Characterization of microplastics in storm water in Örebro, Sweden

(1)

Characterization of microplastics in storm water in

Örebro, Sweden

Isabelle Karlsson Sjögren

Bachelor thesis (15 hp) Örebro University, 2020

Examiner: Mattias Bäckström Supervisor: Anna Rotander

(2)

3.3 Chemical characterization ... 10 3.4 Black fragments ... 11 4. Discussion ... 12 4.1 Method ... 12 4.2 Visual characterization ... 13 4.3 Chemical characterization ... 14 4.4 Black fragments ... 15 5. Conclusion ... 15 6. Acknowledgements ... 16 List of references ... 17 Appendix 1... 20 Appendix 2... 21 Appendix 3... 26

(3)

Abstract

Microplastic is a widespread pollutant in marine and fresh water systems. A major pathway by which microplastics end up in these systems is via storm water. Storm water is generated as precipitation drain off of impenetrable surfaces like paving. Microplastic analysis of storm water make up a good foundation for better understanding what sources and factors contribute to microplastic pollution in marine and fresh water systems. This study puts emphasis on characterization and quantification of microplastics through visual characterization. As visual characterization is a subjective form of analysis, the characterization was performed based on guidelines in order to minimize the risk of identifying false positives. The concentration of microplastic was found to be higher in the current study than in comparison to larger water bodies and storm water streams in less urban areas. Fragments, i.e. irregular shaped particles with the appearance of being broken from a larger piece of litter, were found to be the most abundant type of microplastics, pointing at littering as a major source of microplastics in storm water.

Key words: Chemical characterization, microplastic, pollution, storm water, visual characterization

(4)

1. Introduction

Microplastic – a general term for describing anthropogenic polymers differing in size, shape and color – is a widespread pollutant in marine and freshwater systems (GESAMP, 2016). The presence of microplastics in aquatic systems and the aspect of their biological effects on aquatic organism with possible trophic transfer has led to a matter of concern in recent years. Moreover, it has led to an aspiration of better understanding the sources and pathways by which microplastic end up in aquatic systems (GESAMP, 2016). To this day, marine systems have been given more attention than freshwater systems – this, despite the fact that studies suggest freshwater systems as major pathways of microplastics to the oceans (GESAMP, 2016; Lebreton et al., 2017; Schmidt, Krauth & Wagner, 2017).

1.1 Definition of microplastic

The definition of microplastics is still somewhat diffuse as no upper or lower limit in size, as well as no characterization in terms of type, shape and color, have found international

consensus. Varying from study to study, lower limits of 1 nm and 1 µm, and upper limits of 2 mm and 5 mm are all used. There is an ongoing debate of how to define microplastics in order to facilitate comparison of studies and prevent ambiguity (Frias & Nash, 2019; GESAMP, 2019; Hartmann et al., 2019). At times when visual characterization is the only method of identification or the possibility of performing a complementary method for verification is limited, having a standardized manner of characterization becomes especially important. (Hidalgo-Ruz et al., 2012).

1.2 Sources and pathways of microplastics

Microplastics are subcategorized into primary and secondary microplastics. Primary

microplastics are defined as small plastic particles being deliberately made in a particular size and shape, while secondary microplastics are fragments or byproducts generated from other sources or processes involving plastic, such as construction work, plastic recycling, road wear and abrasion of tires or littering (GESAMP, 2016). The sources of secondary microplastics just described, together with industrial production and handling of plastic pellet and artificial turfs, are all contributors to microplastics found in Swedish storm water (Magnusson et al, 2016). Storm water is successively a pathway of microplastics to the marine and freshwater systems (Magnusson et al., 2016; Piñon-Colin et al., 2020). Microplastic analysis of storm water is thus of great interest as it provides opportunities to better understand to what degree different sources contribute to microplastic pollution in the recipients (Magnusson et al., 2016).

Road wear and abrasion of tires has been confirmed to be a global source of microplastics in the environment (Kole, Löhr & Ragas, 2017). Together with artificial turfs, road wear and abrasion of tire make up the two sectors that emit the most microplastic in Sweden, generating 2 300-3 900 t/y and 13 520 t/y, respectively (Magnusson et al., 2016). However, it is

important to note that more data on microplastics found in Swedish, as well as international, storm waters has to be collected in order to determine to what degree these emissions actually contribute to marine and fresh water pollution (Magnusson et al., 2016; Kole et al., 2017). Storm water is generated as precipitation drain off of impenetrable surfaces such as paving. (Magnusson et al., 2016). The volume of storm water generated in an area is correlated to the extent of impenetrable surfaces. The larger proportion of impenetrable surfaces, the more storm water is created (Valtanen, Sillanpää & Setälä, 2014a; Magnusson et al., 2016). As a consequence, the extent of impenetrability affects the degree of pollution in storm water; the

(5)

more storm water, the more pollutants are carried to the recipient (Valtanen et al., 2014b; Yonkos et al., 2014).

1.3 Common polymers

Depending on their density relative to water, different polymers are expected to be found in a higher degree in some levels than other. Polymers like polypropylene (PP) and polyethylene (PE) are generally expected to be found near the surface as their densities are lower than the density of water, while polymers with densities higher than that of water, such as

polyethylene terephthalate (PET), polyvinyl chloride (PVC) and polyester (PL) are expected to be found nearer the sediment floor (Cincinelli et al., 2017; Song & Andrary, 1991).

Nevertheless, it should not be excluded that low-density polymers can be found sub surface as their density can be altered (Cincinelli et al., 2017; Song & Andrary, 1991). Likewise, high-density polymers can be found near the surface water due to high flow rates and turbulence (Cincinelli et al., 2017).

1.4 Methods of characterization

Visual characterization of microplastics is an important method of identifying microplastics (Hidalgo-Ruz et al., 2012). Although it is a subjective form of identification, with plenty of room for ambiguity, it is not to be underestimated. In coaction with tactile inspection, the analyst is given an opportunity to evaluate and question similarities, as well as differences, in spectroscopic measures (Karlsson et al., 2019). Karlsson et al (2019) emphasize the

importance of tactile inspection when distinguishing paraffin from polyethylene and polypropylene which otherwise have both visual and spectroscopic similarities.

In addition to visual analysis, it is recommended to have at least one additional method in order to verify the visual characterization and in that way reduce the degree of

misidentification. (Hidalgo-Ruz et al., 2012; Karlsson et al., 2019). Many studies use Fourier transform infrared spectroscopy (FTIR) as analysis method for chemical characterization (Cincinelli et al., 2017; Karlsson et al., 2019).

1.5 Aim and objectives

The aim of this study is to characterize and quantify microplastics of size 50 µm - 300 µm found in storm water in Örebro, and to determine the relative abundance of different types of plastic particles.

Factors like flow rate and precipitation at the times of sampling is taken into consideration upon comparison.

2. Method

2.1 Sampling location and method

This study has analyzed samples obtained from Bygärdesbäcken, a stream carrying storm water from the southern parts of Örebro City (figure 1) to Svartån (Örebro kommun, 2005), which subsequently flow into Hjälmaren – the fourth largest lake in Sweden. The sampling was performed using an in situ filtration pump developed by KC Denmark. The pump was suspended from a bridge and lowered into the water in a horizontal position until fully submerged. Stainless steel filters of mesh sizes 300 µm and 50 µm were used but only the 50 µm mesh filters was analyzed during this study (table 1). The 50 µm filters were pumped through for 10 minutes while the 300 µm filters were pumped through for 1.5 h. Different

(6)

sample volumes were thus obtained. Samples were stored in jars that were lined with aluminum foil.

Precipitation data was collected from the Swedish meteorological and hydrological institute (SMHI). Sample 1 and 3 was preceded by seven and six days without any precipitation, respectively. Sample 2 was preceded by a 14.0 mm rain event the day prior to sampling. A summarized table of the precipitation data is presented in appendix 1.

Figure 1. Bygärdesbäcken receives storm water from area 22 and 34 (Örebro kommun, 2005). The dotted arrow indicates the flow of Svartån to lake Hjälmaren.

Table 1. Flow rate measurement was not performed during sampling of sample 1 and 2, but the observation was made that the water was at a standstill during sampling of sample 1.

Sample ID Mesh size

(µm) Volume (dm3₎ Flow rate _(m3_/s) Date of sampling

1 50 272 N/A 25-09-2019

2 50 445 N/A 03-10-2019

3 50 363 0.15 22-04-2020

2.2 Digestion and extraction

The 50 µm stainless steel filters were thoroughly scraped with a spatula and the filtrate was transferred to a glass jar. To facilitate the scraping, water was gently poured over the filter, careful not to spill any filtrate over the edges.

(7)

The filtrate was digested in order to decrease the amount of organic debris by gently pouring 50 mL of 30% H2O2 over the spatulas and into the glass jar. The jar was covered with

aluminum foil and was let to sit for 1h before transferred to an oven, set to 50°C, overnight. The solution wasfiltered through a 10 µm PP membrane filter in a vacuum funnel. Extraction was performed by transferring the filtrate to a separatory funnel with saturated sodium

chloride (NaCl) solution (1.2 g/mL). The solution was let to separate according to density for ~20h. The sediment and supernatant were finally collected through the bottom valve in separate falcon tubes. The extraction was repeated once. The supernatant was ultimately filtered on to a gridded 45 µm cellulose nitrate filter and the filter was analyzed visually for microplastics.

2.3 Visual characterization

The study followed the recommended definition and classifications of microplastics presented by GESAMP in their report “Guidelines for the monitoring and assessment of plastic litter in the ocean” (2019), with the additional sub classification of “lines” and “pellets” (table 2). Visual characterization was performed using a Zeiss Stemi 508 microscope, mounted with an Axiocam ERc 5 for taking pictures.

Table 2. Adapted table of how to classify microplastic presented by GESAMP (2019) with alterations made for “Lines” and “Pellet”.

Physical category Type Description Additional information

Size Microplastic (MP) <5 mm Largest dimension of the object <5 mm

Shape Fragment Irregular shaped hard particle having appearance of being broken down from a larger piece of litter Foam Near spherical or granular

particle, which deforms readily under pressure and can be partly elastic, depending on weathering state

Film Flat, flexible particle with smooth or angular edges Line Long fibrous material that

has a length substantially longer than its width

Lines are sub classified into filament and fibers – filaments being

substantially larger than fibers

Pellet Hard particle with spherical, smooth or granular shape

Round, primary

microplastics with smooth surfaces is classified as “Spheres”

Color Standardized color

(8)

2.4 Chemical characterization

Chemical characterization was performed using attenuated total reflection - Fourier transform infrared spectroscopy (ATR-FTIR). Due to the practical limitations that small particles are difficult to transfer to the spectrometer and that small particles generate low signals, only particles (fibers excluded) with one dimension larger than 300 µm were chosen for spectral analysis. The particles were matched against a spectral library containing reference spectra for nine polymers (table 3). Due to low signals relative to the reference spectra, the desired match of 98% between the sample and reference spectra was overlooked and the spectra were

identified manually. For identifying a particle as plastic, emphasis was put on the spectrum to have characteristic polymer peaks in the range 2 700 cm-1_{- 3 000 cm}-1_.

Table 3. Names and abbreviation of polymers in ATR-FTIR reference library.

Polymer Abbreviation

Polyethylene PE

Polypropylene PP

Polystyrene PS

Polyamide PA

Polyethylene terephthalate PET Polyvinyl chloride PVC Polyhydroxyalkanoate PHA Polyhydroxybutyrate PHB Poly-L-lactide acid PLLA

3. Results

3.1 Quality assurance/Quality control

3.1.1 Contamination control

Three blank samples were carried through the experiment. The blanks were performed by adding 50 mL H2O2 into a clean glass jar. From there on, the blank samples went through

digestion, filtration, extraction and visual characterization in the same manner as the field samples. A limit of detection (LOD) was calculated for each type of particle based on the blank results (table 4). Blank corrections were performed by subtracting the average count from the respective count in each sample.

Table 4. Average count in blank samples. The LOD was calculated by adding the average count to three times the standard deviation.

Particle Average count LOD

Fiber <300 µm 39 60 Fiber 300 µm - 1 500 µm 31 39 Fiber >1 500 µm 6.0 12 Fragment 23 48 Other types of microplastics* 0 -

(9)

3.1.2 Recovery

Two recovery runs were performed, and the results are presented in table 5. The first recovery was performed by spiking 20 PP-fibers and 20 PE-fragments to a field sample before

transferring the filtrate from the 50 µm stainless steel filter to a glass jar. The second recovery sample was spiked directly to the glass jar containing H2O2 as no representative field sample

was available and a stainless steel filter with mesh size 50 µm was used instead of a PP-filter. During the second recovery run, the filters were analyzed before and after each step, in order to estimate where during the process particles get lost. Particles were found to be lost during each step but mainly during vacuum filtration and extraction (table 6).

Table 5. Recoveries of spiked PP-fibers and PE-fragments in the respective recovery run.

Recovery run Recovery PP-fibers (%) Recovery PE-fragments (%)

1 10 60

2 25 20

Table 6. Summary of where, during the second recovery, particles were lost. Particle # of spiked particles # of particles lost after 1st filtration # of particles lost during transfer to separatory funnel # of particles lost in sediment # of particles lost during extraction and 2nd_filtration PP-fiber 20 0 2 1 12 PE-fragment 20 8 0 0 8

3.2 Microplastic concentration in Bygärdesbäcken

The results are presented as microplastic per cubic meter (MP/m3_{), where MP includes}

fragments, fibers, foams, films, pellets and spheres.

During visual analysis a lower and upper boundary count was established in order to account for uncertainties during identification of the smallest fragments. The lower boundary count is defined as particles that could be identified as plastic when magnified 1.6 - 2.0 times, while upper count is defined as particles that were identified as plastic when magnified 4.0 - 5.0 times. The lower and upper boundaries are, in this study, interpreted as the observed minimum and maximum concentrations in each sample, leaving room for possible

misidentification, but with a good chance that the true concentrations are found within the respective interval (figure 2). The major differences between the upper and lower counts were made up of small transparent fragments.

Average concentration of the three samples were calculated for both lower and upper boundary (table 7). Due to the higher level of uncertainty in the upper boundary, the results that are here after presented are based on the result from the lower boundary counts.

(10)

Figure 2. Bars indicating the interval between the upper and lower boundaries in each sample.

Table 7. Average MP concentrations and standard deviation (Stdev) of the upper and lower boundary.

Average conc. (MP/m3₎ _{Stdev (MP/m}3₎

Lower boundary 1 600 1 100

Upper boundary 2 800 1 800

The highest MP concentration was found in sample 2 (2 800 MP/m3_{, see figure 3), which was}

the sample that was preceded by rain events. No comparison between MP concentration and flow rate was made, due to lack of data regarding flow rate.

The major proportion of microplastics found in the samples were composed of fragments (77%) followed by fibers 300 µm - 1 500 µm (17%). Foams, films, pellets and spheres collectively made up the remaining 5.8 %. A detailed description of the composition within each sample is presented in figure 3.

The annual contribution of microplastics to Svartån was calculated to be 7.7 ×109_MP/year.

Calculations were based on an average flow rate of 0.15 m3_{/s obtained by laboratoriet AVÖ}

(1999 - 2004) and the average concentration of the lower boundary.

During the visual analysis it was found that fragments with two dimensions larger than 300 µm had passed through the 300 µm mesh filter that was placed on top of the 50 µm mesh filter during sampling (figure 4). Data based on 30 measurements on three 300 µm filters were provided showing variation in mesh sizes. The large fragments were found to be within the range of the observed mesh size variation and therefore, the fragments were counted and interpreted within the result. These fragments were later used to perform chemical

characterization with ATR-FTIR. 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 1 2 3 MP /m 3 Sample

(11)

Figure 3. Composition of different types of plastic particles in each sample; 1, 2 and 3. Lower and 1 L 1 U 2 L 2 U 3 L 3 U Total 1500 2900 2800 4600 540 940 Sphere 48 63 45 72 8,3 11 Foam 7,4 7,4 Film 7,4 11 9,0 Pellet 130 170 31 58 Fiber 300-1500 um 380 380 220 220 220 220 Fragment 910 2300 2500 4200 310 710 0 500 1000 1500 2000 2500 3000 3500 4000 4500 MP /m 3

(12)

Figure 4. Fragments of different sizes and colors found in sample 2. On the left, a transparent fragment with dimensions larger than 300 µm.

3.3 Chemical characterization

78 particles were observed to have dimensions larger than 300 µm and out of those, 7 particles (9.0%) were analyzed with ATR-FTIR. Out of the seven particles analyzed, three were identified as PE, two were identified as PP and two were unidentified (figure 5). Pictures of particles are provided in figure 6 and figure 7, together with their corresponding spectrum. Even though the largest particles were chosen for spectral analysis, the signal was found to be very low in comparison with the reference spectra, resulting in low matching percentages.

Figure 5. Pie chart of the particles analyzed with ATR-FTIR.

28%

43%

29%

Unidentified PP PE

(13)

Figure 6. Left: Transparent fragment found in sample 1. Right: Spectrum showing characteristic peaks at ~2800 cm-1_{and ~1400 cm}-1_{for PP.}

Figure 7. Top: Green film found in sample 1. Bottom: Sample spectrum (pink) matching the reference spectrum of PE (black).

3.4 Black fragments

During visual analysis, black fragments were found. Some of them behaved as plastic upon tactile inspection, meaning that they did not fall apart when applying pressure. They differed from the remaining black fragments in the sense that they had sticky texture. However, some particles with similar appearance did break under pressure, giving the impression that it was soot. Due to uncertainty in how to identify them they were counted and sorted into a separate category, which is not included in the result previously presented.

The black fragments were divided into two size fractions, <300 µm and >300 µm. All

fragments in the larger size fraction, 35 particles, were tactually inspected and out of those, 29 fragments (83%) were resistant to pressure and 7 of those fragments (24%) were analyzed using ATR-FTIR. One fragment (14%) had a sufficient match to be identified as PE (figure

(14)

to strong transmittance at ~2 800 cm-1 _{(figure 9). These black fragments were present in every}

sample.

If the black fragments (>300 µm), which were resistant to pressure, were to be incorporated within the results they would constitute 1.9% of the total average concentration. The relative abundance of the other types would be 76 % fragments, 17% fibers 300 µm - 1 500 µm and 5.7% foam, film, pellet and spheres. It is important to recognize a high level of uncertainty in this result as the fragments in the lower size range (<300 µm) are not included as they were not tactually inspected, and the black particles were a lot more abundant in the smaller size fraction (364 particles). Upon rough approximation that 83% of the black particles are resistant to pressure also in the lower size range, the relative abundance would shift, and the black fragments were to constitute 17% of the total average concentration.

Figure 1. The largest black fragment, found in sample 2, with its corresponding spectrum to the right. The spectrum is well matched to the PE-reference spectrum.

Figure 2. Potential tire wear particle.

4. Discussion

4.1 Method

It is clear that, after two recovery attempts both yielding a total recovery of less than 40%, there is room for improvements during the sample preparation. Furthermore, the recovery runs also indicated a lack of repeatability within the method as the recovery rates for each of the spiked particles, PP-fibers and PE-fragments, differed substantially between the runs. The greatest difference in recovery was noticed for PE-fragments which had a 60% recovery in the

(15)

first run and was followed by a 20% recovery in the second recovery run. A major part of the particles was found to be lost during vacuum filtration and extraction and thus, a key

improvement to the method would be to optimize these steps. Nakajima et al (2019) has presented the introduction of a small stain-less steel sieve in order to keep the transferring step between containers to a minimum and it was shown that the recovery rate was

significantly increased. To this method, an introduction of a stain-less steel sieve would mean that the H2O2 digestion and first filtration step would be performed within one compartment

and thus, the first vacuum filtration would be excluded. In addition to possibly increasing recovery rates, this could further lower the level of contamination as the PP membrane filter would not be necessary. The PP membrane filter could though easily be substituted to a filter on non-plastic material.

As concluded by other studies (Cincinelli et al., 2017; Renner et al., 2018; Song and Andrary, 1991), it should not be ruled out that polymers of different densities can be found in levels of aquatic systems other than their typical one, due to alteration of densities or high flow rates and turbulence. In the case of high flow rates and turbulence, which are not unlikely events regarding storm water, high-density polymers - such as PVC (1.4 g/mL) - could be found in surface water, even though more commonly found in sediment samples. By solely using NaCl solution of density 1.2 g/mL, there is a risk that, if present, high density microplastic particles have been lost during the extraction. To incorporate another round of density separation upon sample preparation, using a solution of higher density like zinc chloride (ZnCl2) could be of

interest to further develop the method. In addition to the possibility of including more types of polymers, ZnCl2 has shown to yield better recoveries than NaCl during density separation for

several types of polymers in sediment samples (Townsend et al., 2019), including PE and PP that were confirmed to be present in the samples

4.2 Visual characterization

The average MP concentration was determined to be 1 600 MP/m3_{(lower boundary). Bearing}

in mind both the low recoveries and the possible loss of high-density polymers, there is a chance that the results presented are underestimated. This needs to be counterbalanced to risk of identifying false positives during visual characterization, as the level of uncertainty

increases as particles get smaller (Song et al., 2015). Taking both sides into consideration, the choice was made to present the result of the lower boundary count in order to keep the risk of overestimation to a minimum.

Previous studies in Sweden have mainly been focused on analyzing microplastic in marine and larger fresh water systems rather than smaller streams of urban storm waters. Upon comparison with other studies one should identify differences in sampling locations, as well as methods and the way that results are presented, since these are factors that can vary highly among studies. When comparing the average concentration of 1 600 MP/m3_{to other studies}

performed by Örebro university (Rotander & Kärrman, 2019; Vigren, 2019), it is important to notice the differences in sampling locations. Both studies from Örebro university presented lower average concentrations (5.5 MP/m3_{and 14.6 MP/m}3_{) but the sampling was performed}

in larger water bodies like rivers and lakes where the concentrations is expected to be lower as storm water streams is a pathway of microplastics ending up in recipients (Magnusson et al., 2016) and the concentrations will be more dilute in larger water bodies (Schmidt et al., 2017). Furthermore, when comparing the results to other studies performed on storm water streams, it is important to again consider differences in sampling location between studies. Jordnära miljökonsult AB submitted in 2020, a report where samples obtained from storm water streams in Lidköping was sampled and analyzed by Örebro university, using the same sampling equipment as well as close to the same sample preparation as this study. It was

(16)

MP/m3_{and 191 MP/m}3_{. Lidköping is a city roughly one fifth the size of Örebro and thus, it is}

in accordance with other studies(Yonkos et al., 2014) stating that the concentrations can be expected to be higher in more urbanized areas.

This study showed varied result in MP concentrations at the different times of sampling. The highest MP concentration was found in sample 2 (2 800 MP/m3_{) which was the sample that}

followed upon a period of rain events in close proximity to the time of sampling. The correlation was not established on a statistical level, but the observation is in line with previous studies (Lattin et al., 2004; Piñon-Colin et al., 2020; Yonkos et al., 2014). A noticeable difference in MP concentration was observed between sample 1 (1 500 MP/m3₎

and sample 3 (540 MP/m3_{) which both were preceded by a week of no rain events.}

Reservation about the observation that sample 2 contained most microplastics and sample 3 contained the least, should be made with regards to the poor repeatability of the method. A more extensive study, including more sets of samples, is of interest in order to test the correlation between precipitation and concentration.

The annual contribution of microplastics in the size range 50 µm - 300 µm from Bygärdesbäcken to its recipient Svartån, was calculated to be 7.7 × 109_{MP/year. It is}

estimated that Bygärdesbäcken contribute to one third of the storm water pollution ending up in Svartån (Örebro kommun, 2005). The estimation was done in 2005 with regards to heavy metal pollutants and it should not be concluded that the estimation automatically refers to microplastic pollutants. However, there are similarities in that the abundance of both heavy metal pollutants (Valtanen et al., 2014b) and microplastic pollutants (Yonkos et al., 2014; Townsend et al., 2019) in storm water, is affected by land use and land use intensity as well as run off volumes. Noticeable is also the fact that southern district of Örebro has been further urbanized since 2005. The districts contributing to the remaining two thirds of storm water pollution has not been looked at. For even more comprehensive results it should be

considered relevant in future studies to also present the result in terms of weight, if it can be found to be done with ease. This would facilitate comparison, namely between studies focused on different size ranges, as 7.7 × 109_{MP/year could have a different meaning to it if}

it were to be presented in a report focused on microplastics in size range 1 000 µm - 3 000 µm.

The most abundant type of microplastic was found to be fragments (77%). This serves an indicator that littering is a major contributor to microplastic pollution in storm water

(Townsend et al., 2019). In two out of three samples, particles that were identified as “pellet” were found. This was the characterization that was found most suitable as it did not have the appearance of being torn off of a larger piece of plastic, and therefore it was not be

categorized as “fragment”, nor did it have a perfectly spherical shape to be characterized as “sphere“. This called for questioning as pellet are described elsewhere as primary plastics particles in sizes 1 - 5 mm (PlasticsEurope, 2017). As the guidelines used during this study did not specify any size limits, other than that microplastics are plastic particles smaller than 5 mm, the decision was made to keep the particles in question categorized as pellet, as it was found to be the most suitable match. This reflect ambiguities that can arise during visual analysis.

4.3 Chemical characterization

Out of the seven particles that were analyzed with ATR-FTIR, two were found to be unidentified, the rest was identified as either PP or PE. The unidentified particles showed transmittance at wave number 2 700 cm-1_{- 3 000 cm}-1_{but remained unidentified as}

characteristic peaks in the fingerprint region could not be identified, indicating that the reference library might have fallen short with its 9 reference spectra. Due to practical

(17)

particles were biased towards particles with at least one dimension larger than 300 µm. Considering the biased selection and small number of tested particles, it is advised that these results are interpreted with caution. In this report, the result of the chemical characterization is seen more as an estimation of how reliable the visual characterization was, rather than seen as a representative distribution of polymers. A method of chemical characterization, where bias has less influence, would have been desired. GESAMP (2019) present µFTIR and Raman spectroscopy as suitable methods for microplastics of smaller sizes.

As previously stated by Hidalgo-Ruz et al (2012), guidelines for visual characterization becomes especially important as chemical characterization is limited, as was the case of this study. Thus, large amount of time was spent on visual and tactile inspection of particles to minimize the risk of false positives. Digestion with H2O2 was also performed in order to keep

the risk of identifying organic material as plastic to a minimum. To really test the applicability of the guidelines on visual characterization of smaller size fractions, tests should be

performed by several laboratories, using the same guidelines, to identify ambiguities as it has been advised that chemical characterization is of more importance as particle sizes get smaller (GESAMP, 2019)

4.4 Black fragments

The distinctive appearance of the black fragments brought on the suspicion that it was fragments emitted through road wear or abrasion of tires, based on the knowledge that this is the sector that emit most microplastics in Sweden (Magnusson et al., 2016). The choice of leaving the black fragments out from the overall result was mainly due to the high level of uncertainty during visual identification and that the appearance of the fragments did not correlate to the guidelines that this report was established around. Additionally, tire wear particles are complex in the sense of chemical structure. They can contain several types of rubber polymers, numerous fillers, stabilizers and more (Rodgers and Waddell, 2013), and they can pick up material from the surrounding environments (Kreider et al., 2010). Upon comparison with other studies (Lee et al., 2007), where rubber has been the main focus, one can spot resemblance to rubber spectra in the transmittance at ~2 800 cm-1_{in figure 9.}

However, peaks at this wavenumber was (Lee et al., 2007), and should not (Mandal et al., 2006), be used for characterization of rubber. Instead, characterization should be based on characteristic peaks in the fingerprint region, none of which could be identified in this study. Yet another uncertainty arises as one of seven the fragments that were analyzed with ATR-FTIR were well matched to the PE reference spectrum, giving no signs of rubber constituents. Again, it is advised that the chemical characterization is interpreted with caution due to biased selection and the small number of particles being tested.

Strengthening the theory that a majority of the fragments could be particles from abrasion of tires is the fact that Magnusson et al (2020) has presented tire wear particles, specifically styrene butadiene rubber, as one of the most abundant constituents among microplastic pollutants in urban environments through atmospheric deposition.

5. Conclusion

MP concentrations were found to be higher in storm water stream in the current study, in comparison to larger water bodies, as well as storm water streams in less urbanized areas. Even though endued with clear guidelines, a level of uncertainty still arose during visual characterization. This was indicated by the differences in the upper and lower boundaries in visual characterization, as well as by the difficulties found in the characterization of pellets and black fragments. ATR-FTIR is not the most suitable method for chemical characterization for microplastics <300 µm as it can produce biased results.

(18)

This study showed that the most abundant type of particle being transported to the recipient via storm water was fragments, indicating littering as a major source of microplastics in storm water streams.

The finding of the black fragments and the inability to identify them, together with the knowledge that tire wear particles are found to be the most abundant type of microplastic in urban atmospheres, should serve as reasons to further investigate the composition of these type of fragments. Ultimately, if the black fragments were found to be tire wear particles, they would not constitute more than 20% of the total composition.

6. Acknowledgements

Many thanks to my supervisor, Anna Rotander, for the support and guidance I´ve been given throughout this project. A big thanks also goes to Clara Svantesson, for help and consultation during the laboratory work.

(19)

List of references

Cincinelli, A., Scopetani, C., Chelazzi, D., Lombardini, E., Martellini, T., Katsoyiannis, A., Fossi, M. C., Corsolini, S., (2017). Microplastic in the surface waters of the Ross Sea (Antarctica): Occurrence, distribution and characterization by FTIR. Chemosphere, 175, pp.391–400.

Frias, J. P. G. L. & Nash, R., (2019). Microplastics: Finding a consensus on the definition. Marine Pollution Bulletin, 138, pp.145–147.

GESAMP, (2016). Sources, fate and effects of microplastics in the marine environment: part two of a global assessment. Reports and Studies 93, 221p.

GESAMP, (2019). Guidelines for the monitoring and assessment of plastic litter and microplastics in the ocean. Reports and Studies, 99, 123p.

Hartmann, N. B., Hüffer, T., Thompson, R. C., Hassellöv, M., Verschoor, A., Daugaard, A. E., Rist,S., Karlsson, T., Brennholt, N., Cole, M., Herrling, M. P., Hess, M. C., Ivleva, N. P., Lusher, A. L., Wagner, M., (2019). Are We Speaking the Same Language? Recommendations for a Definition and Categorization Framework for Plastic Debris. Environmental science & technology, 53(3), pp.1039–1047.

Hidalgo-Ruz, V., Gutow, L., Thompson, R.C., Thiel, M., (2012). Microplastics in the Marine Environment: A Review of the Methods Used for Identification and Quantification.

Environmental Science & Technology, 46(6), pp.3060–3075.

Jordnära miljökonsult AB, (2020). Mindre mängd mikroplast till Kinneviken: Kartläggning av flöden av mikroplast i vatten från Lidköpings tätort. Lidköping: Jordnära miljökonsult AB https://intranat.lidkoping.se/download/18.5ec0539a17115437fe6322a6/1587116060141/Rapp ort%20Mikroplast%20Lidköping%20Jordnära%20inkl%20bilagor%20200414%20total.pdf Karlsson, T. M. Kärrman, A., Rotander, A., Hassellöv, M., (2019). Comparison between manta trawl and in situ pump filtration methods, and guidance for visual identification of microplastics in surface waters. Environmental Science and Pollution Research, 27(5), pp.5559–5571.

Kole, P., Löhr, A. & Ragas, A., (2017). Wear and Tear of Tyres: A Stealthy Source of Microplastics in the Environment. International Journal of Environmental Research and Public Health, 14(10), p.1265.

Kreider, M.L., Panko J.M., McAtee, B.L., Sweet, L.I., Finley, B.L., (2010) Physical and chemical characterization of tire-related particles: Comparison of particles generated using different methodologies. Science of the Total Environment, 408(3), pp.652-659

Laboratoriet AVÖ (1999-2004). Förlängd utsläppskontroll, Svartån – Bygärdesbäcken, punkt 6. Örebro kommun: Tekniska förvaltningen.

Lattin, G.L., Moore, C.J., Zellers, A.F., Moore, S.L., Weisberg, S.B., (2004) A comparison of neustonic plastic and zooplankton at different depths near the southern California shore.

(20)

Lebreton, L.C.M., Van Der Zwet, J., Damsteeg, J-W., Slat, B., Andrady, A., Reisser, J., (2017). River plastic emissions to the world’s oceans. Nature Communications, 8(1), p.15611. Lee, Y.S., Lee, W-K., Cho, S-G,. Kim, I., Ha, C-S., (2007). Quantitative analysis of unknown compositions in ternary polymer blends: A model study on NR/SBR/BR system. Journal of Analytical and Applied Pyrolysis, 78(1), pp.85–94.

Magnusson, K., Eliasson, K., Fråne, K., Haikonen, K., Hultén, J., Olshammar, M., Stadmark, J., Voisin, A., (2016). Swedish sources and pathways for microplastics to the marine

environment. Ivl Swedish Environmental Research Institute. Number C 183, 87p.

Magnusson, K., Winberg von Friesen, L., Söderlund, K., Karlsson, P-E., Pihl Karlsson, G., (2020). Atmosfäriskt nedfall av mikroskräp. Ivl Swedish Environmental Research Institute Number C 511, 23p.

Mandal, N., Chakraborty, S., Ameta, R., Bandyopadhyay, S., Dasgupta, S., Mukhopadhyay, R., Deuri, A.S., Tyre, J.K., (2006) An overview on application of FTIR. Rubber World, 235(3):36–40.

Nakajima, R., Lindsay, D.J., Tsuchiya, M., Matsui, R., Kitahashi, T., Fujikura, K.,

Fukushima, T., (2019) A small, stainless-steel sieve optimized for laboratory beaker-based extraction of microplastics from environmental samples. MethodsX, 6, pp. 1677-1682 Örebro kommun (2005): Dagvattenstrategi för Örebro kommun, Örebro, Diarienr: 2002-736 https://www.orebro.se/download/18.1d8f9a39155628f738416746/1467966299465/Dagvatten strategi+f%C3%B6r+%C3%96rebro+kommun.pdf

Piñon-Colin, T. J., Rodriguez-Jimenez, R., Rogel-Hernandez, E., Alvarez-Andrade, A., Wakida, F. T., (2020). Microplastics in stormwater runoff in a semiarid region, Tijuana, Mexico. Science of the Total Environment, 704, p.135411.

PlasticsEurope, (2017). PlasticsEurope Operation Clean Sweep. Report, Brussels: PlasticsEurope

Renner, G., Nellessen, A., Schwiers, A., Wenzel, M., Schmidt, T.C., Schram, J., (2018). Data preprocessing & evaluation used in the microplastics identification process: A critical review & practical guide. Trends in Analytical chemistry, 111(1), pp. 229-238 .

Rodgers, B., Waddell, W., (2013). Chapter 9 - The science of rubber compounding. In The science and technology of rubber. Elsevier Inc. pp.417-471.

Rotander, A., Kärrman, A., (2019). Microplastics in Södertälje: From Lake Mälaren to the Baltic Sea.

http://naturvardsverket.diva-portal.org/smash/get/diva2:1317939/FULLTEXT01.pdf

Schmidt, C., Krauth, T. & Wagner, S., (2017). Export of Plastic Debris by Rivers into the Sea. Environmental science & technology, 51(21), pp.12246–12253.

(21)

Song, Y. and Andrady, A.L., (1991). Fouling of floating plastic debris under Biscayne Bay exposure conditions. Marine Pollution Bulletin, 22(12) pp.608-613.

Song, Y. K., S. H. Hong, M. Jang, G. M. Han, M. Rani, J. Lee, W. J. Shim (2015). A comparison of microscopic and spectroscopic identification methods for analysis of microplastics in environmental samples. Marine Pollution Bulletin, 93, pp. 202-209. Sveriges meteorologiska och hydrologiska institut (2020).

https://www.smhi.se/data/meteorologi/ladda-ner-meteorologiska-observationer/#param=precipitation24HourSum,stations=all

Townsend, K. R., Lu, H-C., Sharley, D. J., Pettigrove, V., (2019). Associations between microplastic pollution and land use in urban wetland sediments. Environmental Science and Pollution Research 26(22), pp.22551–22561.

Valtanen, M., Sillanpää, N. & Setälä, H., (2014a). Effects of land use intensity on stormwater runoff and its temporal occurrence in cold climates. Hydrological Processes, 28(4), pp.2639– 2650.

Valtanen, M., Sillanpää, N. & Setälä, H., (2014b). The Effects of Urbanization on Runoff Pollutant Concentrations, Loadings and Their Seasonal Patterns Under Cold Climate. Water, Air, & Soil Pollution, 225(6), pp.1–16.

Vigren, D., (2019). Temporal and spatial variations of microplastic concentrations in surface waters in Gothenburg, Sweden. Master thesis, School of science and technology. Örebro:

Örebro University http://www.diva-portal.org/smash/get/diva2:1353049/FULLTEXT01.pdf

Yonkos, L.T. Friedel, E. A., Perez-Reyes, A. C., Ghosal, S., Arthur, C. D., (2014).

Microplastics in four estuarine rivers in the Chesapeake Bay, U.S.A. Environmental science & technology, 48(24), pp.14195–14202.

(22)

Appendix 1

Table A1. Precipitation volumes (mm) the days prior to sampling. Data collected from SMHI, measuring station 95160.

Number of days prior to sapling

Sample 1 Sample 2 Sample 3

0 0.0 0.0 0.0 1 0.0 14.0 0.0 2 0.0 0.0 0.0 3 0.0 1.8 0.0 4 0.0 1.2 0.0 5 0.0 0.1 0.0 6 0.0 9.8 0.0 7 0.0 0.0 0.4 8 0.4 0.0 0.0 9 0.2 0.0 0.1 10 0.0 0.0 0.2 11 12.8 0.0 2.0 12 0.0 0.0 0.0 13 0.5 0.0 0.0 14 3.4 0.0 0.0 15 0.0 0.0 0.2 16 6.3 0.4 0.0 17 1.6 0.2 0.0 18 3.6 0.0 0.0 19 4.5 12.8 0.0 20 4.4 0.0 4.3 21 11.0 0.5 0.0 22 4.2 3.4 0.0 23 4.4 0.0 0.0 24 8.5 6.3 0.0 25 0.7 1.6 0.0 26 0.0 3.6 0.0 27 12.7 4.5 0.0 28 0.0 4.4 0.0 29 0.0 11.0 0.0

(23)

Appendix 2

Photos taken during sampling and sample preparation.

Figure A2.1. Flow rate measurement.

(24)

Figure A2.3. Sampling site. Filtration pump suspended from the bridge.

Figure A2.4. Filtration pump from KC Denmark.

(25)

Figure A2.6. Rinsing 50 µm mesh filter with spatula and transferring filtrate to glass jar.

(26)

Figure A2.8. Vacuum filtration onto a 10 µm PP membrane filter.

Figure A2.9. Filtrate was transferred from membrane filter to separatory funnel with the aid of a spatula and NaCl solution.

(27)

(28)

Appendix 3

Table A3.1. Concentrations in MP/m3_. Fragment Fiber

<300 µm Fiber >300 -1500 µm

Fiber

>1500 µm Pellet Film Foam Spheres

1L 906 379 <LOD 129 7.4 7.4 48 1U 2321 <LOD 379 <LOD 165 11 7.4 63 2L 2461 218 <LOD 16 23 2U 4187 <LOD 218 <LOD 29 9.0 36 3L 307 220 <LOD 8.3 3U 714 <LOD 220 <LOD 11

Table A3.2. Color distribution, lower boundary count. All samples included.

Color Fragment Fiber <300 µm Fiber >300 µm -1500 µm Fiber >1500

Pellet Film Foam Sphere

Transparent/ White 1238 296 21 49 1 2 33 Black/Grey 51 32 5 3 Blue 99 28 1 Green 28 5 1 Red 23 6 Orange 9 Yellow 33 10 Pink 27 3 Purple 1

Table A3.3. Color distribution, upper boundary count. All samples included.

Color Fragment Fiber <300 µm Fiber >300 µm -1500 µm Fiber >1500 µm

Pellet Film Foam Sphere

Transparent/ White 2433 152 296 21 71 6 2 49 Black/Grey 80 14 32 5 7 Blue 131 3 28 1 Green 38 3 5 1 Red 29 6 Orange 18

(29)

Yellow 42 10

Pink 50 3

Characterization of microplastics in storm water in Örebro, Sweden