSURE AND RISK ASSESSMENT
Key parameters for polymer identification, grouping, and environmental fate There is a clear need to develop standard identifiers for polymers to avoid ambiguity in risk assessment; identifiers based on the key physicochemical properties summarized in Figure 2 may be useful in differentiating polymers formed from the same monomer units, which would otherwise not be distinguishable from just, for example, name and CAS number. A number of these descriptors have also been highlighted by ECETOC (2019), including MW (MW N , MW W , and MWD), T m , T g , and solubility, among others.
However, it is still unclear which parameters may be most important for polymer grouping and exposure assessment, given the complexity and potential overlap of factors in influencing environmental behavior. Development of grouping approaches based on correlation between key parameters and environmental behavior is necessary, which will likely require data from experimental fate and ecotoxicology studies for a FIGURE 3: Impact of polymer properties, analytical techniques, and fate parameters for solid and dissolved polymers in development of an environmental exposure assessment framework. pKa = dissociation constant; q = charge or charge distribution; t 1/2 = degradation half-life; k deg = degradation rate constant; ρ = density; K OW = octanol-water partition coefficient; K d = soil-water partition coefficient; K OC = soil organic carbon-water partition coefficient; R h = hydrodynamic radius; η = viscosity; k u = uptake rate constant; k d = depuration rate constant; α = attachment efficiency; PSD = particle size distribution; S A = surface area; S q = surface charge; ζ = zeta potential; T m = melting point; T g = glass transition temperature; FGEW = functional group equivalent weight; MW N = number average molecular weight; MW W = weight average molecular weight; MWD = molecular weight distribution.
Environmental exposure assessment of polymers-Environmental Toxicology and Chemistry, 2022;41:515-540 wide range of polymers. Assessing the ability of key parameters to predict environmental behavior of polymers is likely to be achieved through a combination of experimental fate studies and modeling; for example, Min et al. (2020) established key predictors for surface erosion and degradation of marine plastic debris based on physical properties and molecular structure. Similar analyses for other polymers and endpoints, based on use of experimental data, intrinsic properties, and key parameters to inform predictive modeling, are likely to be extremely useful in environmental exposure assessment and grouping. Further research into the relative extent that certain properties may influence hazard and fate, with establishment of a hierarchy of features to predict environmental behavior , as well as how these properties may interact to mitigate or exacerbate hazard, is warranted. Filling this research gap would also supplement development of QSARs and read-across approaches, as well as prioritization efforts for polymers and identification of data needs for risk assessment. Development of QSARs for polymers will also further consolidate grouping approaches and establishment of key parameters for environmental exposure assessment of polymers.
Research into cutoff points for solidity and solubility is also warranted given the potential ambiguity that may arise for polymers which are not clearly either solid or dissolved (e.g., waxes). For polymers of sufficiently LMW, parameters that would normally only be relevant for LMW chemical substances and oligomers (such as P and BCF) may become relevant, so it may be important to define MW cutoff points for such parameters. In addition, as knowledge develops of which properties of particles may confer hazard, such as shape and surface properties (see Della Torre et al., 2014;Frydkjaer et al., 2017), the relative importance of these parameters for grouping of micro-and nanopolymers may become apparent.

Polymer degradation and implications for fate
Many of the current standard test methods for degradation study different transformation pathways in isolation or under specific sets of conditions (e.g., OECD, 2004aOECD, , 2008; however, it is likely that in the environment these processes will occur in tandem and may interact. Therefore, simulation tests which closely mimic environmental conditions (e.g., OECD, 2004b), to study net degradation processes and products, are likely to be more useful in characterizing complex polymer degradation. Such tests are frequently employed in environmental exposure assessment and have been applied to a number of polymer classes. In particular, environmental exposure and risk assessments have been conducted for alcohol ethoxylates, alcohol ethoxysulfates, and polycarboxylate homo-and copolymers as part of the HERA project (HERA, 2004(HERA, , 2009(HERA, , 2014a(HERA, , 2014b, with degradation data for these classes of polymers being summarized as part of these risk assessments. In addition, Duis et al. (2021) gathered available data for several polycarboxylate polymers, polyethylene glycols, and polyquaterniums.
In the present review, we have further summarized the aforementioned collated degradation data for these polymer types, to provide a comprehensive overview of the available degradation data and test results for these polymers, presented in Table 2. Full details are presented in the Supporting Information. We have focused on available data relevant to environmental exposure assessment for water-soluble polymers, given the vast pool of studies available on degradation of marine plastic debris, which frequently employ varied and nonstandard methods.
While there are degradation data in a range of media available for many of these polymer groups (Table 2), it should be noted that these groups cover only a small fraction of the polymer types in current use, and degradation data for environmental matrices (surface waters, soils, and sediments) are limited. There are also few data available for polyquaterniums as a class (Duis et al., 2021), despite potential concerns relating to the environmental hazard of cationic polymers (see USEPA, 1997). In addition, a lack of information on experimental methods limits assessment of the quality of some results (Duis et al., 2021) as well as comparison and verification between studies, highlighting the need for transparency and standardization of methods for adequate risk assessment.
In general, it can be observed that alcohol ethoxylates, alcohol ethoxysulfates, and polyethylene glycols often exhibit higher rates or levels of degradation than polycarboxylates and polyquaterniums, although there are high levels of variation due to the wide ranges of polymers summarized together in the present review. Importantly, many studies focus on the extent of degradation and associated biodegradability endpoints (Table 2), whereas full environmental exposure assessment will in many cases require treatment of the degradation products formed. In addition, tests focused on measures such as CO 2 evolution may underestimate degradation for some HMW polymers which may undergo extensive fragmentation into lower MW polymer chains before complete mineralization; similarly, measurement of loss of a parent material may overlook the presence of persistent polymer chains of lower MW. Analysis of degradation products will likely require additional parameters and a wide array of analytical techniques to describe their fate. However, it may not always be feasible to characterize the full range of polymer degradation products, particularly given the constraints of current analytical methodologies for analysis of nano-scale polymer particles; therefore, further research into optimum methods by which polymer degradation can be characterized, which product types are most significant in terms of environmental risk, and how polymer properties can be predictive of degradation products (see Min et al., 2020) is warranted.

Characterization of polymers and degradation products
A further key consideration for polymer exposure assessment is the analytical tools available to characterize polymer fate and degradation processes. The applicability of existing standard test methods to analysis of polymer properties and fate parameters has been evaluated (ECETOC, 2020), and thus  "not inherently biodegradable"; "moderately/partly eliminated from water; virtually eliminated from water by, e.g., sorption to activated sludge"; "removed from waste water by, e.g., strong sorption on activated sludge" (Continued ) we present a holistic overview of how analytical tools could be deployed and further developed to better characterize polymer-specific fate properties and degradation products. Fate and degradation studies may involve use of complex environmental matrices, which will often require extraction or separation prior to analysis. A number of methods exist for extraction of micro-and nanoplastics from soils, sediments, and biota, including density separation and chemical or enzymatic digestion (see Hurley et al., 2018;Karlsson et al., 2017). However, these treatments may alter the particle analytes (Enders et al., 2017;Hurley et al., 2018;Rist et al., 2017), and thus methods should be tested and validated for the polymers in question. For analysis of LMW chemical compounds in complex environmental matrices, various solvent extraction techniques are typically used (see Basheer et al., 2005;Berlioz-Barbier et al., 2014;Martínez-Parreño et al., 2008), which may be developed and optimized for dissolved polymers (see ćAnti et al., 2011).
A number of reviews of available techniques for analysis of micro-and nanoplastics in the environment are available (Fu et al., 2020;Li et al., 2018;Nguyen et al., 2019;Schwaferts et al., 2019;Silva et al., 2018). The advantages and limitations of some key analytical methods for solid polymers and their degradation products are summarized in Table 3.
Microscopy, particularly light microscopy and scanning electron microscopy, is commonly used in visualization of plastics, allowing characterization of size and shape of particles (see Hernandez et al., 2017;Oriekhova & Stoll, 2018;Ter Halle et al., 2016) and surface degradation of macropolymers (Gómez & Michel, 2013;Musioł et al., 2017). However, unequivocal chemical identification of the analyte is essential and relies on combination with spectroscopic methods such as Fourier-transform infrared (FTIR) and Raman spectroscopy Cabernard et al., 2018), which may also provide information on chemical changes with degradation (Da . Automation can provide faster and more reliable results and reduce issues with bias and sample representativeness, for example, in focal plane array-based micro-FTIR Primpke et al., 2017). However, spectroscopic techniques are unable to give chemical information on particles below the micro-scale.
Information on PSD can also be obtained from scattering or diffraction-based techniques, which can be applied to nano-scale particles (see Gigault et al., 2016;Lambert & Wagner, 2016a;Mintenig et al., 2018). Laser diffraction instruments in particular have the potential to cover a wide particle size range (Keck & Müller, 2008;Witt & Röthele, 1996), and dynamic light scattering (DLS) and nanoparticle tracking analysis are useful for characterizing particle aggregation (see Besseling et al., 2017;Filipe et al., 2010;Gigault et al., 2017). However, such techniques typically utilize spherical models to describe particles (see Eshel et al., 2004;Frydkjaer et al., 2017;Lambert & Wagner, 2016b), which may influence analysis of irregularly shaped secondary particles. Techniques such as DLS and multi-angle light scattering (MALS) may also require preseparation of particles into specific size fractions, which can be achieved using asymmetric flow fieldflow fractionation (AF4; see Filipe et al., 2010; Gigault      Eshel et al. (2004), Keck and Müller (2008), Kokalj et al. (2018), Lee et al. (2014), Witt and Röthele (1996) MALS      Mintenig et al., 2018); however, it has been highlighted that many AF4 techniques have been optimized using primary particles and that secondary particles may behave differently (Schwaferts et al., 2019). Chromatographic techniques utilized in nanoparticle separation and analysis that have the potential to be adapted for plastic particle analysis have also been highlighted by Schwaferts et al. (2019), including hydrodynamic chromatography and high-performance liquid chromatography.
For chemical analysis of nano-sized particles, mass spectrometric techniques are crucial. Pyrolysis gas-chromatography mass-spectrometry (py-GCMS) has been used to identify polymer types of plastic particles (Fries et al., 2013;Hermabessiere et al., 2018;Ter Halle et al., 2017) and may reveal changes resulting from degradation (Ter Halle et al., 2017). Thermal extraction desorption gas-chromatography mass-spectrometry can be used to directly analyze and potentially quantify plastic particles in an environmental sample (Dümichen et al., 2014(Dümichen et al., , 2015(Dümichen et al., , 2019, making it a potentially powerful technique for analysis of polymers in environmental matrices in fate and degradation studies. Other techniques are available to determine additional key properties for polymer fate analysis. For example, MW information can be obtained using gel-permeation chromatography, which has been used in analysis of microplastics down to 10 μm (Hintersteiner et al., 2015). Differential scanning calorimetry can give information on thermal properties including T m and T g (Deroiné et al., 2014;Musioł et al., 2017).
Although most studies have focused on analysis of solid plastic polymers, particularly microplastics, most chemical identification techniques will also be suitable for dissolved polymers, as highlighted by Arp and Knutsen (2020). In addition, scattering methods have been used to characterize the hydrodynamic radius of polymers in solution (Armstrong et al., 2004). However, while some analyses of water-soluble polymers in environmental matrices have been carried out (see Antić et al., 2011), overall few techniques have been developed for environmental analysis of dissolved and water-soluble polymers (Huppertsberg et al., 2020), presenting a key research need for environmental exposure assessment.
Each technique has a workable size range ( Figure 4) and provides different levels of information, emphasizing the importance of addressing the research need in question (Elert et al., 2017). It is likely that full characterization of a polymer and its degradation products for fate and exposure assessment will require a combination of techniques, which should be tailored to the nature of the polymer in question. For a solid polymer, this may include all or a combination of chromatographic, spectroscopic, scattering, and spectrometric techniques. For example, Mintenig et al. (2018) recently combined AF4-MALS with py-GCMS to characterize both particle size and polymer type of nanoplastics in environmental samples within a suggested framework for micro-and nanoplastic analysis. Use of multiple techniques may aid in analysis of diverse polymer degradation products in standard degradation tests when characterizing full rate and route (see OECD, 2002OECD, , 2008 as well as facilitating development of new standard test methods for polymer-specific properties and fate parameters. For example, DLS and spectrophotometry may be useful in establishing standardized methods for determining α of polymer particles to describe aggregation with suspended particles  as an alternative to partition coefficients. However, the need for full sample characterization should be balanced with time and cost effectiveness and the level of information needed for adequate risk assessment. As methods and data relating to polymer risk assessment continue to develop, the key properties, polymer types, and degradation products dictating fate and hazard may be elucidated and used to refine and focus risk-assessment methodologies and analytical technique development. Analytical techniques developed for nanoparticles and microplastics will be useful in solid polymer risk assessment; however, it has been recognized that a previous lack of standardization and adequate quality control of techniques for microplastic analysis has hindered progress in assessing their environmental risk . Moving forward in polymer analysis, further development and standardization of techniques is required for robust risk-assessment methodologies, with improvement and adaptation of the techniques discussed in the present review as well as development of novel methods likely being necessary.

Fate and exposure models for polymers
Given the differences in applicability and importance of fate parameters to polymers compared with LMW compounds, development of methods for prediction of fate properties as well as higher-tier exposure models for polymers which incorporate both measured and predicted fate parameters is warranted. Although some efforts have been made to predict environmental fate of polymers based on their intrinsic properties  and QSARs have been developed for algal toxicity of polymer particles (Nolte, Peijnenburg, et al., 2017), further development of robust data sets for model development to establish an array of QSARs for polymer environmental fate is warranted. Adaptation of QSARs for engineered nanoparticles may also be useful for application to polymer particles.
Exposure models for engineered nanoparticles have now been developed and range in complexity from emission-based mass-balance models (see Gottschalk et al., 2009) to multimedia (e.g., Meesters et al., 2014) and spatiotemporally resolved (see Domercq et al., 2018;Quik et al., 2015). Recently, fate models have also been applied to micro-and nanoplastics (see Besseling et al., 2017;Nizzetto et al., 2016), with the unique combination of low density, wide size range, persistence, and variable shape of plastic particles distinguishing them from other particle types in fate and exposure modeling (Kooi et al., 2018). Research on environmental exposure to dissolved polymers remains scarce, and exposure models may again require development of additional input parameters, given the additional properties of polymers which are not applicable to LMW chemical compounds.

CONCLUSIONS AND RECOMMENDATIONS
Given the widespread and increasing use of both solid and liquid or water-soluble polymers and their subsequent release into the environment, development of ERA approaches is essential. The unique and complex nature of polymers, including their high and distributed MWs, potentially complex matrix properties, and the presence of various additives, means that adaptation of current risk-assessment approaches is warranted.
In environmental exposure assessment, use of key fate parameters is essential for fate characterization and modeling; however, some parameters established for LMW chemical compounds are unlikely to be relevant to polymers. In the present review, an assessment of the relevance of typically used fate parameters to polymers has been performed, revealing that solidity and solubility of polymers are key to the applicability of such parameters and providing a useful basis for development of an environmental exposure assessment framework. Additional parameters and parameters describing the unique properties of polymers compared to LMW compounds have been suggested, many of which may be useful in higher-tier fate and exposure assessments of polymers.
Incorporation of these parameters into an environmental exposure assessment framework for polymers has been suggested in the present review based around this categorization, highlighting which parameters may be most important both in polymer identification and grouping and for exposure assessment and fate modeling. However, it is clear that limitations and knowledge gaps remain; key research needs to develop environmental exposure assessment methodologies for polymers are identified and highlighted as follows: • Standard identification methods for polymers which incorporate their complexity and key properties should be developed. In addition, the relative significance of key fate parameters, particularly in polymer identification and in impacting fate behavior, should be assessed to establish a base set of parameters for screening-level assessments as well as provide insight on which parameters are most significant for higher-tier assessment. This will facilitate prioritization efforts for polymers and subsequent in-depth exposure assessments. • Research into characterizing and defining polymer solidity and solubility to reduce ambiguity in classification is essential. • The potential for polymers to further expose the environment to a complex mixture of degradation products with altered fate parameters should be accounted for in exposure assessment. To incorporate degradation products into a risk assessment, a deeper understanding of the pathways and products of polymer degradation under environmentally relevant conditions is required, with particular focus on potential changes in key fate parameters and environmental risk. • There is a clear need to develop, adapt, and standardize validated and reliable analytical methods for characterization of polymers and their degradation products, to measure properties relevant to exposure assessment as well as characterize degradation processes and products for exposure characterization and modeling. For full characterization, multiple techniques tailored to the polymer analyte in question may be required in tandem; for example, all of chromatography, scattering or microscopy, and spectroscopy or spectrometry may be required for complete characterization of a nonhomogeneous mixture of polymer particles. However, as knowledge of key polymer types, properties, and degradation products implicating risk assessment improves, methods can be refined and focused to provide sufficient levels of information with minimum application of techniques. • While simple lower-tier models may be appropriate for polymer exposure assessment, higher-tier exposure models that account for the unique properties and fate characteristics of polymers should be developed. Adaptation of models from analysis of engineered nanoparticles may be useful for application to micro-and nanopolymer particles, such as microplastics; and a combination of modeling approaches from both LMW compounds and nanoparticles may be necessary for characterizing the fate of both a solid parent polymer and its chemical degradation products. This will be further supplemented by development of QSAR approaches and data sets for polymers. • Further research into the critical fate properties of watersoluble polymers and their breakdown products is warranted to better characterize their risk to the environment. This would help to prioritize data-generation needs and identify polymers for further investigation.
Approaches to polymer environmental exposure and risk assessment should incorporate and allow for the complexity of polymers. Developing knowledge of how polymer properties influence fate, and therefore which are most important in characterizing risk, as well as methods to incorporate complex degradation products in exposure and hazard assessment is essential to develop adequate and robust risk-assessment methodologies for polymers.
Supporting Information-The Supporting Information is available on the Wiley Online Library at https://doi.org/10.1002/ etc.5272.