Chemical Activity ‐ Based Loading of Arti ﬁ cial Sediments with Organic Pollutants for Bioassays: A Proof of Concept

: Persistent organic pollutants (POPs) pose a risk in aquatic environments. In sediment, this risk is frequently evaluated using total or organic carbon ‐ normalized concentrations. However, complex physicochemical sediment characteristics affect POP bioavailability in sediment, making its prediction a challenging task. This task can be addressed using chemical activity, which describes a compound's environmentally effective concentration and can generally be approximated by the degree of saturation for each POP in its matrix. We present a proof of concept to load arti ﬁ cial sediments with POPs to reach a target chemical activity. This approach is envisioned to make laboratory ecotoxicological bioassays more reproducible and reduce the impact of sediment characteristics on the risk assessment. The approach uses a constantly replenished, saturated, aqueous POP solution to equilibrate the organic carbon fraction (e.g., peat) of an arti ﬁ cial sediment, which can be further adjusted to target chemical activities by mixing with clean peat. We demonstrate the applicability of this approach using four polycyclic aromatic hydrocarbons (acenaphthene, ﬂ uorene, phenanthrene, and ﬂ uoranthene). Within 5 to 17 weeks, the peat slurry reached a chemical equilibrium with the saturated loading solution. We used two different peat batches (subsamples from the same source) to evaluate the approach. Variations in loading kinetics and eventual equilibrium concentrations were evident between the batches, which highlights the impact of even minor disparities in organic carbon properties within two samples of peat originating from the same source. This ﬁ nding underlines the importance of moving away from sediment risk assessments based on total concentrations. The value of the chemical activity ‐ based loading approach lies in its ability to anticipate similar environmental impacts, even with varying contaminant concentrations. Environ Toxicol Chem 2024;43:279 – 287. © 2023 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.


INTRODUCTION
Environmental risk assessment of persistent organic pollutants (POPs) has been primarily based on substance concentrations in environmental matrices.The concentration-based approach is useful for describing POP concentrations in different ecosystem components, such as sediment.However, it has limitations when it comes to effectively assessing the actual risks posed to biota living in the sediment.The reason for this is that the sorption strength to different types of sediment organic carbon varies considerably.For example, strong sorption to highly condensed carbonaceous materials, such as black carbon, can substantially reduce bioavailability, uptake, and bioaccumulation of harmful organic substances (Rust et al., 2004).Kupryianchyk et al. (2011) showed varying levels of mortality among benthic invertebrates due to POP exposure in sediment with different black carbon content, even when the total sediment concentrations of these substances are the same.Abiotic partitioning processes of POPs in the environment, such as the distribution and solubilization of chemicals from sediment into the water column, show a similar dependence on physicochemical sediment characteristics (Cornelissen et al., 2005).For example, sediment with a higher relative contribution of marine versus terrestrial organic matter was found to have a higher sorption capacity of POPs and therefore lower release into the water column (Nybom et al., 2021).
Fugacity-based concepts, such as the chemical activity of a substance, characterize the bioavailability and environmental partitioning of POPs more reliably than total concentrations (Di Toro et al., 1991;Gobas et al., 2018).The concept of chemical activity (a), or relative saturation, dates back to the early 20th century, when Gilbert N. Lewis defined it as an accurate measure of a substance's ability to undergo spontaneous reactions, such as partitioning into another environmental phase (Lewis, 1907).Chemical activity is a unitless metric describing the chemical's reaction potential compared with a reference phase.For POPs, chemical activity can be approximated as a fraction of the concentration of the substance and its solubility in any matrix, such as water (Equation 1): where C is the concentration and S is the solubility of the compound in a given matrix.
In the environment, dissolved POPs behave like liquids, even though the substances exist as solids in their pure form at ambient temperature.Hence, the reference state of a solid POP has to be adjusted to represent the hypothetical, subcooled liquid of the substance (Equation 2): where S L is the solubility of the subcooled liquid of the compound of interest.
No such adjustment is necessary for POPs that are liquids at the experimental temperature (Ferguson, 1939).As a result, the activity can reach unity for such liquid compounds.For POPs in the solid state at the experimental temperature, the maximum chemical activity (a max ) is always <1 due to the higher solubility of a subcooled liquid compared with the crystalline solid.The specific value for a max can be approximated through the fugacity ratio of the compound, which in turn is a function of its melting point T m and the environmental temperature T (Equation 3; Mayer & Holmstrup, 2008): There are several advantages of describing the occurrence of a POP in the environment by its chemical activity instead of total concentration.A POP's chemical activity expresses its tendency to leave the current matrix and partition into another, such as the transfer from sediment particles into water or biota tissues (Di Toro et al., 1991;Ferguson, 1939).Total concentrations cannot adequately estimate the potential for bioaccumulation or adverse effects, but chemical activity can.The applicability of chemical activity to account for the bioavailability of POPs in sediment can be exemplified by the effect of the sediment's organic or black carbon content.When sediment organic or black carbon content increases, the POP solubility within this sediment also increases.Consequently, the chemical activity will be lower at constant total POP concentrations (and increased organic or black carbon content due to increasing denominator in Equations 1 and 2).Hence, chemical activity can capture the difference in bioavailability between sediments with similar POP total concentrations but different organic or black carbon content.When the chemical activity is equal in sediment and biota, the thermodynamic potential is equal between these two matrices despite possibly different concentrations.Hence, using chemical activity-based assessments facilitates data comparability and provides a better understanding of the bioaccumulation risk and/or adverse effects (Gobas et al., 2018;Reichenberg & Mayer, 2006).
To some extent, the lipids in biota act similarly to the organic carbon in sediments, with higher body lipids resulting in higher POP solubilities and thus lower chemical activities at the same total concentrations.Lower chemical activities imply a lower effective concentration or reactivity of the chemicals, making them less likely to engage in biochemical reactions within the organism that may lead to toxic effects.Consequently, POP chemical activities correlate positively with adverse effects observed in exposed organisms.Moreover, compounds that exert non-specific effects (i.e., baseline toxicity or narcosis) show a consistent onset of adverse effects within a narrow chemical activity range of a = 0.01 to 0.1, corresponding to 1% to 10% of their saturation level in the organism (Gobas et al., 2018;Schmidt et al., 2013;Smith et al., 2013).Because of the normalization of a compound's concentration to its compound-specific thermodynamic property, such as its subcooled liquid solubility, the unitless chemical activity of POPs in any matrix is additive, simplifying the assessment of complex chemical mixtures (Smith et al., 2013).Thus, the additivity of a makes it a tool for assessments of the effects of chemical mixtures.
Per definition, chemical activities are equal in all phases of a system in equilibrium.Because POP solubility data are readily available for aqueous solutions but not for sediments or organisms, the simplest approximation of the chemical activity in a non-aqueous environmental matrix is through the freely dissolved concentration (C free ) in water that has been equilibrated with the environmental matrix in question.Freely dissolved water concentrations can be accurately assessed by equilibrium passive sampling methods, which are not influenced by accidentally sampled particulate matter or cosolvents, such as dissolved organic carbon (Mayer et al., 2014).Numerous studies have shown that passive sampling provides concentrations that can be more reliably linked to the toxicity of sediment-associated contaminants compared with total or organic carbon-normalized concentrations (Li et al., 2013;Lyytikäinen et al., 2003;You et al., 2008).
Total sediment concentrations measured conventionally in sediments can be converted to C free values and chemical activities using partition ratios (e.g., K OC ).More advanced and complex models also consider other factors, such as organic carbon quality, which, together with its quantity, can affect the predicted C free values in sediment pore water (Cornelissen et al., 2005;Hawthorne et al., 2007;McGrath et al., 2019).Such models have higher accuracy for specific sites but are still limited by the variability of sediment characteristics and lack of consistency in literature values for the physicochemical properties of POPs and organic carbon quality used in these calculations (Hawthorne et al., 2006(Hawthorne et al., , 2011)).
Using chemical activity as a metric in environmental risk assessment can provide more environmentally relevant and comparable information about the risk of contaminants and reduce the reliance on models with a high degree of uncertainty.In monitoring settings, passive samplers can be used to measure chemical activities directly.However, it is important to substantiate these measurements with reliable data on POP toxicity at different chemical activities.
A commonly reported toxic range for POPs is a = 0.01 to 0.1, primarily based on water-only exposures and a limited range of chemicals and test organisms.Available guidelines can be adjusted to use chemical activity as a dose metric for assessing median lethal or effective concentrations (e.g., EC50) of POPs in sediment (e.g., Organisation of Economic Coordination and Development [OECD], 2004[OECD], , 2007, or the US Environmental Protection Agency [USEPA], 2000).In these guidelines, the source of organic carbon is peat, a natural substrate that can exhibit a broad spectrum of physicochemical properties attributed to different degrees of decomposition of the original plant material (Moore & Shearer, 2003).In particular, the sorption capacity and organic carbon stability can thus vary substantially between different peats (Jesus et al., 2022).Consequently, these guidelines fall short in specifying organic carbon in the test sediment for determining the bioavailable fraction of a spiked compound.Due to the even more complex composition of natural organic carbon, greater variability in observed toxicities can be expected for natural sediments.Using lethal or effective chemical activities instead of concentration-based metrics can help overcome this issue (Reichenberg & Mayer, 2006).These chemical activitybased dose metrics inherently account for the thermodynamic properties that affect sorption behavior and thus POP bioavailability in a given matrix.The knowledge of the matrix properties is not necessary when equilibrium partitioningbased approaches are used to either confirm the exposure a (by equilibrium passive sampling) or to load the sediments to a target a value.However, no guideline is currently available for an equilibrium partitioning-based loading of test substances in sediment-water systems.
We propose an equilibrium partitioning-based loading approach for sediments instead of conventional spiking protocols (OECD, 2004(OECD, , 2007;;USEPA, 2000), which introduce a test substance based on their total concentration.Our approach ensures a test system loaded at an intended C free , and therefore at a known chemical activity, similar to passive dosing techniques used in water-only exposures (Smith et al., 2010).We present a proof of concept for this loading technique, focusing on the organic fraction of artificial sediments associated with peat used as a model organic substrate.To demonstrate the applicability of the approach, we saturated the peat with a mixture of four polycyclic aromatic hydrocarbons (PAHs) as model compounds: acenaphthene (ACE), fluorene (Flu), phenanthrene (PHE), and fluoranthene (FluO).A desired target a can then be reached by mixing (i.e., diluting) the saturated peat with clean peat.Assuming approximately linear sorption isotherms, the final a of the mixture can be calculated as the fraction of the known activity at saturation (a max ) equal to the dilution ratio.The peat at target a can be used in conjunction with inorganic materials to create artificial sediments for sediment-water bioassays as, for example, specified in OECD guidelines (OECD, 2004(OECD, , 2007)).Adding these inorganic constituents to the peat has negligible effects on the a values in the final sediment because their hydrophobic character means that sorption interactions with the organic fraction are several magnitudes higher (Talley et al., 2002).

Chemicals and equipment
Standards of ACE, Flu, PHE, and FluO and their deuterated variants were purchased at >97% purity from Sigma-Aldrich Sweden AB.SupraSolv ® -grade ethyl acetate, n-hexane and acetone (Merck AB) were used for extraction and sample preparation.Polyethylene (PE) sheets at 35 µm thickness were acquired from Rani Plast.Natural peat (Solmull Naturtorv) without additives was purchased from a local gardening store.

Peat preparation
The fresh peat was freeze-dried (FreeZone 70040; Labconco Corporation) for 72 h and ground using an electromagnetic micro pulverizer (FRITSCH).Particles >250 µm were sieved out before the use of the peat.The total organic carbon content of the peat was determined from a small subsample treated with 5% HCl to remove inorganic carbon.After drying overnight at 60 °C, the peat samples were weighed into tin capsules and run on an Elemental Combustion System (Costech 4010; Costech Analytical Technologies).This initial elemental analysis of the dry, pristine peat showed a relatively high organic carbon content of 49%, which could indicate that the peat was harvested from a deeper, more degraded deposit (Wang et al., 2017).

Equilibrium PAH-loading concept for organic matter
The basic concept of loading the peat on a thermodynamic equilibrium basis takes advantage of the fact that-even without knowing either the concentration or solubility of the PAH in the peat-its chemical activity at saturation is known (a max ; Equation 3).Saturating the peat can be achieved by exposing it to saturated PAH solutions.However, care should be taken to avoid depletion of the saturated loading solution.For the saturation of passive dosing silicone donors, saturated methanol suspensions are used to avoid depletion, (e.g., Smith et al., 2013).This allows large quantities of PAHs to be dissolved while the remaining crystals in the suspension form a reservoir that balances losses through partitioning to the clean receptor phase.Utilizing methanol to load peat designated for formulated sediments in bioassays is not recommended, owing to both the toxicity of methanol and the potential dissolution of peat-associated organic substances.Furthermore, a suspension cannot be used directly to load the peat because separating peat particles and PAH crystals would be impossible.Hence, water appears to be the only appropriate solvent for this application.
A constant replenishment of PAHs in the aqueous solution is required to keep it at saturation after introducing the peat, following the basic principle of a generator column (Dillon & Burton, 1991).For this purpose, we designed a semipermeable PE pouch that can be filled with an aqueous suspension of PAH crystals before being sealed and submerged into a vessel filled with a water-peat slurry (Figures 1 and 2).This design allows the dissolved PAHs to diffuse out of the pouch, functioning as a reservoir that prevents depletion of the water phase while PAH crystals are never in contact with the peat particles.The saturated peat can then be diluted to the intended target activities using clean peat from the same batch.Assuming linear sorption isotherms in the target concentration range (Endo et al., 2009) ensures that the chemical activity will be proportional to the ratio of saturated to clean peat (e.g., halved activity at a 1:1 mixing ratio).

Peat equilibrium loading setup
The PE pouch consisted of two 35-µm PE-sheet squares of ca. 8 × 8 cm heat-fused over three edges.The pouch was filled with Milli-Q (MQ)-water and PAH crystals and sealed, keeping as little air inside as possible.Before making the pouch, the PE sheets were washed sequentially in ethyl acetate, acetone, and n-hexane.With each solvent, the sheets were kept overnight on a roller mixer.After the n-hexane wash, the material was spread out on acetone-rinsed aluminum foil and left to dry briefly in a fume hood.The pouches were placed into 750-ml glass bottles containing a peat slurry (MQ-water and peat); a separate bottle was used for each of the four PAHs.The bottles were closed with Teflon-lined lids, wrapped in aluminum foil, and agitated on an orbital shaker at room temperature.
Choosing a setup in which the four PAHs were loaded into the peat in four individual bottles allowed omitting a loading blank.Sample contamination can be assessed by measuring all four PAHs in each of these bottles.At each time point, only the target PAH was detected, with the three other PAHs falling below the detection limit confirming that there was neither initial contamination present in the peat nor cross-contamination occurred throughout the experiment.
The peat loading was tested in two separate experimental runs, hereafter referred to as Batches A and B. Batch A featured a higher sampling frequency at the beginning of the loading process and was terminated after ca.1000 h, whereas Batch B was set up as a long-term experiment with a duration of up to ca. 3000 h and lower sampling frequency (Table 1).The two batches were conducted with peat from the same supplier sampled on two different occasions; it was dried and ground in the same manner.At each sampling point, a small sample of  the peat slurry (100-200 mg dry wt) was removed using a glass pipette and placed on a pre-combusted (450 °C) glass fiber filter to remove excess water.Each sample was split into a duplicate subsample for chemical analyses (n = 1-2 for Batch A, n = 2 for Batch B) and another subsample for dry weight determination.For the chemical analyses, approximately 10 to 30 mg (dry wt; Quintix 224-1S, Sartorius) per replicate were weighed into glass tubes and mixed with anhydrous Na 2 SO 4 (combusted at 450 °C) until a free-flowing powder was obtained.For the dry weight determination, a single replicate (10-30 mg dry wt) was weighed and subsequently oven-dried overnight at 105 °C.Method blanks were processed along with the peat samples and consisted of Na 2 SO 4 only.

Chemical analyses
After drying with Na 2 SO 4 , the peat samples were spiked with mass-labelled (deuterated) ACE, Flu, PHE, and FluO and extracted using accelerated solvent extraction (Thermo Fisher Dionex 350 ASE; Thermo Fisher Scientific) using acetone:n-hexane (1:1 v/v) at 100 °C and 1500 psi.Due to the expected high PAH concentrations, three cycles with 10 min of static time were used.To fill up empty space in the 22-ml extraction cells, the sample volume was increased with diatomaceous earth (Thermo Fisher Scientific).The instrumental analysis was carried out on a Thermo Trace 1310 gas chromatograph (GC) coupled to a single quadrupole mass spectrometer using electron ionization (ISQ LT, Thermo Fisher Scientific) operated in selective ion monitoring mode.One quantifier ion and one to two qualifier ions were measured for each substance.The separation of analytes was achieved on a 30-m DB-5 mass spectrometry (MS) column using a temperature gradient from 60 to 310 °C.The samples were injected at 60 °C into a PTV inlet, which was subsequently ramped up to 280 °C for a full transfer in splitless mode.

Data acquisition and analysis
The raw GC-MS data were evaluated with Chromeleon 7 (Thermo Fisher Scientific).Each analyte was quantified through its surrogate standard pair using a 10-point calibration curve.Instrument blanks (n-hexane) were run after five samples, and a single calibration point was run after each 15 samples as the quality control sample.Only correctly qualified peaks were evaluated (less than 20% qualifier ion ratio deviation as averaged in the calibration points).The results were processed with Excel 2019 (Microsoft) and GraphPad Prism 9 (GraphPad Software).The saturation curves for the peat over the loading time were calculated using a three-parametric asymptotic exponential growth model (Equation 4): where Y and Y max are the concentration at the time point t (h) and the maximum concentration achieved at saturation (both in mg PAH/g dry peat), that is, the solubility of the PAHs within the peat.The initial concentration (Y 0 ) was constrained to 0 in the model.The calculated asymptote (Y max ) was assumed to represent the absolute solubility of the PAHs in the peat.The time point at which the lower boundary of the 95% confidence interval for Y max is reached was assumed to be the time required to reach an equilibrium (t 95% ).The model fit was assessed using the root mean squared error (RMSE).The RMSE <0.75 was considered a very good fit, while values of 0.75 to 1.00, 1.00 to 2.00, and >2.00 were considered good, satisfactory, and not satisfactory, respectively.Substituting the concentration C in Equation ( 2) with the calculated solubility (Y max ) and a with a max (Equation 3) allows for the determination of the subcooled liquid solubility S L of each PAH in the peat.This can be used to convert the concentrations measured at subsaturation levels (e.g., at earlier time points during the loading process) into chemical activities.

RESULTS AND DISCUSSION
The concentrations and chemical activities of the four PAHs in peat increased with time, as demonstrated in Figure 3.The regression parameters are presented in Table 2.The RMSE indicated that the fit of the chosen model was good for ACE (0.30), Flu (0.32), and PHE (0.65), and acceptable for FluO (1.03).Saturation of the peat was evident only in the long-term experiment (Batch B), with t 95% of 2326, 2715, 1715, and 806 h for ACE, Flu, PHE, and FluO, respectively, and solubilities (Y max ) of 6.31, 5.99, 6.25, and 4.34 mg/g, respectively.For PHE and FluO, the observed concentrations in Batch A exceeded the estimated solubilities in Batch B after 1000 h, while for ACE and Flu, they reached comparable levels in a shorter timeframe.The saturating concentrations in Batch A would therefore have been notably higher than in Batch B. Thus, we hypothesize the two peat batches had different sorption properties, causing the differences in PAH concentrations and chemical activity (Figure 3).
A range for the final (saturated) PAH concentrations in peat can be predicted using literature partition ratios between organic carbon and water (K OC ) and water solubilities.The saturated concentrations obtained in this pilot study fall within or close to the range of calculated values using K OC and solubility data from Schwarzenbach et al. (2017) and the Swedish Environmental Protection Agency (Naturvårdsverket et al., 2011; Table 2 and Figure 3).For ACE, Flu and PHE, the predicted concentration range at saturation obtained from the two used literature sources spans a factor of approximately 2 (2.88-6.29, 2.28-5.51, 4.63-9.37, respectively), while for the heaviest PAH used in the present study, FluO, the range is higher (factor 4, 3.96-16.4mg/g).In line with these observations, the variability of published and peer-reviewed sediment EC50 values for PAHs increases with increasing PAH molecular weight (Fisher et al., 2011;Lotufo, 1998;Tani et al., 2021;Verrhiest et al., 2001).Measured and calculated K OC values follow the trend of increasing variability with increasing compound molecular weight or hydrophobicity.Unique properties of different organic carbon types have been suggested as the most likely cause for this divergence of reported partition ratios and toxicity parameters (Hawthorne et al., 2006;Nguyen et al., 2005).The larger PAH molecules are also the more hydrophobic compounds, and with that comes a higher importance of their sorption to organic carbon in governing bioavailability and toxicity in sediment.Analytical artifacts could also contribute to the differences in partition ratios between the low-and highmolecular-weight PAHs reported in the literature due to the increasing difficulty of accurately measuring aqueous FIGURE 3: Temporal dynamics of polycyclic aromatic hydrocarbon (PAH) concentrations in two separately loaded peat batches.Each of the four model PAHs (acenaphthene, fluorene, phenanthrene, and fluoranthene) was loaded independently.Batch A was terminated earlier to allow a higher sampling frequency within the first 1000 h, whereas Batch B was continued for 3000 h.All curves are displayed with their 95% confidence intervals.Dotted horizontal lines represent the lowest and highest literature values reported for PAH solubility in peat.The right y axis shows the calculated chemical activity (Equation 2) using the experimentally obtained values (Table 2); these values apply only to Batch B because no saturation level could be determined for any PAH in Batch A.  2017) and the Swedish EPA (Naturvårdsverket et al., 2011).c Subcooled liquid solubilities (S L ) of PAHs in peat were calculated using Equation (2) with experimentally determined solubilities (Y max ).
Y and Y max are the concentration at the time point t in hours and the maximum concentration that is achieved at saturation, that is, the solubility (both in mg PAH/g dry peat).The initial concentration (Y 0 ) was constrained to 0 in the model.The time required to reach equilibrium with reasonable certainty was defined as t 95% , which is where the curve reaches the lower 95% confidence interval (CI) for Y max .Solubility of PAHs in peat was calculated using K OC values from Schwarzenbach et al. (Schwarzenbach et al., 2017) and the Swedish EPA (Naturvårdsverket et al., 2011).concentrations of compounds with high hydrophobicity.To overcome these challenges, chemical activity-based exposures can be utilized because they inherently internalize factors such as the K OC for the specific organic carbon of a given test medium.As presented in our study, the equilibrium-driven loading approach could prove especially useful for highly hydrophobic POPs with highly variable reported physicochemical parameters.
In Batch B, the concentrations and chemical activity for the two lighter PAHs, ACE, and Flu, increased slowly, even after 1700 h.By contrast, the heavier PAHs, PHE, and FluO, had a more defined asymptote, that is, concentrations increased less after 1700 h.The observed longer equilibration time for ACE and Flu may seem surprising due to their lower molecular weight and higher water solubility, which generally facilitates diffusion and mass transfer (Mäenpää et al., 2011).We cannot fully explain this phenomenon based on the data obtained in the present study.A plausible cause could be the faster diffusion of PAHs with lower molecular weight within the peat particles, as was shown for diffusion in polymers (Rusina et al., 2010).To our knowledge, no data on diffusion coefficients for PAHs are available for natural organic particles, but polymers have been identified as a reasonable synthetic proxy (Mäenpää et al., 2011).The main effect of such intraparticle diffusion is the slow movement of PAHs toward sorption sites within the particle, which frees up the readily available outer sorption sites that quickly become occupied by the dissolved PAHs.Consequently, the total PAH concentration within a particle rises.This process slows down as the intraparticle diffusion rate decreases.When using loaded peat in bioassays, one should remember that the kinetics of these sorption processes also apply to desorption.The relatively small fraction of PAHs absorbing into the peat within the last weeks of the loading period can be expected to be desorbing at a low rate, thus less bioavailable (Cornelissen et al., 2005).In natural sediments, this is referred to as the slowly desorbing fraction of POPs (Werner et al., 2010).
The exposure duration in sediment-water bioassays is usually relatively short (<30 days; OECD, 2004OECD, , 2007)), therefore it is unlikely that the fraction of POPs with a slowly increasing concentration in the peat would influence the assessment outcome.The equilibrium loading approach presents an additional advantage for multiple testing as long as the peat is stored with the PE pouch containing the POP reservoir.In this case, the storage time can vary without affecting the exposure levels because the more readily desorbing fraction is constantly replenished and kept at saturation.Conventionally spiked peat, on the other hand, would be expected to show decreasing chemical activities and thus decreasing toxic potential over time.
Degradation of peat organic matter could partly explain a rise of PAH concentrations, even after several weeks of equilibration time, as observed in Batch B (Figure 3).The loading vessels and peat were not sterilized, thus the microbial activity would be expected to break down organic matter, especially given the favorable experimental temperatures for microbial growth (Bergman et al., 1999;Pind et al., 1994).As demonstrated in multiple studies by Hawthorne et al. (2006Hawthorne et al. ( , 2007Hawthorne et al. ( , 2011)), an increasing degree of condensation in organic matter, as a result of increasing degradation, can lead to higher affinities and thus sorption capacities for particularly POPs with a planar structure, such as PAHs.However, to the best of our knowledge, the impact this could have on POP-sorption within a single, degrading batch of organic material has not yet been quantified.Despite this uncertainty, the impact on the chemical activity in our equilibrium partitioning-based loading approach can be expected to be minimal.As long as the PE bag remains in contact with the peat slurry, all compartments in the system (water and peat) are kept at saturation, and therefore at the compound-specific a max .An increased sorption capacity due to degradation is balanced by an increased release of solubilized PAHs from the crystalline reservoir.
The difference in POP sorption behavior between the two peat batches used in the present study (originating from the same supplier) highlights the advantage of an equilibriumpartitioning based loading approach.If instead the target chemical activity had been calculated based on literature data on partition coefficients, the loaded sediments/peat would have shown different actual chemical activities depending on the peat subsample used.This aligns well with the aforementioned two-to fourfold difference in PAH solubility data calculated from literature values (Figure 3), which would correspond to an equal difference in actual chemical activities of PAHs at a given equal concentration.Due to this high variability of organic carbon quality and the resulting variations in published K OC values, the predictability of chemical activity-based sediment POP concentrations and organic carbon content is often insufficient to reliably predict contaminant bioavailability to organisms (McGrath et al., 2019;Neff et al., 2005).To circumvent the difficulties in predicting the bioavailability within a spiked sediment, empirically determined partition ratios (K d ) for a particular test sediment have been suggested, for example by Neff et al. (2005).Using specific K d values allows the bioavailable share of the spiked concentration to be calculated, helping to increase the comparability of results between different chemicals and test sediments.Even though obtaining K d values is time-consuming, it is necessary due to the strong impact that differing K d values could have on results obtained from sediment-water bioassays.With the approach to load sediment presented in our study, the determination of K d and sediment preparation can be combined while monitoring the saturation of the loaded peat.Another, sometimes overlooked, aspect in spiked sediment tests is the equilibration time required, especially for hydrophobic compounds.To introduce the test chemicals into the sediment, they are usually coated onto quartz sand particles (OECD, 2004(OECD, , 2004) ) or onto the walls of the test vessels (USEPA, 2000).It has been recommended that the spiked sediment is allowed to age for at least a month to allow the partitioning of the test substance into and within the organic fraction (USEPA, 2000).Due to our loading approach yielding a sediment in which the test substance is immediately associated with the organic fraction, this ageing time can be shortened or omitted.In practice, it is not necessary to monitor the saturation during the loading process if the loading is allowed to continue well beyond the estimated t 95% .
Following common practice with passive sampling methods, the equilibration times established in a pilot experiment or elsewhere can be used, allowing for one or two extra weeks to ensure complete saturation (Wernicke et al., 2022).Should the sediment preparation time be shortened, monitoring POP concentrations during the loading process is recommended to assess whether the equilibrium is reached.

IMPLICATIONS
We demonstrated the feasibility of dosing organic matrices with POPs at a target chemical activity, enabling chemical activity-based exposure experiments in ecotoxicology and environmental risk assessment.The loading approach ensures that the intended chemical activity range is reached in a reproducible and reliable way, regardless of the properties of the loaded matrix.Furthermore, the loading kinetics and resulting saturated concentrations in two peat batches highlight that the solubility of a POP within an organic matrix and therefore its chemical activity at a given concentration are variable and hard to predict.Hence, converting conventionally applied POP concentrations into chemical activities is challenging.Our concept avoids this issue by loading chemicals into sediments and letting the thermodynamic properties of the POPs drive the exposure concentrations.As a result, the organic matter in sediment would have a lower effect on the exposure levels and bioassay outcome, making the biological responses more reproducible and comparable across the studies.These methodological advancements can help assess POP mixture effects and facilitate the regulatory acceptance of chemical activitybased risk assessment by stakeholders.

FIGURE 1 :
FIGURE 1: Schematic illustration of peat loading.A sealed polyethylene (PE) pouch is filled with a polycyclic aromatic hydrocarbon (PAH)-water suspension, keeping the water saturated (at a max ), and the dissolved PAHs can diffuse through the PE into the surrounding water.Once peat particles are introduced to the water, they start absorbing the PAHs.Owing to the constant replenishment from the PE pouch, the sorption continues until the peat is saturated and chemical activity is equal in the peat and the aqueous solution.

FIGURE 2 :
FIGURE 2: Practical implementation of the loading concept described in Figure1.A sealed polyethylene bag, filled with water and crystalline polycyclic aromatic hydrocarbons (left) is placed into larger vessels (center) that can be filled with the target matrix suspension, such as peat (right).ACE = acenaphthene; Flu = fluorene.

TABLE 1 :
Sampling time points for each of the two peat batches

TABLE 2 :
Parameters of the non-linear regression Y = Y max -(Y max -Y 0 )exp(−kt) used to fit the measured concentrations over time a Calculated using Equation (3).b Solubility of PAHs in peat was calculated using KOC values fromSchwarzenbach et al. (