Measuring microplastics in mussel tissue with fluorescence and stimulated Raman scattering

(1)

MSc Chemistry

Analytical Sciences

Master Thesis

Measuring microplastics in mussel tissue with fluorescence and stimulated

Raman scattering

by

Thijs Oussoren 11207299

August 2019

Supervisor:

Examiner:

Liron Zada MSc Freek Ariese dr. Assoc. Prof.

Second reviewer:

Govert W. Somsen Prof. dr.

Biophotonics and medical imaging/

Vrije Universiteit Amsterdam

(2)

2

Preface

During my research project at the department “biophotonics and medical imaging” I did research on the accumulation of microplastics in the mussel through exposure experiments. Although it was intended to measure microplastics with SRS, an unconventional measurement method was used due to unforeseen circumstances.

I want to thank dr. Assoc. Prof. F. Ariese and MSc. L Zada for giving me the opportunity to do my project at the laserlab and for putting a lot of time into guiding and helping me during my project. I also want to thank the fellow students, doing also their research project at the department, for their pleasant conversations.

I hope you will enjoy reading the report! Thijs Oussoren

(3)

3

Abstract

Over the years plastic waste has become a worldwide environmental problem, due to a constant supply of plastic waste into the environment. An enormous amount of these plastics end up in the oceans and on the beach were these plastics are broken down into very small pieces called microplastics. Although scientists are nowadays very well aware of these particles it is still not very well understood what the effects of microplastics are on the environment and on the health of humans and animals. Studies have confirmed that these microplastics can be found in the atmosphere, oceans and even your drinking water. Biota samples are taken to measure the accumulation of microplastics in marine life for toxicological effects and as indicator for environmental pollution in oceans, rivers, lakes etc. Also new detection methods for microplastics are in development like stimulated Raman scattering (SRS) microscopy.

The mussel is very efficient in accumulating particles from the water and is used as indicator for different kinds of (chemical)pollution. In this research thesis a spectroscopic technique new to this field, called Stimulated Raman Scattering, is used to measure microplastics from mussel samples obtained from laboratory exposure experiments and mussels obtained from the sea. However, due to some technical problems with the SRS setup, SRS was replaced with fluorescence spectroscopy (one of which was the Varioskan Flash) for measuring the sample obtained from laboratory exposure experiments. In the sample preparation step, the mussels were digested using the enzyme Proteïnase K. As an alternative method, hemolymph was extracted from the posterior adductor muscle. An advantage of the hemolymph extraction method, is that in principle only filtration is needed to separate the beads from the rest of the sample.

With regards to the laboratory exposure mussels, reasonable separation could be accomplished for fluorescence microscopic analysis by using fluorescent beads as replacement for microplastic beads. In this preliminary investigation, the blue mussels were exposed to two different concentration of microbeads (7143 and 71430 microbeads/ L) with sizes of 15 micron. It could be observed that the accumulation was higher when the mussels were exposed to the higher concentration of beads. However, after depuration in clean water the results were inconclusive, because of the similar results between the control aquarium and the other aquariums. From the hemolymph results, no uptake of microbeads in the hemolymph was observed for the bead sizes of 15 micron. With regards to the mussels obtained from the sea, no results could be obtained. The field mussels were not comparable, in regards to sample preparation, with the mussels obtained from the exposure experiments. Filtration after digestion was not a suitable method, due to clogging of the filter and an intermediate step, like density separation had to be introduced. Although this solved the filtration problem it created a new problem for measurement with SRS, due to a white substance on the filter that buried the microplastics and distorted the measurement. This problem is not yet solved.

(4)

4

1. Introduction

1.1. The plastic problem

Since the mid-20th century plastics gained an increasingly important role in every day’s aspect of society. Today plastics have become indispensable in our lives, with a world production reaching 350 million tonnes in 2017.[1] Plastic consumption can be found in the packaging industry, construction industry, electronics, transport, agriculture etc. Noticeable uses of plastics in everyday life are, for example, food packaging, plastic bags, toys, bottles and everyday uses of objects in households, schools and offices.[2] Plastics refer to a group of polymer materials like polyvinyl chloride (PVC), polyethylene terephthalate (PET), polystyrene, polyethylene and many more. Every type of plastic material has different purposes for which they are used. Different types of plastics can be modified to create plastics with improved or new chemical and physical properties. This can be done by adding different kinds of additives, fillers and/ or polymers. Another advantage of plastics is that it can be easily shaped in any kind of form and that the material has a very long usable lifespan due to their resistance against biodegradation.[3] However, due to the enormous use of

plastics worldwide and discarding plastic materials after usage, a lot of the plastics ended up in the environment. Because plastics are not broken down quickly in comparison to organic waste (plants, animals etc.) and the constant supply of plastic waste over decades, plastic has become a serious environmental problem worldwide. Over the years an enormous amount of plastics has eventually

accumulated in the oceans reaching an amount of 150 million tonnes of plastic waste drifting in the ocean waters, called the “plastic soup”. An estimated amount of 4.8 to 12.7 million tonnes of plastic litter ends up each year in the oceans (figure 1.1).[4] Marine waste materials are mostly accumulated at the 5 subtropical gyres of the oceans and at certain coastlines. A gyre is a large circular ocean current that pushes the debris to central convergence zones.[5][6] Due to the continuously growing amount of plastic debris in the environment, animals on land and in the sea suffer from the consequences. Marine animals can get entangled thereby losing freedom of movement and, if oxygen is needed by the animals, drown. Plastic materials have been found in stomachs of different animals, like birds, whales etc.. Due to excessive consumption of plastics by animals, they will lose the impulse to eat sufficiently. Additionally, animals and humans can get exposed to chemicals present in the plastic by directly swallowing plastics or can get exposed to plastics via the food chain.

Figure 1.1: Plastic debris in the oceans. Image taken from the European Union-EP.[4]

(7)

7 1.2. Breakdown of plastics in the environment

Although plastics have a high resistance against degradation, eventually they do age under the continuous influence of environmental factors (also called weathering) and will break down into smaller and smaller pieces. Aging symptoms can be recognized by an increase in stiffness, deterioration in strength, embrittlement, discoloration etc. Plastics can breakdown due to the influence of chemicals, microbial degradation, thermal degradation, ultraviolet (UV) induced photo degradation and mechanical abrasion. Generally the UV induced photo degradation with oxygen has a significant part in the degradation process of plastics before microbial degradation takes place. Photo degradation is induced by UV light (e.g. from the sun), whereby alkyl radicals are formed by homolytic cleavage of a C−H sigma (σ) bond. Alkyl radicals react with oxygen into peroxy radicals. The peroxy radicals react with the polymer and release new radicals which in turn react with oxygen to form again the peroxy radicals. The main steps are described in the reaction sequence below in which P stands for a polymer. This process is also called auto-oxidative degradation:[7][8]

Step 1. Initiation: PH → P• and H• Step 2. Propagation: P• + O2 → POO•

POO• + PH → POOH + P• Step 3. Termination: POO• + POO• → POOP + O2

P• + P• → P−P PO• + H• → POH P• + H• → PH

Regarding the weathering of plastics, the marine environment can be separated into zones with different weathering conditions: the beach, the surface water and the deep water/ benthic environment (table 1.1). It was observed that the rate of degradation in the marine environment is significantly slower than on land.[9] A number of factors that influence that rate of degradation can be suggested. The first reason is the lower temperatures of the plastics floating in the sea. The seawater will just cool down the plastic in comparison to plastics on land (and on the beach). The second reason is the exposure of plastics in the sea to the formation of a biofilm and eventually biofouling that will shield the plastic from sunlight. Also colonization and growth of marine biota causes plastics to sink. Beneath the surface, seawater weakens the impact of sunlight (UV light) on plastics due to absorption and reflection on the surface layer of water.[7]

(8)

8

Table 1.1: Comparison of the availability of weathering conditions in the different zones within the marine environment. This table originates from the book “Marine Anthropogenic Litter”.[7]

Weathering agent Land Beach Surface water Deep water or

sediment

Sunlight Yes Yes Yes No

Sample temperature High High Moderate Low

Oxygen levels High High High/moderate Low

Fouling No No Yes Yes

1.3. Microplastics

Exposure to UV light, heat and continuous mechanical abrasion from wave action can cause larger particles to break down into smaller pieces. These plastic pieces can be categorized based on their size in macroplastics, mesoplastics, microplastics and even nanoplastics. In this thesis the following definitions will be adopted, although there is still no consensus in the scientific field about the definition of these different categories (you can also look at figure 1.2 for the definition of these terms). The term Nanoplastics is used for small pieces of plastics that are smaller than 1 μm,[10] microplastics is used for small pieces of plastics that are smaller than 5 mm,[11] mesoplastics is used for pieces of plastics that are smaller than 2 cm and macroplastics is used for pieces of plastics that are bigger than 2 cm. The total amount of plastic particles floating on the surface of the oceans is estimated to be 5.25 trillion particles, weighting 268,960 tonnes.[12] Figure 1.2 shows the comparison between the mass of particles (left chart) and the number of particles (right chart) found in the ocean surface waters. The plastic mass consists for the most part of macroplastics, but the amount of plastic particles is dominated by small and large microplastics.[13]

As shown the concentration of microplastics in the environment is growing and is slowly becoming a global subject of concern by scientists. It is understood that microplastics ends up in nature by two principles, namely by direct deposition of industrially manufactured microplastics and degradation or fragmentation of macroplastics. The microplastics that are industrially manufactured are called primary microplastics. These microplastics are often used in cosmetics and shampoos as exfoliants for example. Sewage treatment plants clean 50% to 90% of the microplastics out of the sewage water, but a significant part still enters the environment via discharged water and partly due to the re-use of sludge (as compost).[14] Microplastics that are formed due to degradation or fragmentation of macroplastic debris are called secondary microplastics. As already described in par. 1.2, due to long-term exposure to UV light and physical abrasion by water movement the plastic debris becomes brittle which facilitate fragmentation into secondary microplastics. Studies have confirmed that microplastics are

(9)

9

everywhere. The particles are found in freshwater, tap water,[15] in the atmosphere[16] and in animals and humans.[17] Because the particles are of very small size they are (almost) not visible to humans, which make their possible negative effects on the environment and on the health of humans and animals life less obvious. It is still not very well understood what the effects of microplastics are on the environment and on the health of humans and animals, but recently it was announced that “unique” research projects will be launched into the effects of micro- and nanoplastics on human health.[17]

Figure 1.2: The comparison between the plastic levels by mass (left chart) and the plastics by particles count (right chart) on the ocean surface waters. Microplastics are smaller than 5 mm (0.33 – 1 mm in yellow and 1 – 4.75 mm in gray color). Mesoplastics are smaller than 2 cm (red) and macroplastics are bigger than 2 cm (blue). Chart is taken from the site “Our World in Data”.[13]

1.4. Quantification and identification of microplastics

It is now known that microplastics can be found everywhere and that research is done worldwide to track the movement and accumulation of these plastic particles throughout the environment. Water, sediment and biota samples are taken from lakes, rivers or oceans to measure the concentrations in the aquatic environment. Biota samples are taken to measure the accumulation of microplastics in marine life for toxicological effects and as indicator for environmental pollution. A call for standardization of sample preparation methods and quantification[18] has been made, but also new detection methods for microplastics are in development like stimulated Raman scattering (SRS) microscopy.[19]

(10)

10

Most samples taken from the environment have to go through sample preparation before quantification and identification can take place.[18] A common practice is to remove or breakdown non plastic material (e.g. by density separation and acids, bases or enzyms) such that only microplastics stay behind on a filter for analysis. The microplastics which are left on the filters can be then quantified and identified. The most common method for quantification is visual counting by making use of a microscope. By using a microscope the color, size and shape can be determined but not the type of polymer. Moreover, it can be very hard to distinguish between a plastic particle and other, non-plastic particles. Due to this problem, only using visual counting can be very inaccurate. To find out if a particle is a plastic particle visual counting is combined with techniques, like Fourier-transform Infrared spectroscopy (FT-IR), Raman spectroscopy or Scanning electron microscopy (SEM). Also other techniques to identify plastics can be used, like Pyrolysis gas chromatography/ mass spectrometry (Pry-GC/ MS).[20] A short explanation of the techniques is given in the paragraph below. In this thesis, the number of microplastics were determined by measuring the fluorescence intensities of the sample, by using fluorescence microscopy and SRS microscopy (method of detection new for microplastics). Fluorescence, Raman and SRS are in more detail explained in Chapter 2 Theory.

IR spectroscopy is based on the absorption of infrared radiation by molecules. Absorption takes place when the vibrational frequency of the molecule matches the frequency of the incoming radiation. In accordance with the frequencies that are absorbed an IR spectra is created. IR spectroscopy is a good technique for identifying microplastics, because plastics do have a very specific IR spectra that is characteristic for each plastic. Before measurement, samples must be dried because of the high absorption of IR radiation by water.[21] SEM is based on scanning the surface of an object using electrons. The atoms/ molecules of the sample interact with these electrons to emit various signals that can be detected. With SEM the size, shape and composition of particles can be determined. In Pry-GC/ MS the sample is burnt in the absence of oxygen. Due to decomposition of the sample (including the plastics) smaller volatile fragments are formed that can be separated by gas chromatography. By using GC/ MS structural information about the plastics is obtained. The disadvantage of this techniques is that the sample is destroyed and no information is obtained about the color, size, shape and number of microplastics present.[20]

(11)

11

1.5. Monitoring microplastics using blue mussels (Mytilus edulis)

Stimulated Raman Scattering microscopy is a technique new to the field for monitoring microplastics by means of identification and quantification. So far, SRS was only used to measure five of the most common manufactured microplastics that can be found in the marine environment. These five plastics are Nylon, Polyethylene terephthalate (PET), Polystyrene (PS), Polypropylene (PP) and Polyethylene (PE). SRS had already been used successfully to measure microplastics from sediment samples taken from the Rhine estuary.[19] Because microplastics also accumulate in marine animals an interesting step is to identify and quantify accumulated microplastics in biota with SRS.

Marine animals that are used as indicators for pollution of harmful chemicals are perhaps also suitable candidates to monitor the microplastics pollution in the marine environment. Exposure experiments with microplastics are already being conducted in the laboratory on these marine animals for ecotoxicological studies or studies focused on accumulation and uptake, depuration pathways.[37] A marine animal that is commonly found around the globe in temperate seas is the blue mussel, also called Mytilus edulis. The blue mussel was one of the first animals that was used to assess the water quality by monitoring the chemical pollution. Blue mussels are suitable for monitoring water pollutions because they provide location-specific information, are easily collected, are tolerant to a broad range of environmental conditions, one mussel provides enough tissue material for analysis and they efficiently accumulate chemical pollutants by filtering large volumes of water over their gills.[22] When enough food is available, blue mussels constantly filter the water to feed themselves. By filtering the water for food they also will accumulate microplastics, chemical pollutants and other contaminants. In figure 1.3 a schematic representation is shown of the pathways for the transport of microparticels inside the mussels. According to the prominent pathway shown in figure 1.3 the microparticles are captured by the gills and via mucus transported to the mouth and the digestive system. Indigestible parts can be excreted as faeces, after passage of the gastrointestinal tract. Non edible parts are not eaten and are excreted before the digestive system as pseudofaeces. The kinetics of the uptake and depuration of contaminants in mussels is often studied by assuming steady state conditions. A complex balance between uptake and depuration of contaminants determines whether accumulation in the mussel will take place.

(12)

12

Figure 1.3: A schematic representation of the prominent pathway for internal transport of contaminants inside the blue mussel.[23]

1.6. Aim of this research project

Eventually, the end goal of the project is to monitor microplastics in coastal water with wild mussels using stimulated Raman scattering. For Stimulated Raman Scattering microscopy measurements with the blue mussels a practical sample preparation step, to extract the microplastics from the mussel tissue, has to be found. Two methods of extraction are being tested: digestion of the whole soft tissue by means of enzymatic digestion and extraction of the hemolymph from the adductor muscle of the mussel. For a preliminary investigation with blue mussels an exposure experiment was done with different concentrations of microplastic polystyrene beads to look for a possible correlation between polystyrene bead concentration and accumulation. The blue mussels used for the exposure experiment are obtained from a mussel farm in Yerseke (Barbé Groep/ Aquamossel). After the exposure experiments the microplastics are extracted from the mussels to measure the ‘type’ of plastic and amount by using SRS. However, due to some technical problems with the SRS setup it was not possible to measure the microplastics by using stimulated Raman scattering. To keep my research project ongoing, I replaced SRS with fluorescence and I replaced PS microbeads with fluorescent PS microbeads. By combining fluorescence with microscopy it was also possible to scan (or map) the whole filter with the same microscopic setup and software also used for SRS. Of course fluorescence can’t be used for environmental samples to measure microplastics and as such this method should be seen as an emergency solution to measure the uptake of microplastics by mussels in a laboratory setup. For the last phase of this research project, field mussels from different locations

(13)

13

were obtained from Rijkswaterstaat. These mussels were already freeze-dried and pulverized and only needed to be digested and filtrated for measurement with SRS.

(14)

14

2. Theory

2.1. Fluorescence spectroscopy

Fluorescence (and also phosphorescence) is the phenomenon in which a molecule emits light with a longer wavelength after absorbing light with a lower wavelength. When UV-light is absorbed by a molecule an electron in a molecular orbital of the ground state is excited to an unoccupied molecular orbital. Various molecular orbitales exist: the bonding orbitals (σ orbitals and π orbitals), non-bonding orbitals (n-orbital) and anti-bonding orbitals (π*

orbitals and σ* orbitals). In an electronic transition an electron is promoted to a non-bonding or anti-bonding orbital by absorption of enough energy. These transitions are generally in the following order:

lower Δ E n→π*

(S1) < π→π* (S2) < n→σ* (S3) < σ→π* (S4) < σ→σ* (S5) higher Δ E

The process of absorption and emission of light is described in a Jablonski diagram. A Jablonski diagram is shown in figure 2.1. The Jablonski diagram is made out of singlet electronic energy states (depicted by S0, S1, S2, etc.), triplet electronic energy states (depicted by T1, T2, etc.) and

vibrational energy levels. After electronic excitation various conversion, non-radiative relaxation and photon emission mechanisms occur. An electron in a singlet excited state can flip its spin by intersystem crossing (ISC) to a triplet state. In a triplet state the electron is parallel to the electron in the molecular orbital from which it is promoted. The triplet state has a lower energy in comparison to its singlet state. Decay from T1 state to S0 (ground state) by emitting a photon

(light) is called phosphorescence. Decay from S1 state to the ground state by emitting a photon is

called fluorescence. Fluorescence is often observed with flat, conjugated molecular structures.

Figure 2.1: A Jablonski diagram. The bold black lines are the electronic energy states and the thin black lines are representatives for the vibrational energy levels. Blue arrows represent absorption of light and the green arrows represent fluorescence. Conversion and relaxation mechanisms are shown with wavy lines in different colors. The diagram is taken from the site “Edinburgh Instruments”.[23]

(15)

15

In this thesis, exposure studies were carried out with fluorescent polystyrene microplastics (Ø 15 µm) on mussels. For each fluorescent material an excitation and emission spectrum can be made. For the polystyrene beads the excitation and emission spectrum are shown in figure 2.2. The excitation spectrum is measured at a fixed emission (415 nm) wavelength and the emission spectrum is measured at a fixed excitation wavelength (365 nm). The concentration of the fluorescent molecules is proportional to the fluorescence intensity of the emitted light at selected excitation and emission wavelength.

Figure 2.2: The excitation (in blue) and emission (in red) spectrum of FluoSpheresTM, polystyrene beads, Ø 15 µm, blue (Ex. 365 / Em. 415). The excitation spectrum is measured at a fixed emission wavelength and the emission spectrum is measured at a fixed excitation wavelength. The spectrum is taken from the site “Thermofisher”.[24]

The instrument Varioskan Flash was used to measure the fluorescence intensities of the polystyrene beads. The Varioskan Flash is capable of measuring well plates of different sizes, ranging from the 6 well plate to the 1536 well plate. A schematic block diagram in figure 2.3 shows the principle of the system for measuring the fluorescence intensity of a sample. In the excitation optics light of selected wavelengths are produced. A xenon flash lamp produces light that passes in the following order a cut off filter, a double monochromator and a bandwidth selector with a used bandwidth of 12 nm. The principle of the double monochromator is explained in figure 2.4. The double monochromator is based on two diffraction gratings in serial separated by an intermediate slit. The grooved surface of the grating reflects each wavelength in a different angle. By rotating the grating the diffracted colors will pass in turn the intermediate slit to the next grating. The double monochromator has the advantage that unselected wavelength that also passes the exit slit (also called stray light) is reduced. After the bandwidth selector the light passes through the measurement optics that creates a high definition excitation beam that is directed to the samples inside the well plates. The fluorescence light emitted from the samples is also collected by the measurement optics to redirect it to the emission optics. The emission optics measures the intensity of a selected wavelength. Inside the emission optics the light will passes again through a double monochromator and a cut off filter to be detected by a photo multiplier tube (PMT).[25]

(16)

16

Figure 2.3: A schematic block diagram for measuring the fluorescence intensity with the Varioskan Flash instrument. The diagram is divided into four parts: 1. the excitation optics produce the excitation light. 2. the measurements optics create a beam directed at the sample and collect the light emitted from the sample. 3. the well plate with the fluorescence sample. 4. the emission optics detects the light of a selected wavelength. Block diagrams taken from the user manual of the Varioskan Flash.[25]

Figure 2.4: A simplified picture showing the principle of a double monochromator. In short, The double monochromator is based on two diffraction gratings in serial separated by an intermediate slit. A double monochromator reduces the stray light passing the exit slit. Image taken from the user manual of the Varioskan Flash.[25]

(17)

17 2.2. Raman spectroscopy

Like fluorescence, Raman scattering is the phenomenon in which a molecule scatters light when being irradiated. For fluorescence, the difference with Raman scattering is that light is first absorbed by an electron to be excited to a higher molecular orbital. Absorption of a photon by a molecule happens when the energy of the incoming photon matches the energy difference between the ground state and the electronic energy states and/ or the vibrational energy levels (The blue arrows in the Jablonski diagram in figure 2.1 represents absorption). In the case of Raman scattering the incoming photon is not absorbed. Instead, the observed Raman scattering is based on the inelastic scattering of light.

The process of Raman scattering is described in the Jablonski diagram in figure 2.5. In Raman spectroscopy, to induce the Raman effect, a laser with monochromatic light in the visible or near-infrared region is irradiated on a sample. The photons don’t have enough energy to excite the electrons to an existing electronic state but instead are immediately scattered after reaching a so-called virtual (imaginary) state. If the scattered photons has the same energy as the incoming photons (no energy exchange takes place) the phenomenon is called elastic scattering or Rayleigh scattering. If energy exchange takes place between the photon and the vibrational levels of the molecule inelastic scattering occurs. The energy difference between the incoming photons and the inelastic scattered photons is called the Raman shift and corresponds to the energy difference between the ground state and the vibrational energy levels. Relative to the incoming wavelength of the radiation, the inelastic scattered light can have a longer wavelength, called Stokes shift. This happens when part of the photon energy is transferred to the vibrational energy levels. When the inelastic scattered light has a lower wavelength, called anti-Stokes shift, the vibrational energy levels transferred its energy to the photon. The Raman spectrum is presented in Stokes shift signals only because anti-Stokes shift signals, that give the same information, are much weaker. At room temperature the fraction of the molecules that are in a higher vibrational state is lower in comparison to the fraction that are in the ground state according to the population ratios calculated by the Boltzmann equation. The intensity of the anti-Stokes signals also decreases exponentially with a constantly lower wavelength of the scattered light according to the Boltzmann equation. Figure 2.6 gives an illustration of a Raman spectrum with Stokes and anti-Stokes shift. The values on the horizontal axis (the units are given in cm-1) are relative to the wavelength number of the Rayleigh scattering and corresponds with a molecular vibrations of the molecule. Raman spectroscopy is not a very sensitive technique, because there is not a large probability to induce the Raman effect. Rayleigh scattering has the highest probability, because most of the molecules are in the ground state and the most probable event is to return to the ground state without energy transfer. The Raman effect (and the signal intensity) is dependent on the change of the polarizability of the electron clouds in the molecule. The polarizability gives the tendency of how well an electron cloud can be deformed.[26][27]

(18)

18

Figure 2.5: A Jablonski diagram describing the Raman scattering process. (A) Absorption of light to an excited vibrational energy level (v=1), a process that happens in Infrared spectroscopy. (B) Rayleigh scattering or elastic scattering. (C) Raman effect in the form of a Stokes shift. An inelastic scattering process. (D) Raman effect in the form of an anti-Stokes shift, also an inelastic scattering process.[26]

Figure 2.6: An illustration of a Raman spectrum with Stokes and anti-Stokes shift. The values (presented in cm-1) on the axis are relative to the wavelength of the Rayleigh scattered light. The Raman spectrum is normally only shown in Stokes shift, because the anti-Stokes shift gives the same information with much weaker intensity signals.[26]

Raman spectroscopy is often used for identification of microplastics after visual sorting to prevent false positives. It doesn’t, however, prevent the problem of false negatives. For a more representative result, a solution is to count every particle in a sample. One way to do this is to map/ scan a whole filter (or a section of the filter) after sample preparation. A motorized stage with on top a filter is moved in small increments under a stationary laser to analyze the filter point by point.[28] However, Raman mapping is rather slow and takes 38 hours to scan 1 mm2 of

(19)

19

a filter.[29] Besides that, Raman spectroscopy is not a very sensitive technique and it also suffers from fluorescence interference due to impurities in the sample or from the microplastic particles. Due to fluorescence, the baseline of the spectrum will be raised with the possibility to conceal the Raman signal.

2.3. Stimulated Raman Scattering

Raman spectroscopy is often used as an extra identification method in combination with visual counting (through a microscope) to prevent wrong determination of particles as microplastics. The particles are counted and observed through the microscope to be identified one by one using spectroscopy, what makes this method time-consuming and biased. Another spectroscopic identification technique for microplastics based on the concept of Raman scattering is called Stimulated Raman Scattering (SRS). In comparison with Raman spectroscopy, SRS has a much higher sensitivity and is capable of circumventing the biased selection that might occur with the use of visual sorting by mapping an area at a fast speed.

As explained earlier, Raman spectroscopy is based on the inelastic scattering of light. A laser with a certain wavelength is irradiated on a sample. Light is inelastic scattered from the sample with a very low probability. The difference in energy between the incoming and the inelastic scattered light is transferred to the vibrational energy levels. If the focus is placed on the stokes shift, scattered light with a longer wavelength will be observed. In a Jablonski diagram the transfer of energy to the different vibrational energy levels (in the case of an amorphous solid) is depicted in figure 2.5. However, SRS microscopy is capable to specifically transfer the energy to a certain vibrational energy level. focus the energy transfer specifically on a certain vibrational energy level. In other words, in SRS the focus is specifically on a certain energy difference between the incoming light and the inelastic scattered light which corresponds to a vibrational energy level. In order to excite a particular molecular vibration two lasers (with an energy difference to each other corresponding to a vibrational energy level) are irradiated on the sample. The first laser is called the Pump beam and the second laser is called the Stokes beam. The Jablonski diagrams in figure 2.7 looks similar to the process as described in ‘spontaneous’ Raman spectroscopy, but actually explains the process for SRS. The blue arrows represents the photons from the Pump beam and the red arrows represents the photons from the Stokes beam. Just like in ‘spontaneous’ Raman spectroscopy, one of the photons from the Pump beam has to transfer a part of its energy to excite the molecular vibration to a higher vibrational energy level. As expected, the photon is scattered with the same wavelength used for the Stokes beam. The intensity of the Pump beam becomes lower (stimulated Raman loss, SRL) and the intensity of the Stokes beam is increased (stimulated Raman gain, SRG). Due to the stimulated vibrational excitation that is generated by the two laser beams, the Raman signal is greatly enhanced.[30]

(20)

20

The change in intensity of both laser beams after stimulated Raman scattering is small in comparison to the intensity of those laser beams. A modulation transfer method is shown in figure 2.8 to detect the change in intensity (Raman signal) of the Pump beam (SRL). By modulating the intensity of the input Stokes beam, the intensity of the output Pump beam will be modulated due to the alternating stimulated Raman loss. The modulated intensity of the Stokes beam will have a stimulated Raman gain. Before detection the modulated Stokes beam is blocked by a filter.

In Comparison to Raman spectroscopy, SRS is able to scan a lot faster. The SRS setup covered in Chapter 3 Method from Zada et al. has a scanning speed of 2.7 min. per 1 mm2. The only downside is that this SRS setup has less flexibility, because it is only able to target one desired vibrational frequency at a time.[19]

Figure 2.7: The two Jablonski diagrams describe the Stimulated Raman Scattering process. The blue arrows represents the photons from the Pump beam and the red arrows represents the photons from the Stokes beam. The difference in energy between the photons of the Stokes beam (with frequency wp) and the photons of the Pump beam (with frequency ws) matches the

vibrational frequency of a chemical bond. By the stimulated energy transfer of the Pump photon to the chemical bond, the photon is scattered with the same frequency as the Stokes photons.[31]

(21)

21

Figure 2.8: Schematic presentation for the modulation transfer method for detection of the change in intensity due to Stimulated Raman Scattering. After irradiation of the sample the Pump beam will have a modulating intensity due to stimulated Raman loss. The modulating intensity of the Pump beam is created by the modulation of the Stokes beam. The diagram is taken from the University website of Harvard, department of Chemistry.[30]

(22)

22

3. Method

3.1. Aquarium exposure setup

A simple aquarium setup for exposure experiments is designed. Blue mussel (Mytilus edulis) is used for the exposure experiments. The aquariums used for the exposure experiments consist of 6 times 2 liter beakers (Sigma Aldrich). The beakers are filled with 1,4 L of demineralized water (demi water). For healthy living conditions of the blue mussels in the aquarium a salt concentration of 32‰ is used. In comparison the North Sea has a salt concentration between 32‰ and 35‰. Artificial sea salt (Red Sea Salt, 4 kg bag) is used and bought online from zeeaquarium-winkel.nl. The temperature of the aquariums is held at a constant temperature of around 16/ 17°C. To maintain a constant temperature the beakers are placed inside a white container filled with water as shown in photos 3.1 and 3.2. The water in the container is cooled due to evaporation by blowing compressed air from the lab onto the water by using a hosepipe as shown in photo’s 3.1, 2 and 4. The room temperature is constant around 21°C and due to evaporation the water temperature in the aquariums is dropping 4 á 5°C.

To prevent the micro plastics from sinking to the bottom of the beaker and to keep them distributed a magnetic stirrer with stirring rod (3/ 3,5 cm) and compressed air from the lab with air stones for bubbles is used. The stirring rate of the magnetic stirrer is in between 500 rpm and 750 rpm. Bubbling by air stones also retains a healthy oxygen content in the water for the mussels to stay alive. With a salinity of 35‰, a pressure of 1 bar and a temperature between 15°C and 20°C, the solubility of oxygen in seawater is between 7,9 mg/L and 7,2 mg/L.[32] The mussels feed themselves by filtering the water. Due to the filtering the mussels can have filtered all the water and the microplastics within 10 min. and due to the filtering create an additional flow in the aquarium. Blue mussels, with a shell length size of around 6,5 cm, can filter the water with a filtration rate of around 10 L/h (aquarium conditions: 3 to 9 mgO2/L, 20‰ salinity and

18,2°C).[33] The mussels will be lying inside a holder, made from chicken wire, so they are not disturbed by the stirring rod at the bottom of the beaker. An illustrated picture is shown in figure 3.1 and in photos 3.3 and 3.4. To prevent microplastic contamination and loss of water from the aquarium by evaporation the beakers are closed off with tinfoil.

(23)

23

Photos 3.1 – 3.4: Installation of aquarium exposure setup for mussels. 1. Picture of the whole laboratory exposure setup. 2. Top view of three exposure beakers in a container with water to regulate the water temperature inside the beakers. 3. Top view of one exposure beaker without foil on top and with a chicken wire holder inside the beaker. 4. Side view of an exposure beaker with chicken wire holder next to the beaker.

Figure 3.1: Schematic presentation of a beaker in the exposure setup.

Air Air Stone Magnetic stirrer Stirring rod Blue mussels Metal wire container

1 2

(24)

24 3.2. Microplastic sedimentation

Before exposure studies a first look was taken at the sedimentation (the tendency for particles in suspension to settle out of the fluid in which they are entrained and come to rest against a barrier) of polystyrene microplastics in the aquarium setup. In the first experiment the aquarium was filled with 1.4 L demi-water. A stirring rod for water movement and an air stone for air support was used. Figure 3.1 shows a schematic presentation (without the metal wire container and mussels) of the setup. A polystyrene microplastic (Ø 20 µm, not fluorescent) was used. In order to examine the sedimentation, we carried out an experiment with two different conentrations, one with a concentration of 714 microplastic/L and one with a concentration of 7142 microplastic/L. 10 mL water samples were taken from the aquarium after 5 min., 1, 3, 5 hours and 1 day.

In a second experiment the metal wire container was added to the aquarium and a concentration of 714285 microplastic/L was used. In the second experiment, two experiments were simultaneously carried out, an aquarium with and an aquarium without a stirring rod. It should be noted that the samples are taken from the middle of the aquarium. The point of water sampling, if the metal wire container is included in the aquarium, is almost the same without the metal container.

For the evaluation of the number of microplastics present, all water samples were filtered through an Anodisc filter. The Anodisc filters were prepared for measurement with SRS according to figure 3.5. The Anodisc filter is placed on a very small water droplet on the microscope slide. Next the cover slide, with a very small droplet on the underside, is placed on the Anodisc filter. To avoid evaporation of water between the two slides, the gap is sealed by covering it with nail polish at the edge of the cover slide. The particles were detected with SRS and the number of particles was determined by using the Matlab software as described in section 3.5. The result of a few water samples are shown in table 4.1 in the ‘Results and discussion’ chapter. Most of the water samples couldn’t be measured due to the unfortunate breakdown of the laser necessary for SRS.

3.3. Extraction of micro plastics from exposed mussels.

Two methods have been considered to extract micro plastics from exposed mussels. The first method is enzymatic digestion of the whole soft body of the mussel (all the tissue except the shell and white posterior adductor muscle) and the second method is the extraction of hemolymph from the white posterior adductor muscle. The protocol used for enzymatic digestion is taken from an article of T.M. Karlsson et al.[34] In the article and in our protocol the proteinase K enzyme is used.

(25)

25 3.3.1. Enzymatic digestion and filtration

The protocol used for enzymatic digestion is as follows. The whole soft body of the mussels will be freeze-dried for 2 to 4 days whereupon the freeze-dried tissue will be ground with a mortar for homogenization or stored in closed petri dishes in the freezer (-20°C) to be homogenized the next day. From a pooled sample of mussels, triplicate subsamples of 0.2 grams are collected after the homogenization step and added to a glass bottle (250 ml) with a plastic screw cap on top. In the next step 15 ml of a homogenization buffer (400 mM Tris-HCl, pH 8, 0,5% SDS) is added and the samples are incubated for 60 min. at 60°C. After incubation a mixture of 8 mg CaCl2 and

7,5 mg proteinase K (3.0 – 15.0 units/mg solids) in 1 ml demi-water is added. The samples are then incubated for at least 2 hours at 50°C, shaken for 20 min. at room temperature and incubated for 20 min. at 60°C. In the next step 30 ml of H2O2 (30%) is added to digest the

remaining tissue overnight at room temperature. The slight change in the protocol is to incubate the remaining tissue with the added hydrogen peroxide overnight at a temperature of 40°C. The chemical ratios used in the protocol will remain the same in respect to the amount of mussel tissue used for digestion.

After enzymatic digestion the samples are filtered by vacuum filtration. The first vacuum filtration setup is shown in figure 3.2. The sample is filtered through an inorganic, Anodisc membrane filter (pore size: 0,2 µm, Ø 25 mm) made from aluminum oxide with a polypropylene support ring. The Anodisc membrane filter is very fragile/ brittle, so it is supported by a grid at the end of the glass tube to prevent rupture of the filter.

Anodisc filter Vacuum pump Support grid for filter

Figure 3.2: Schematic presentation of the vacuum filtration setup.

(26)

26

Because of the very small pore size (Ø 0.2 µm) of the Anodisc membrane filter it needs to be verified that the digested mussel sample will go through the filter. Changes in the protocol may be needed to avoid clogging of the filter. The mussels used for this experiment are bought from the local supermarket (Jumbo, verse Zeeuwse mosselen, large, 2 kg). Dead mussels and mussels with broken shells are thrown away. The mussels that are alive are killed by opening the shells entirely by tearing loose the white posterior adductor muscle. The whole soft tissue is digested according to the described protocol and filtered through the Anodisc filter.

Another filtration method that can be used, when filtration through the Anodisc filter proved challenging, is to filter first through a filter with a bigger pore size to prevent clogging. The sample is first filtered through an hydrophilic Isopore membrane filter (pore size: 5 µm, Ø 25 mm) made from polycarbonate. Instead of a funnel shaped glass tube a cylindrical glass tube is used. After filtering through the Isopore filter the cylindrical glass tube, the Isopore filter and the glass tube with the support grid, is placed as a whole upside down on a new glass tube with a support grid with in-between the new glass tube and the cylindrical glass tube the Anodisc filter. The glass tube with support grid on top is replaced by the funnel shaped glass tube. Back-flushing with demi-water and ethanol should move the residuals (with focus on the microplastics) from the Isopore filter to the Anodisc filter. The back-flushing procedure is illustrated in figure 3.3. Anodisc filter Isopore filter 2 Back-flushing 1 Isopore Anodisc filter Isopore filter 1 2 Back-flushing Move upside-down

Figure 3.3: Schematic presentation of the back-flush procedure in vacuum filtration.

(27)

27

Four samples were made to try this filtration approach and to test if the proteinase K for digestion is necessary:

1. Enzymatic digestion following the protocol described.

2. Enzymatic digestion following the protocol without adding mussel tissue (blank). 3. Enzymatic digestion following the protocol but without adding proteinase K

4. Enzymatic digestion following the protocol but without adding proteinase K and mussel tissue.

3.3.2. Efficiency of filtration using back-flushing

Using back-flushing, the microplastics and residues are collected on the Anodisc filter. In the filtration step some microplastics might be lost or can linger on the polycarbonate filter after back-flushing. For this reason the number of microplastics on the Anodisc filter is not representative for the number in the sample before filtration. To properly estimate the initial number of particles in the sample, the efficiency of the filtration and back-flushing should be known. The efficiency is measured by comparing the number of microplastics collected on the Anodisc filter to the initial amount spiked into the sample before filtration. Fluorescence beads are used for the representation of microplastics. The microbeads used are FluoSpheresTM, polystyrene, Ø 15 µm, blue (Ex. 365 / Em. 415) from the brand Invitrogen by Thermo Fisher Scientific. The number of beads on the filter will be estimated by measuring the fluorescence intensity.

To estimate the number of beads by its fluorescence intensity a calibration curve is made. The concentration of fluorescent beads is 100 times diluted to 10,000 beads/mL from an initial concentration of 1,000,000 beads/mL. The diluted sample (made by adding 10 µL of the emulsion to 990 µL of demi-water for a total of 1000 µL) is used for the calibration curve by filtering the volumes shown in table 3.1. Two calibration curves are made, one based on the polycarbonate filter and the other based on the Anodisc filter.

Table 3.1: The volumes (taken from the 100 times diluted samples) used for the two calibration curves. The (nominal) number of beads is the calculated value in the filtered volume that should be left over on the filter after filtration.

Number of beads Filtered volumes (µL)

200 20 400 40 800 80 1600 160 2000 200 2500 250

(28)

28

After vacuum filtration the filters are laid flat inside a polystyrene six well plate to measure the intensities of a whole well with the instrument Varioskan Flash using the software Skanlt Solftware 2.4.5 RE. The Varioskan Flash was mainly used because of its very fast measuring time of a six well plate (20 min. for one six well plate). of An example of a measured well plate with the Varioskan Flash is included in Appendix 1. The example shows the method for obtaining the fluorescence intensity of one well. Because of the fluorescence background of the material of the well plate some tests were done with aluminum foil and black paint to diminish the background. Instead of the Varioskan Flash (which only measures intensities) the Lumascope 620 was used to capture some fluorescence images of clean filters and filters with fluorescent beads. However, the Lumascope 620 was not useful for the Anodisc filter for quantitative measurements. In Appendix 3 fluorescent images from the polycarbonate filter and Anodisc filter taken with the Lumascope 620 are shown. An image from the Anodisc filter with and without fluorescent beads are shown in photos A3.5 and A3.6. Both images look the same. For an unknown reason the Anodisc filter without beads also has fluorescent dots with the same size as the fluorescent beads. No distinction can be made between the dots to decide whether one is a fluorescent bead or not. In case of the polycarbonate filter the beads can best be observed and counted by using the 4× objective. By using the 10× objective the image can’t be made sharp anymore and by using the 1.25× objective it’s hard to distinguish the beads from each other. In table 4.4 in the ‘results and discussion’ chapter the fluorescence intensities are measured with the Varioskan Flash in a treated well plate. From these results it can be concluded that the Anodisc filter is indeed fluorescent.

To include the standard error four calibration curves (made from different diluted samples) were averaged. These calibration curves are only based on the polycarbonate filters. The 200 times diluted sample is made by adding 60 µL of the emulsion to 11940 µL of demi-water for a total of 12 mL. The diluted sample is used for the calibration curves by filtering the volumes shown in table 3.2.

Table 3.2: The volumes (taken from the 200 times diluted samples) used for the four calibration curves. The (nominal) number of beads is the calculated value in the filtered volume that should be left over on the filter after filtration.

Number of beads Filtered volumes (µL)

100 40 200 80 400 160 800 320 1600 640 3200 1280

(29)

29 3.3.3. Hemolymph extraction

The other method, for extracting microplastics from mussels, is to consider the extraction of hemolymph from the white posterior adductor muscle. The location of the muscle in the mussel is shown in figure 3.4. The extraction of hemolymph from the mussel is less time consuming and cheaper in comparison to enzymatic digestion. To extract the hemolymph, first the mussel should be checked if it is alive. Additionally, the mussel should have closed its shell when the mussel is removed from the aquarium. Then the shell has to be opened slightly to drain the mussel from seawater and to make room for the needle (21G×1”, 0,6×26 mm or 23G×1½”, 0,8×40 mm). A syringe of only 1 or 2 ml is needed for slow and controlled extraction of the hemolymph from the adductor muscle. Hemolymph samples are stored in small glass bottles in the freezer (-20°C). Immediately after extraction, to verify that the extracted fluid is hemolymph, a test can be done by placing a small drop on a glass plate and to let it incubate at high humidity for 20 min. The cells of the hemolymph will stick to the glass plate and can be seen with a microscope. After extraction, the hemolymph is treated with H2O2 (30%) overnight and vacuum filtered through the

polycarbonate filter (figure 3.2).

White posterior adductor muscle

Figure 3.4: Dead blue mussel portrayed from 3 different angles showing the position of the white posterior adductor muscle.

Figure 3.5: Preparation for SRS measurement of the Anodisc filter placed in-between a microscope slide and a cover slide.

Brush doped in nail polish

(30)

30 3.4. Exposure experiments

Two exposure experiments were carried out according to the setup described in paragraph 3.1, photo’s 3.1-4 and figure 3.1. Six aquariums are used for each exposure experiment and named as follows: 1A, 1B, 1C, 2A, 2B and 2C. The experiments were done in duplo (hence the naming of the aquariums 1A, 2A etc.). The exposure experiments were done with two different concentrations and a blank. The following (nominal) calculated number of beads were added to each aquariums.

First experiment:

- Aquarium A: 10,000 beads (7143 microplastics/ L) - Aquarium B: 1000 beads (714 microplastic/ L) - Aquarium C: zero beads (blank, control) Second experiment (with fluorescent beads):

- Aquarium A: 100,000 beads (71429 microplastics/ L) - Aquarium B: 10,000 beads (7143 microplastic/ L) - Aquarium C: zero beads (blank, control)

The microbeads used for the first experiment are made from polystyrene (Ø 20 µm, not fluorescent). The exposure experiment started with 4 mussels in each aquarium and are collected after 24 hours. The mussels were bought from the local supermarket (Jumbo). The microbeads used for the second experiment are FluoSpheresTM, polystyrene, Ø 15 µm, blue (Ex. 365 / Em. 415). The second exposure experiment started with 6 mussels in each aquarium. 3 mussels were collected from each aquarium for measurement after being exposed to microplastics for 24 hours. The 3 mussels left over in the aquariums are exposed for 3 days to clean water without microbeads and were also collected. The mussels were obtained from a mussel farm in Yerseke (Barbé Groep/ Aquamossel). From all the collected mussels hemolymph was extracted. The mussels were digested following the digestion protocol described. Dead mussels were thrown away after hemolymph extraction. The following information about the mussels were collected: shell length, shell weight, wet weight tissue, wet weight mussel and dry weight tissue. The condition index was calculated with the following formula: (wet or dry weight of the tissue / weight of the shell) × 100% and is a measure for the overall health of the mussel.

(31)

31 3.5. Monitoring field mussels

The end goal of the project is to monitor microplastics in coastal water with field mussels using stimulated Raman scattering. To start testing with such mussels, four samples (already freeze-dried and pulverized soft mussel tissue) from different locations were borrowed from Rijkswaterstaat. Two of the four samples (sample 1: 2015004719, sample 2: 2015004716) were used in the laboratory to test the digestion and filtration method again. The sample preparation method with enzymatic digestion and back-flush filtration, as described in ‘Chapter 3 Method’ was used. The Anodisc filter is placed between a microscope- and a cover slide, according to figure 3.5, to prepare for measurement with SRS. The SRS microscopy setup is described in ‘section 3.6. Counting microplastics’.

3.6. Counting microplastics

After digestion and filtration of the sample (mussel tissue, hemolymph or water samples from the exposure) the number of microplastics were counted on the concerning filter. The intention was to use ‘Stimulated Raman Scattering’ (SRS) for detection of the microbeads, but due to the unfortunate breakdown of the laser necessary for SRS the technique could not be used for detection. In the second exposure experiments fluorescent beads were used. As an alternative technique to estimate the number of beads ingested by the mussels, fluorescence was used in combination with the microscope (LSM 7MP, Zeiss, originally used for SRS) to scan the whole surface of the filter. The dry polycarbonate filter was placed on a glass plate and illuminated from below with UV lamp 4 (from CAMAG) with a wavelength of 366 nm. A microscope with a 10× objective was used to scan the filter. A longpass absorption filter of 370 nm (cutoff filter) was used in the setup. After scanning the filters the beads were counted from the image manually and in some cases using the program ImageJ (Fiji). Manual counting was done using software ZEN 2011. Counting with Fiji was done using the option “Analyze Particles” from the “Analyze” menu.

(32)

32

Figure 3.6: Microscope setup for detection of fluorescent beads. 1. Absorption filter. 2. Objective (10×). 3. Illumination (was only used for SRS measurement). 4. Detection. 5. Software (ZEN 2011). From below a UV lamp 4 illuminates a polycarbonate filter on a glass plate. The microscope scans the whole surface of the polycarbonate filter from above.

The SRS microscopy setup that is used in the laboratory for measuring microplastics is shown in figure 3.7. A frequency-doubled Nd: YAG laser generates a 8 ps laser pulses with an 80 MHz repetition rate and a wavelength of 532 nm. The laser is tuned to a desired wavelength by the optical parametric oscillator within a range of 790 – 950 nm and acts as the Pump beam. A second laser with a wavelength of 1064 nm is modulated by an acousto-optical modulator (AOM) at 3.636 MHz and acts as the Stokes beam. The applied laser powers used to irradiate the sample are for the Stokes beam 20 mW and for the Pump beam 10 mW. The Pump beam and the Stokes beam are combined through a dichroic mirror and is irradiated onto a sample by using a laser scanning microscope (LSM 7MP, Zeiss) with a 32X objective. The laser scanning microscope is capable of scanning a large surface frame by frame (tile scan) with the use of the software ZEN 2011 on the PC. The laser light below the sample is collected and the Stokes Beam is blocked by a filter. The Pump beam is collected with a photodetector and the intensity modulation of the Pump beam is measured with a lock-in amplifier. The lock-in amplifier extracts the Raman signal from the Pump beam and with the software ZEN used to construct the images. In figure 3.7 for the Pump beam six selected wavelengths are in turn used to measure five targeted polymers. The wavelength of 893 nm is used for the blank and does not target a polymer vibration. In the following order the wavelength of 906, 908, 909, 921 and 935 nm or used for nylon, polyethyleentereftalaat (PET), polystyrene (PS), polypropylene (PP) and polyethylene (PE).[19]

(33)

33

Figure 3.7: The Stimulated Raman Scattering setup used in the laboratory to measure microplastics.[35]

(34)

34

4. Results and discussion

4.1. Micro plastic sedimentation

Table 4.1 represents the results of a few water samples to get an indication of the sedimentation of microplastics (Ø 20 µm, not fluorescent). The water samples originate from experiment 1 and 2. The results for experiment 2 are obtained by making use of the stirring rod. After 5 days the walls of the aquariums in experiment 2 were scraped with a glass rod. This was done to observe if there were microplastics sticking to the wall due to the centrifugal force created by the stirring rod. SRS is the technique used to detect the microplastics, but due to the unfortunate breakdown of the laser necessary for SRS many of the samples couldn’t be measured. A plot to represent the sedimentation of microplastics over time for the aquarium couldn’t be made due to these missing data. There is a big difference in the nominal concentration and the measured concentration of the samples “a test” and “BM15”. The measured concentration in the sample “a test” is much higher than the nominal concentration. For the sample “BM15” the measured concentration is much lower than the nominal concentration. However, from the obtained results in table 4.1 it can be observed that after 1 day there is, in both cases, an enormous decrease in microplastic concentration in the water.

Table: 4.1: Results from the microplastic sedimentation, in the aquarium setup for experiment 1 and 2. The results for experiment 2 are obtained by using the stirring rod in the aquarium.

Experiment 1 1 2 2 2*** 2***

Sample name a test a test 1 day BM15 AM11D BM15DR BM15DR

Sampling time 1 hour 1 day 5 min. 1 day 5 day 5 day

Metal wire container No No Yes Yes Yes Yes

Use of stirring rod Yes Yes Yes Yes Yes Yes

Nominal concentration (mp/ L)* 7142 7142 714285 714285 714285 714285 Concentration sample (mp/ L)** 24300 500 44800 2300 2700 2800

* The nominal concentration (micro plastics/ Liter) pipetted into the aquariums. ** Measured concentration. The particles were detected with SRS.

*** Sample measured 2 times with SRS. Water sample was taken after scraping the walls with a glass rod.

(35)

35 4.2. Extraction of micro plastics from exposed mussels 4.2.1. Enzymatic digestion and filtration

In the first extraction method enzymatic digestion is used to digest the mussel and vacuum filtration to separate the digested mussel from the micro plastics.

Enzymatic digestion (incubation with proteinase K for 2 hours at 50 °C) is done without any change in the protocol. When filtering the sample on the Anodisc filter using vacuum filtration, the filter was clogged fast. The exact amount filtered was not measured but when given an estimation it should have been around 15 ml of sample going through the filter (there is a total of 45 ml of sample after each enzymatic digestion). With a subsample of the freeze-dried and ground mussels bought from Jumbo the enzymatic digestion and vacuum filtration through the Anodisc membrane filter was repeated. Instead of 2 hours at 50 °C incubation for digestion by proteinase K, the sample was now incubated for 24 hours at 50 °C. After vacuum filtration of 15 ml sample (the filter was not clogged), 3 ml H2O2 (30 %) was added on the filter for half an hour

without vacuum filtration. After half an hour the hydrogen peroxide was filtered using vacuum filtration. The remaining sample (30 ml) was filtered for 3 hours. The filter was, after the 3 hours, fully clogged and an amount of 20 ml was still on top of the filter. In total an amount of 25 ml sample of the 45 ml was filtered. To repeat the procedure again (adding hydrogen peroxide on top of the filter) until all of the sample is filtered takes a lot of time and was not tested. Vacuum filtration of the digested mussel is fast in the beginning but slows down very fast to a very slow filtration speed, hence the necessary 3 hours of waiting before confirmation of a fully clogged filter. At the same time the entire enzymatic digestion steps were done without the addition of 0,20 g ground mussel (blank). The blank was also filtered through the Anodisc membrane filter by vacuum filtration. The whole amount of blank was filtered through the Anodisc filter without problems in a short amount of time.

After it is confirmed that filtration through the Anodisc filter didn’t work due to clogging, the proposition of first filtering the sample through a filter with bigger pore size (Isopore filter, Ø 5 µm) was tested with the back-flush filtration approach. Four samples were made with slight changes in the protocol as described in paragraph 3.3.2. In Photo 4.1 the results are shown of the four samples after following the enzymatic digestion procedure.

(36)

36

Note that there is no visible mussel tissue to be seen in sample 3 due to the hydrogen peroxide treatment. Before adding hydrogen peroxide to samples 3 and 1 there was a clear visible distinction between the two samples. In sample 3 the mussel tissue was still seen as wet clumped tissue. In sample 1 after proteinase K treatment the wet clumped tissue had disappeared into a brown liquid. In table 4.2 the results are shown of the filtration through the Isopore filter with back-flushing to the Anodisc filter. All the other samples were filtered through the Isopore filter without problems in a short amount of time. Using back-flushing, no clogging occurred on the Anodisc filter. Sample 3 was skipped, because there were no filtration problems with the enzymatic digestion protocol.

Table 4.2: Results from the back-flush filtration of the four samples.

Sample Filtration with the use of back-flushing 1 Good filtration, no clogging of both filters 2 Good filtration, no clogging of both filters

3 -

4 Good filtration, no clogging of both filters

It should be noted that SDS does react with CaCl2 in water or hydrogen peroxide. After the

reaction, in sample 4, a white emulsion can be seen and eventually precipitates to the bottom of the beaker. The white slurry after filtration on the filter doesn’t have a noticeable effect on the rate of filtration of the sample.

4.2.2. Efficiency of filtration using back-flushing

Using back-flushing the microplastics and residues are collected on the Anodisc filter. In the filtration step some microplastics may be lost or will linger on the polycarbonate filter after back-flushing. To give an estimate of the initial number in the sample the efficiency of the filtration and back-flushing should be known. The efficiency is measured by comparing the

1 2

3 4

Photo 4.1: Results of the four test samples after enzymatic digestion. Sample 1: following the described protocol. Sample 2: protocol without adding mussel tissue (blank). Sample 3: protocol without adding proteinase K. Sample 4: protocol without adding proteinase K and mussel tissue.