Effects of sediment manipulation on freely dissolved concentrations of hydrophobic organic chemicals

freely


Introduction
The freely dissolved concentration (C free ) of hydrophobic organic chemicals (HOCs) in pore water of sediments and soils has been shown to better predict bioaccumulation and toxicity of HOCs in benthic organisms than total, solid phase-associated concentrations (Lydy et al., 2014).Hence, C free is considered the most relevant exposure metric for benthic communities (Mayer et al., 2014) and an essential parameter in the risk assessment and management of contaminated sediments and soils (Greenberg et al., 2014), e.g., when judging remediation necessity and effectiveness.
C free of HOCs can be determined through partitioning-based, non-depletive sampling with a specific polymer ('passive sampling').The polymer is exposed to a sediment or soil slurry and HOCs present in the sample will partition from the pore water into the polymer, from which they can be extracted and analyzed.Passive sampling can be performed in the field (in situ) and in the laboratory (ex situ).The latter approach is often preferred, because it is faster, less laborious, and cheaper as compared to in situ sampling (Ghosh et al., 2014).Ex situ passive sampling requires that a sediment or soil sample be collected in the field and then transported to the laboratory for determination of C free .In the laboratory, the sample will be stored and possibly treated ('manipulated') before starting the passive sampling measurements.Storing can be done at room temperature, in a refrigerator (4e7 C), or in a freezer (À20 C), and samples may be manipulated by homogenization, sieving, drying and/or grinding.
Degradation of chemicals during storage should be minimized, such that the concentrations occurring in the field are maintained.Therefore, storing sediment or soil samples at room temperature is discouraged (unless temperatures in the field are close to room temperature), as microbiological processes may intensify at higher temperatures.For this reason, storing samples in a refrigerator is common practice, but storing in a freezer will be even more effective in preventing any degradation of chemicals.For practical reasons, the practitioner may consider freeze-drying if the samples are frozen, because processing the resulting dry, powdered samples is more convenient than processing wet samples.After all, dry samples do not require dry weight determinations and are more simple to handle when preparing passive sampling equilibration systems (Jonker et al., 2020).To remove any lumps created during the freeze-drying process and to homogenize the sample, grinding in a mortar is a suitable and often applied final step in sample treatment.Alternatively, sieving may be performed with dried samples for homogenization purposes.This technique may however also be applied to wet sediment and soil samples, in order to remove e.g., stones, twigs, and leaves (Jonker et al., 2020).
Removing larger objects, particularly stones, is critical, because these may damage the passive sampler (in particular fragile Solid Phase Micro Extraction (SPME) fibers), which is applied for the C free determination (Jonker et al., 2020).Sample manipulation by gentle sieving or manually picking out these objects ('size separation') thus is inevitable.Because C free of HOCs is controlled by sorption of the chemicals to the organic carbon fraction, which is primarily associated with the fine fraction and not with stones, removing larger (no/low-carbon) objects probably will not substantially affect the metric.Whether the other sample storage and manipulation steps (i.e., freezing, freeze-drying, and grinding) can affect C free of HOCs is largely unknown.Information on any effects is crucial though, since the objective of ex situ passive sampling often is to determine a C free , which represents field conditions.In contrast to the 'size separation' step, freezing, freeze-drying, and grinding are not essential for a proper C free determination and if the metric would be substantially affected by any of these treatments, they should not be applied.
The present study was performed to investigate if sample manipulation by freezing, freeze-drying, and grinding has an effect on C free of HOCs in sediments.Five field-contaminated sediments were sampled and subsequently manipulated in the laboratory in different ways, after which C free of native polycyclic aromatic hydrocarbons (PAHs) was determined using SPME fibers.
Polydimethylsiloxane (PDMS)-coated, disposable SPME fiber (glass fiber core diameter 110 mm, PDMS coating thickness 28.5 mm) was obtained from Poly Micro Industries (Phoenix, AZ, USA).Prior to use, the fiber was cut into pieces of 5 cm length, which were washed three times for 30 min each with methanol and three times for 30 min each with Milli-Q water, respectively.

Sediments
Using a single manual grab sampler, the upper ~10 cm surface layer from five different sediments in the Netherlands was sampled.An overview of the characteristics of these sediments (sampling location, texture, likely PAH source, and organic carbon content) is presented in Table 1.Bioavailability in several of these sediments has been studied before (Muijs and Jonker, 2012).

Sediment treatments
Sediments were transported to the laboratory in 5 L buckets.
Here, the PAH-containing sediments were gently sieved (2 mm), homogenized with a mechanical mixer, and stored in a refrigerator.From each bucket, four subsamples were taken and placed in 200 mL glass jars.One jar served as 'untreated, reference' sample and was used for PAH C free determinations (see below) as such.The second jar was placed in the freezer (À20 C) for two days, after which it was thawed and used for C free determinations.The third jar was placed in the freezer for two days and then freeze-dried during 28 h.Due to the limited availability of sediment 5, this treatment was omitted for this sediment.The fourth jar was also frozen and freeze-dried, but the resulting dry material was additionally mortared manually for exactly 3 min, using a mortar and pestle.A schematic representation of these procedures is shown in Fig. 1.

C free determinations
C free in all samples was determined with SPME fibers in fourfold.The determinations were performed several years ago, when a standard protocol was not available yet.Therefore, the procedure slightly differed from the recently-published standard procedure (Jonker et al., 2020).Sediment samples representing 2 g of dry weight were brought into 7 mL amber glass vials and Milli-Q water containing 200 mg/L sodium azide was added to the level of 7 mL, leaving sufficient headspace for proper mixing.SPME fibers were added and the vials were equilibrated on a rock and roller apparatus for 4 weeks at approximately 30 rpm.The fibers were then collected, cleaned with wet paper tissue, placed in autosampler vials containing acetonitrile and internal standard (2methylchrysene), and extracted by vortexing for 1 min.PAH concentrations in all sampler extracts were analyzed by HPLC as described before (Jonker and Muijs, 2010).

Effects of sediment manipulation on C free
The results of the C free determinations in the sediments are presented in Fig. 2. The Figure shows the effects of the different treatments (i.e., freezing, freezing/freeze-drying, and freezing/ freeze-drying/grinding) on the C free of the chemicals in a specific sediment, by normalizing the C free values to those quantified in the untreated, reference sample.As such, the effect of a specific treatment is expressed as a factor, representing the change in C free as a result of the treatment.In most of the cases, clear effects are observed, many of which are statistically significant (paired t-tests; P < 0.05); however, the effects are different for the different sediments, chemicals, and treatments.The largest significant effects occurred for the 3-ring PAH phenanthrene as a result of the 'extensive treatment' (freezing/freeze-drying/grinding): the C free of this chemical increased up to 16 times in sediment 1.For the 6-ring PAHs, C free in all sediments significantly increased by a factor of approximately 1.5e3 as a result of this treatment.
The effects of freezing and freezing/freeze-drying on C free are much smaller as compared to the effects of the extensive treatment, and are primarily observed for the 3-and 4-ring PAHs (phenanthrene e chrysene), for which increases in C free with up to a factor of 4 are observed.For the 5-and 6-ring PAHs (benzo[e]pyrene e indeno[123-cd]pyrene), hardly any effects of these treatments are visible (barring several exceptions, in particular in sediment 3, where significant increases up to a factor of 1.6 are seen).However, although small, it is important to stress that effects are actually observed for these treatments.After all, this indicates that simply freezing (and freezing/freeze-drying) can cause an increase in C free .In the presently investigated sediments, the maximum increase is a factor of four, but larger increases in other sediments or soils cannot be excluded.Interestingly, in most cases, the impact of both treatments was similar.This suggests that freezing, and not the freeze-drying process, is responsible for causing the increase in C free , as freeze-drying in addition to freezing generally did not further increase C free .
Considering the differences in effects observed for different sediments, both in intensity and in terms of 'PAH fingerprint' (i.e., relative differences between the different PAHs), the effects of the treatments will most probably be related to the OC content or type (i.e., PAH source -soot, diesel, creosote, coal, or oil) of the sediment.In other words, the observed differences most likely will be caused by differences in the way and the extent of binding of the chemicals to the matrix.This implies that by simply freezing (and thawing) a sediment, the binding strength of (3 and 4-ring) PAHs can be decreased.The change appears irreversible, because after freezedrying and resuspension in water, similar C free values, thus sorption strengths, were determined.Generally, differences in (OCnormalized) HOC sorption affinity for soils and sediments are explained by differences in the chemical nature of the matrix, such as the degree of aromaticity or the content of black carbon or oil (Cornelissen et al., 2005).However, a change in the chemical composition as a result of freezing is impossible.Therefore, the sorption matrix most likely will be physically changed, e.g., due to structural 'damage' caused by ice crystals, which may break open the matrix and expose inner sorption sites, enhancing the potential for PAH desorption and consequently increasing C free .Apparently, the influence of this hypothesized structural change is the largest for the relatively less hydrophobic (3-and 4-ring) PAHs.The exact cause of this chemical size/hydrophobicity-dependent behavior is unclear.Presumably, grinding has a similar (physical) effect, leading to an increased desorption surface or reduced diffusion path lengths for the sediment-bound PAHs that are sampled by the polymer.Previously, Ter Laak et al. (Ter Laak et al., 2007) also observed a reduction in the sorption of 3-and 4-ring, but not 5-and 6-ring PAHs to a specific soil (containing weathered tar particles as the PAH source), after grinding the sample.Because the uptake kinetics of all PAHs in SPME fibers were higher in the ground sample, the authors explained this phenomenon by differences in redistribution kinetics and hypothesized that for the less hydrophobic PAHs equilibration processes might still be ongoing.
In the present study, grinding, even though performed for only 3 min, generally had a larger impact than freezing (up to a factor of 4).Assuming a physical effect, this observation could be explained by a higher level of energy introduced.Therefore, here a relation with the PAH source would be expected even more, as crushing e.g., coal particles would be expected to enhance PAH availability more than grinding (liquid) PAH-containing creosote or diesel.However, although the exact PAH source was not determined for the test sediments, this characteristic is unlikely to explain the trends in the observed manipulation effects, as sediments 2 and 3 most probably have the same PAH source (they were sampled from the same canal, only 2e3 m apart), but display different 'PAH fingerprints' (see Fig. 2), thus responded to the treatments differently.It should be mentioned though, that sediment 2 has a 10 times lower OC content than sediment 3, and much higher (20e90 times) C free values (data not shown).This sediment therefore has an inherent high PAH availability and the treatments may have been less effective in increasing it further.Actually, this hypothesis may play a role in explaining the effect of the extensive treatment observed for several 3-and 4-ring PAHs, as the magnitude of the effect is inversely related to C free of these PAHs in the untreated sediments (see Fig. 3).For the other PAHs and treatments; however, clear relationships as those shown in Fig. 3 are not observed.Likewise, a relationship between the magnitude of the observed effects and OC content is also absent: the largest effects are found for sediment 1, having the second lowest OC content, whereas generally the lowest effects are observed for sediment 2, which has the lowest OC content.Unraveling the exact mechanisms underlying the increased C free values would therefore require additional sediment characterization analyses, such as quantifying the black carbon and NAPL content, (change in) particle size distribution, total solid phase-associated PAH concentrations (enabling the calculation of sorption coefficients), and/or PAH desorption kinetics.However, such analyses were beyond the scope of the present study, which merely aimed to identify possible effects of manipulations on C free and to provide practical advice on whether or not to apply these manipulations.

Implications
The present study demonstrates that certain storage conditions and sediment manipulation steps can significantly increase C free of HOCs.Although observed in the laboratory, the effects may also be environmentally relevant.After all, certain soils and sediments may get frozen in the winter.Admittedly, in the field the matrices generally are not suddenly exposed to À20 C and may freeze more 'gradually', but the formation of ice crystals, which hypothetically may damage the OC matrix, will occur anyway.In this respect, freezing in the field does not differ from freezing in a freezer and upon defrosting in spring, the availability of native contaminants may be enhanced.Whether or not repetitive (yearly) freeze-thaw cycles are able to progressively increase C free of HOCs remains to be investigated.Likewise, the present laboratory results may be extrapolated to environmental manipulations or geo-engineering activities, such as rolling, heavy truck use or perhaps even dredging, which all induce physical stress or crush the matrix.Possibly, these activities are therefore also able to affect the availability of HOCs in the respective soil or sediment systems.
Regarding ex situ C free determinations in the laboratory, the implications of the present results are straightforward.Soil or sediment samples taken from the field should not be stored in the freezer, but kept in a refrigerator.Further, they should not be freeze-dried or ground.All these sample treatments may potentially increase the availability of target chemicals and the final results may thus not reflect actual conditions in the field.This way, risks of the contamination may be overestimated, (partly) nullifying the merits of determining C free in the risk assessment procedure (Greenberg et al., 2014).

Declaration of competing interest
The author declares that he has no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 2 .
Fig. 2. Changes in the freely dissolved concentrations (C free ) of PAHs, as a result of different sediment manipulations: freezing (white bars), freezing/freeze-drying (gray bars), and freezing/freeze-drying/grinding (black bars).All effects are expressed as a factor, relative to the C free values in the untreated, reference sediments, as indicated by the horizontal dashed lines.Gray bars are missing for sediment 5, due to the limited availability of this sediment.Error bars represent standard deviations of replicated measurements.Asterisks indicate statistically significant differences as compared to the untreated, reference sediments.

Fig. 3 .
Fig. 3. Relationship between the increase in the freely dissolved concentrations (C free ) of three PAHs in the five test sediments as a result of the extensive treatment (freezing/ freeze-drying/grinding) and the chemicals' C free in the untreated, reference sediments.

Table 1
Characteristics of the Dutch field sediments investigated in the present study.
a Determined according to methods described in (Van der Heijden and Jonker, 2009).