The development of a detection method for microplastics in seaweed

(1)

MSc Chemistry

Analytical Sciences

Master Thesis

The development of a

detection method for

microplastics in seaweed

by

Naomi Dam

11811374

March 2020

48 ETCS

February 2019 – January 2020

Supervisor:

Examiners:

Dr. Leo van Raamsdonk

Prof. dr. Govert W. Somsen

Dr. Rob Haselberg

Wageningen Food Safety Research

(2)

The development of a detection

method for microplastics in

(3)

(4)

Abstract

The dominant presence of plastic in the environment is considered an increasing problem. Degradation and wearing of plastic will result in the formation of microplastics (MPs). Due to their smaller size, these MPs may enter the food chain and can eventually cause health problems. To allow monitoring of MPs in feed and food products, the aim of this research was to develop a method for detecting MPs smaller than 300 µm by means of microscopy, allowing cheap, fast and easy detection. In the view of the easy dispersal and transferability of MPs in the aquatic (marine) environment, and of the scarcity of available data, seaweed was chosen as matrix.

Optimal detection of MPs was considered a crucial and necessary first step for proper inspection of the effects of the modifications of the sample pre-treatment method. A set of ten types of MPs was used, mostly of a globular shape. First tests revealed that solely microscopy provided insufficient detection possibilities. The use of Nile Red (NR) as fluorescent stain allowed easier detection. An optimal staining procedure was developed with NR in hexane, performed in Eppendorf tubes in an aqueous environment. The samples were inspected when dry, avoiding the use of mounting media. A cover glass fixed with nail polish was applied in order to reduce contamination and allow safer storage of the slides.

The sample pre-treatment focused on the effects on both the matrix and the MPs. Allowing easier detection of MPs, most of the matrix material should be removed and remaining matrix should not interfere with the MP detection. In other words, maximal weight loss should be obtained, with lowest autofluorescence and fluorescence after staining of the matrix. Minimal effects should be obtained for the MPs. Three reagents were tested for effects on the matrix and MPs, where nitric acid resulted in maximal weight loss of the matrix, lowest autofluorescence of the matrix and highest fluorescence of MPs after staining with NR. Sample pre-treatment of the samples with the aid of a microwave (MW) was chosen. An incomplete design with twenty-four experiments was worked out with varying temperature, time and concentration of the reagent in order to find optimal parameter settings for the MW-assisted digestion. An optimal balance between maximal removal of the matrix and minimal interference caused by (auto-)fluorescence intensity was obtained for 1 M nitric acid at 100 °C for 20 min. From sample pre-treatment up to detection, six steps were performed in the initially developed work-flow. This first version of a procedure was relatively time-consuming and susceptible for contamination. The work-flow was critically re-viewed and a new work-flow was proposed with only four steps to be performed. This four-step procedure allowed sample processing within one day, as shown for spiked matrices.

The current study focused on a broad range of aspects for the detection of MPs with NR and fluorescence microscopy. New information is described, such as live records of the staining procedure as indication for staining progression. An elaborate set of experiments was performed to learn more about microwave-assisted digestion of seaweed, sufficiently affecting seaweed with minimal effect on MPs. Further research is required to optimize the complete work-flow, allowing minimal contamination and expanding its applications to more matrices and MPs before real samples are going to be tested for presence of MPs.

(5)

Page 5 of 48

Acknowledgements

When I stepped foot into the building of Wageningen Food Safety Research (WFSR), I never imagined how I would become a part of this great team. I would like to thank dr. Leo van Raamsdonk for being my supervisor, letting me work on this amazing project and the discussions we could have for hours. I felt welcome when knocking on your door and enjoyed our weekly meetings. Thank you for letting me grow continuously. Secondly, I would like to thank the group of microscopy, where I have spent almost an entire year. Thank you, Theo, Corina, Jef and Bruno for the great times we have had together. You have made me feel so welcome and we had lots of fun together. In addition, you all have taught me so many things, both related to work and personal growth. Thank you for encouraging me and listening to me when I needed it. Thank you all so much.

Thanks to all the people in WFSR that have listened to me, have answered my questions and helped me finding the random things that I needed during my research. Thanks to John, Sandra, Nienke and Greet for helping me during my work in the heavy metals lab. You were always so kind and helpful and have taught me so much. Thanks to Eva, Thijs and Carmen for walking into our office every now and then, for listening and helping when I was feeling stressed and reading my many drafts. Additionally, I would like to thank prof. dr. Govert W. Somsen and dr. Rob Haselberg for their time evaluating this thesis.

In my close environment, I want to thank my parents Jan and Jolanda, for always listening to me and helping me. For supporting me and being there for me, always. Thank you so much. Finally, I would like to thank Helmoed, for listening and your personal support that you always provided. You were always there for me and allowed me to clear my head when I needed that. And everyone, thank you for believing in me.

Diemen, March 2020 Naomi Dam

(6)

Abbreviations

BF Bright field

DIC Differential interference contrast DoE Design of experiments

HNO3 Nitric acid

H2O2 Hydrogen peroxide

H2SO4 Sulphuric acid

KOH Potassium hydroxide LDPE Low density polyethylene MP Microplastic

MW Microwave NR Nile Red

NRA Nile Red in acetone NRHX Nile Red in hexane NRM Nile Red in methanol PA Polyamide

PE Polyethylene

PET Polyester, polyethylene terephthalate PHB/PHV Polyhydroxybutyrate/polyhydroxyvalerate PMMA Polymethylmethacrylate PP Polypropylene PS Polystyrene PVC Polyvinyl chloride PVDC Polyvinylidene chloride

UHMW PE Ultra-high molecular weight polyethylene TICT Twisted intra-molecular charge transfer

(7)

Page 7 of 48

1 Introduction

A life without plastic is already difficult to imagine for the current generation. Being used in consumer products such as packaging material, bottles, bags and cups and for industrial use, it is commonly found all over the world. This leads to large amounts of plastic: in 2015, the European demand was 49 million tons of plastic [1]. From this, 39.9% comprised packaging, often meaning these plastic products have a limited lifetime. Only 30% is collected for recycling and single-use plastics are often littered, ending up in the environment. Marine litter is heavily affected by plastic waste, between 60 and 80% of marine litter consists of plastic debris [2]. Under the influence of light or abrasion, plastics can degrade to microplastics (MPs). The term microplastic was first used by Thompson et al. in 2004 [3], where in 2009, MPs were defined as plastic particles smaller than 5 mm [4]. Lower limits vary in literature from 100 nm up to 335 µm, allowing a distinction between micro- and nanoplastics [5]. The importance of the lower cut-off values regarding size are discussed by van Raamsdonk et al. [6] Further distinction is made between primary and secondary MPs [7]. The first category refers to MPs that are produced as such, for instance microbeads in facial scrubs. The second category refers to MPs that are formed by abrasion and wearing, originating from e.g. clothes, bottles or packaging material.

Research has been continued since the term MP was first mentioned. MPs have been detected in numerous abiotic environmental compartments, including sediment in the UK and Belgium [8, 9], in lakes in Mongolia and Canada [10, 11] and from Dutch rivers to wastewater treatment plants [12]. Having a similar size as plankton or grains of sand, MPs can be ingested by organisms and enter the food chain [13]. Ingestion by organisms has already been demonstrated by experiments on lugworms [14], oysters, periwinkles, isopods [15] and mussels [16]. In fact, surveys proved the presence of MPs in crabs [17], mussels [18, 19], fish [20] and in amphipods at depths from 7000 to 10890 m [21]. Thus, the ability of MPs entering the food chain has already been confirmed, as well as their presence in the environment in many different areas. These results allowed further estimations by more studies that focused on the possible human exposure. For instance, a study by Cox et al. focused on data available from literature and combined these with U.S. dietary data to estimate the human MP consumption [22]. American adults and children could be exposed to 81 000 up to 123 000 MPs per year, based on recommended or average consumption of the analysed items. Tap water could result in an intake of 4000 MPs annually, increasing to 90 000 if only bottled water is consumed. The possible exposure of MPs is confirmed by Schwabl et al., who described the presence of MPs in human stool [23]. This raises the question whether the widespread occurrence of MPs is of concern to human and environmental health.

Till date, effects of exposure from MPs have only been investigated on animals. Polystyrene (PS) MPs, for example, were able to modify the microbiota composition in the gut of mice, while a hepatic lipid disorder was also induced [24]. In addition, PS MPs led to gut damage and alterations in the gut metabolome and microbiome in zebrafish [25]. Similar MPs accumulated in the liver, kidney and gut of mice [26]. These and other references are further discussed in a review by van Raamsdonk et al., having a deeper focus on the potential health effects of MPs [6].

MPs have been detected in a broad range of matrices, requiring different sample pre-treatments. This includes density separations for sediment [27] and sieving [28] or filtering [29] of aqueous samples. More complicated matrices require more elaborate sample pre-treatments. A digestion step for removal of organic material is most applied. Various methods include using potassium hydroxide (KOH) for digestion of amphipods [21], bivalve tissue [30] and mussels, crabs and fish [7], or hydrogen peroxide (H2O2) for mussels [16, 19].

Enzymatic digestion has been applied to mussels [31] and fish [32]. Detection can be performed visually using a microscope [33, 34], with additionally applying the fluorescent dye Nile Red (NR) [35–37]. Other more complicated methods for pre-treatment and detection include the use of pyrolysis gas chromatography coupled

(10)

to mass spectrometry (Py-GC/MS) [38]. Literature focuses mainly on simple matrices such as soil, sediment, water, fish and mussels. Several articles show that MPs are able to adhere to seaweed, which could provide another pathway to enter the food chain [39, 40].

Wageningen Food Safety Research (WFSR, previously known as RIKILT) is the institute for monitoring food and feed safety in the Netherlands and is currently involved in the complete method development for the detection of MPs in the food production chain, including sample pre-treatment. Aiming for a more complicated matrix, seaweed was chosen for its applicability in food and future prospects as feed. However, a specific sample pre-treatment of seaweed material in the scope of MP detection is still lacking from literature. In addition, detection is mainly focused on relatively larger MPs. Therefore, this research focuses on the development of the sample pre-treatment of seaweed, to extract and detect smaller MPs (≤ 300 µm) with microscopy, enabling follow-up identification with FTIR. The ability of MPs to adhere to seaweed has opened a new research gap to be filled.

Although different methods for detection have been applied in literature, microscopy was preferred here, allowing cheap, fast and easy detection. This study will start with development of the detection of MPs, applying Nile Red as fluorescent stain for easier detection. An optimal detection procedure is required for determining the effect of the sample treatment on both the matrix and on the MPs. The sample pre-treatment aims to achieve maximal removal of the matrix, determined by weight loss, while obtaining minimal effect on MPs and its detection. In addition, remaining matrix should not interfere with detection, taking into account the autofluorescence and fluorescence after staining with NR. Aiming for a relatively fast sample pre-treatment, microwave-assisted digestion will be tested. As a final step, the method aims to apply micro-FTIR for identification of MPs.

(11)

Page 11 of 48

2 Principles and outline

2.1 Nile Red

Nile Red, 9-diethylamino-5H-benzo[α]phenoxazine-5-one, has previously been used as a stain to detect intracellular lipid droplets [41]. The structure of NR is shown in Figure 1.

Figure 1: Structure of Nile Red

NR is soluble in different organic solvents but is poorly soluble in water. Its fluorescence is quenched in water, which is not the case for organic solvents. NR is a solvatochromic dye, which means that in a polar environment, a red shift is observed [42]. This is caused by NR undergoing a large dipole moment change during the transition between two electronic states. A charge transfer follows between the electron donating di-ethylamino group and the electron accepting aromatic system, which produces a twisted intra-molecular charge transfer (TICT) state. The latter would be stabilized in more polar solvents. Such a twist usually results in a lowering of the activation barrier [43]. The activation barrier would decrease with increasing polarity of the solvent. The TICT state may result in nonradiative relaxation instead of fluorescence [44]. When NR is in a non-polar environment, the TICT formation is thermodynamically unfavourable: as a result, both the fluorescent lifetime and the quantum yield increase. On the other hand, a decrease in the latter two can be caused by hydrogen bonding between NR and solvent molecules.

The application of NR in polymers was already described in 1996, acting as a probe to estimate the micropolarity of polymers [43]. Since empty spaces exist between the chains of polymers, relatively small molecules such as NR are able to be incorporated in those spaces and therefore act as a probe.

2.2 Reading guide

This research involves the development and optimization of different aspects of a complete method to detect microplastics in feed and food products. In general, a method for the detection and identification of MPs includes a sample pre-treatment, detection, observations and identification. This is the work-flow. However, the order of development differed from the work-flow, as visually depicted in the flow-scheme in Figure 2.

(12)

Figure 2: Flow scheme or visual representation of the current research.

Detection is the first part to be addressed. In order to know the effect of the sample pre-treatment on the appearance of the microplastics and the effect of removal of the matrix, the detection needed optimization first. Without an optimal detection, it is difficult to determine whether the sample pre-treatment affects the matrix sufficiently, without interference of detection. Detection, described in 4.1, involves the establishment of the need to stain the MPs, the staining protocol itself, parameter settings and location of the staining. The effect of the incubation time is discussed, and a solvent and concentration are determined. Additionally, the effects of a mounting medium and a cover glass follow, after which the characteristics of the stained MPs are discussed, regarding colours and intensity. With an optimal detection, the sample pre-treatment is discussed in 4.2. Different reagents for treatment of the sample material are discussed first, followed by optimization of the microwave procedure, determining an optimal concentration of the reagent, temperature and time. This is determined by observing the weight loss (%), autofluorescence and fluorescence after staining, post-treatment. Both the matrix and MPs are considered, shown in the table in the flow-scheme. The initial work-flow consists of six steps and was relatively time-consuming. Moreover, only a fraction of the sample could be inspected as a limitation of the initial work-flow. The work-flow is reviewed in 4.3 and adjusted to include less steps that are less time-consuming and allow inspection of the whole sample. Finally, for proof-of-principle, the new work-flow is tested with spiked samples as described in 4.4.

(13)

Page 13 of 48

3 Materials and Methods

3.1 Standards and reagents

3.1.1 Microplastics

Different microplastic reference standards were obtained. Polymethylmethacrylate (PMMA) and polystyrene (PS) were obtained from Sigma-Aldrich (Buchs, Switzerland). Blue polyethylene (PE) microspheres were obtained from Cospheric (Santa Barbara, USA). Polyamide (PA), polyvinylidene chloride (PVDC), ultra-high molecular weight polyethylene (UHMW PE), low density polyethylene (LDPE), polyester (PET), polyhydroxybutyrate/polyhydroxyvalerate 2% (PHB/PHV) and polypropylene (PP) were obtained from Goodfellow (Huntingdon, UK). More specific details can be found in Appendix A: Microplastic standards. Other microplastic material was obtained by cutting or fragmenting raw material. The following materials were processed for getting micro-fragments: cellulose acetate (packaging material for bread), polyvinyl chloride (PVC) tubing, rolling paper made of cellulose (Glass Clear Rolling Papers, Rolling Supreme), fishing line (Extra Line Mustad classic, monofilament 100% polyamide) and Styrofoam from internal packaging material.

3.1.2 Nile Red

Nile Red was obtained from Sigma-Aldrich (technical grade, Steinheim, Germany). n-Hexane (PEC grade), acetone (AR grade) and methanol (Ultra LC-MS grade) were obtained from Actu-All Chemicals (Oss, The Netherlands).

3.1.3 Seaweed

Brown seaweed was kindly provided by Siebren van Tuinen. The seaweed was first washed with demineralized water, to remove as many contaminants from the outside as possible. About 40 L per 500 g seaweed was used. After washing, the seaweed was freeze-dried and grounded to the smallest particle sizes possible using a Retsch Grindomix GM200.

For sample pre-treatment of brown seaweed, hydrogen peroxide (H2O2, 30%, Merck, Darmstadt, Germany),

nitric acid (HNO3, 67-69%, Carlo Erba reagents, Val de Reuil Cedex, France) and sulphuric acid (H2SO4, 0.1

M, Sigma-Aldrich, Steinheim, Germany) were used.

3.2 Equipment

3.2.1 Microscope and fluorescence filters

For fluorescence microscopy, an Olympus BX51 microscope was used, with filters WIBA (excitation 460-490 nm, emission 510-550 nm), WIB (excitation 460-460-490 nm, emission >515 nm) and WIG (excitation 520-550 nm, emission >580 nm). Photos were taken with an Olympus SC30/SC50 camera using cellSens. In addition, a Canon 550D equipped with a Tamron 90 mm macro lens was used for live staining.

3.2.2 Microwave

The seaweed was pre-treated in a Discover SP-D 80 microwave, using 80 mL quartz tubes with Teflon caps.

3.2.3 FTIR

For identification, an Agilent Cary 600 series FTIR microscope was used, equipped with an Agilent Cary 600 FTIR spectrometer. The computer was equipped with Resolutions Pro software and siMPle software was applied for data processing (www.simple-plastics.eu).

(14)

3.3 Protocols for establishing optimal parameters

Standard lab procedures were applied to determine the optimal parameters from Figure 2.

3.3.1 Solvent and concentration of Nile Red

Stock solutions of NR were made in a concentration of 50 mg/L in methanol (NRM) and acetone (NRA). From there, further dilutions were made in concentrations of 10 and 1 μg/mL. From the stock solution of 50 mg/L NR in acetone, dilutions of 10 and 1 μg/mL in hexane (NRHX) were made. The stock solutions were stored in brown flasks.

3.3.2 Staining protocol

Staining on slide

A small number of plastic particles (or any sample) was placed on a microscopic slide with a spatula. Then, about 2-3 drops of the NR stain were added to the slide, enough to cover the sample. To prevent contamination and reduce evaporation, the microscopic slides were covered with an hour-glass. When the solvent was evaporated, the slides were visually inspected with fluorescence microscopy. The slides with samples were initially not covered with a cover glass, nor was a mounting medium applied.

The PS and PMMA standards were available in dispersion. To obtain dry MPs, the dispersion was shaken manually, and a few drops were added to a microscopic slide with a Pasteur’s pipette. This was left to dry. Then, for further staining experiments, the MPs were scratched from the microscopic slide and carefully transferred.

Live staining

A small amount of sample was applied to a microscopic slide with a single cavity (Karl Hecht Assistent, ~ 76×26×1.35 mm), followed by 1-2 drops of NR solution. This was performed while the slide was under the microscope. The process of staining and evaporation was viewed live at a magnification of 4× by looking through the ocular and using cellSens. Photos and videos were taken of the process.

Staining in Eppendorf tubes

A small amount of sample was transferred to an Eppendorf tube (sufficient to cover the bottom of the tube). Then, about 0.75 mL of demineralized water was added. Next, about 0.25 mL of NR stain was added, the tube was shaken manually and left to stand for at least one hour. Afterwards, some of the sample was added to a microscopic slide with a spatula or Pasteur’s pipette. With the use of a heating plate at a temperature of approximately 40 °C, the sample was left to dry on the slide.

3.3.3 Mounting medium and cover glass

Glycerol (bidistilled, 99.5%, VWR Chemicals, Leuven, Belgium), chloral hydrate (Merck, Darmstadt, Germany, dissolved in a solution of 10%) and paraffin oil (Sigma-Aldrich, Steinheim, Germany) were tested as mounting media. The sample was transferred to a slide and 2-3 drops of mounting medium were added. Then, a cover glass was carefully placed on top.

For dry inspection of samples, a cover glass was attached with nail polish as follows. The sample was transferred to a microscopic slide with a spatula or Pasteur’s pipette. Nail polish was carefully applied around the sample, following the contours of a cover glass. The slide was then placed on a heating plate at a temperature of about 40 °C, allowing to dry. Usually, three or four layers of nail polish were applied before adding the cover glass, allowing each layer to dry separately. The layers added height and prevented the nail polish from running out and touching the sample.

(15)

Page 15 of 48

3.3.4 Optimal circumstances for sample pre-treatment

Reagent

Three reagents (H2O2, H2SO4 and HNO3) were tested. H2O2 was initially tested in Erlenmeyer flasks at room

temperature and at 40 °C, using a heating plate and a water bath, at different concentrations, namely 5, 10, 15 and 30%. Further testing was performed in the microwave for all concentrations except 30%. H2SO4 and

HNO3 were initially tested in the microwave, starting with 0.1 M acid, 100 °C for 10 min.

Design of Experiments

The microwave method included three parameters: temperature, time and concentration. For each parameter, three levels were chosen. An incomplete design of experiments (DoE) was made for the following levels: 80-100-120 °C, 10-15-20 min, and 0-0.1-1-2 M acid. Out of the total number of 36 possible combinations, a set of twenty-four experiments was performed, to find the optimal balance of parameters.

Determining weight loss

Before the experiments, folded filters (VWR folded qualitative filter paper, 302, Leuven, Belgium) were coded with pencil and then weighed. After sample pre-treatment, the samples were poured over the filter, then carefully rinsed with demineralized water to achieve a pH higher than 4.5. The quartz tubes were carefully rinsed with demineralized water to transfer remaining particles to the filter. Then, the samples were dried in an oven at 55 °C to determine the weight post-treatment. The samples were allowed to dry at least overnight.

Autofluorescence and fluorescence of the seaweed after staining

After drying and weighing, seaweed was stained in Eppendorf tubes with NR in all three solvents. To determine the autofluorescence post-treatment, blanks were included as follows. To a small amount of microwave-treated seaweed in an Eppendorf tube, about 0.75 mL of demineralized water and about 0.25 mL of hexane, methanol or acetone was added. After shaking, this was left to stand for at least one hour. Slides were prepared for dry inspection.

3.4 Spiking seaweed

Brown seaweed was spiked with 2.5, 5 and 10% of UHMW PE and PET, single and combined. According to the optimized sample pre-treatment, the samples were microwaved and then stained in the quartz tubes. After incubation for at least one hour, the samples were filtered using a vacuum filtration system with glass microfiber filters (Whatman 1920-090, 90 mm diameter, particle retention 1.6 µm). The filter was transferred to a 90 mm plastic petri dish and then visually inspected using fluorescence microscopy with filter WIB. For every sample, a set of photos was takenat similar positions along a grid using a special designed method as explained in Appendix O: Method for representative visual inspection, using an exposure time of 1000 ms or lower, if applicable. In addition, extra photos were taken if the overview was not representative for the whole sample.

3.5 Pre-treatment for FTIR identification

Prior to identification with FTIR microscopy, samples were filtered over an Anodisc 25 filter (0.2 μm, Whatman, Buckinghamshire, UK). A small piece of a VWR filter paper was cut out and placed on top of the Anodisc filter, to prevent particles from reaching the covered area during filtering. That area could be used for background measurement later. After filtration, the filters were placed onto another piece of VWR filter in a glass Petri dish. The piece of filter would prevent the Anodisc from sticking to the glass. Then, the Anodisc filters were placed in an oven at 55 °C to dry overnight.

(16)

4 Results of method development and discussion

4.1 Detection

4.1.1 Need for using a staining agent

Visual inspection and sorting using a microscope is often applied for the detection of microplastics (MPs) [45]. However, visual inspection is mostly applied for relatively large MPs. Therefore, it is important to determine whether it is possible to visually recognize MP particles smaller than 300 μm. Hence, visual inspection of the MP standards was first attempted using bright field (BF), applying a polarizing filter and differential interference contrast (DIC).

Microplastics under bright field

Visual inspection shows that there is no distinct shape, surface or colour that would distinguish MPs, as illustrated in Table 1. The surface, shape and colour of MPs may differ depending on the original condition and the degree of weathering. In addition, there is no specific pattern visible when a polarizing filter is applied. In other words, there are no

characteristics that were observed for all MP standards, allowing them to be recognized as MPs. However, the absence of any organic structures such as cell walls and cell membranes in the MP particles is in line with the criteria defined by Norén, which are: 1) the absence of cellular or organic structures; 2) a fibre should be equally thick and have a three-dimensional bending and 3) particles should be clear and homogeneously coloured [45]. Nevertheless, Norén focused mainly on larger particles and fibres were included. The research of the current study is intended to focus

on smaller particles and solely these criteria do not suffice to find MPs. Consequently, a different method is required to simplify the detection of MPs.

Nile Red

MP detection becomes more difficult when the particle size decreases, where matrix interference increases. In addition, MPs may be missed in visual inspection when they are covered by other particles. NR may facilitate in these cases by acting as a fluorescent probe. Andrady was the first to mention the use of NR for the detection of MPs [46]. Cole developed a reproducible and replicable method to prepare microfibers, allowing researchers to produce suitable microfibers for toxicity testing on organisms [47]. NR stain was applied to confirm the applicability of the microfibers in a feeding experiment with brine shrimp. An NR staining protocol optimized by Shim et al. allowed discrimination between MPs and sand, resulting in a similar recovery rate as obtained with FTIR [36]. Similar results were obtained by Maes et al., where particles were recognized as MPs by both NR staining and IR-microscopy [37]. In addition, the NR staining protocol

Table 1: MPs under bright field (BF), polarization (pol) with (pink) and without (black) red filter, and under differential interference contrast (DIC), 20× magnification, mounting medium: paraffin oil.

BF Pol black Pol pink DIC pink DIC black DIC grey

PS UHMW PE PA PET LDPE

(17)

Page 17 of 48

was applied on sediment samples. More matrices and the effect of NR staining were discussed by Erni-Cassola

et al., where NR allowed detection of MPs in sediment and water samples, applying μ-Raman for

identification [48]. NR was applied for detection of MPs in bottled water by Mason et al. [35], sediment by Wang et al. [49], plankton samples by Hitchcock et al. [50] and in mussels by Catarino et al. [18]. Moreover, NR allowed characterisation of the fibre composition of wet wipes, as described by Pantoja Munoz et al. [51]. These nine literature references were available when this research started and were used as a guideline for the method development of an NR staining procedure. This is further explained in the following section.

4.1.2 Development of staining protocol

Different staining protocols, concentrations, solvents and fluorescence filters were described in literature. All information was taken into account and visualized to get an overview. The use of different concentrations and solvents is visualized in Figure 3. The visualization in Figure 3 shows that there is no consensus for the solvent or concentration used. In addition, the use of the different fluorescence filters will further influence the observations. The excitation and emission wavelengths of the fluorescence filters from literature are visually depicted in Appendix C: Overview fluorescence filters from literature , together with the fluorescence filters available at WFSR. All this information did not exclude any options, so the start of the development was rather broad. However, it became clear that some decisions were based on deformation or co-staining of the filter membrane. Therefore, the aim was to find a staining protocol that allows fluorescence with high intensity for MPs and which is easy and fast to apply.

The optimization involves three different parameters: the staining protocol, concentration and solvent. The

information obtained from literature was used as guideline, hence the concentrations 1 and 10 µg/mL were tested in methanol, acetone and hexane. The three fluorescence filters available (WIBA, WIB, WIG), were all three applied, to get as much information as possible. Other factors include the incubation time, the use of a cover glass and the use of a mounting medium. These are discussed next.

Staining protocol

Initially, staining on a slide was performed with 10 μg/mL NR. This protocol provides both advantages and disadvantages. One of the advantages includes a simple procedure: application of the sample to a slide, addition of a few drops of the stain, cover and wait until the solvent is evaporated. The slide is placed under the microscope and is visually inspected. However, transferring and storing the slides resulted in some MPs

Figure 3: Overview of the concentrations and solvents used in literature, using Nile Red to stain MPs.

500 10 5 2 1 500 10 5 2 1 500 10 5 2 1

Methanol Acetone Hexane

Concentration (µg/mL) & solvent

Concentrations and solvents used for Nile Red stain

(18)

falling off of the slides, requiring a different procedure. Besides, the influence of the solvent evaporation was unclear. It was unknown whether NR is able to probe the MPs before full evaporation of the solvent. Therefore, live staining was performed to screen the effects of the incubation time and solvent evaporation. Live staining was performed while the slide and sample were under the microscope. Fluorescence filter WIB was applied, because it provides the broadest spectrum of colours to be observed. The process was captured by taking photos and videos, not only using cellSens, but also a Canon 550D through the ocular. Due to the solvatochromic behaviour of NR, different colours were expected for NR in different solvents and for stained MPs. These differences are indeed observed: hexane shows yellow/green emission, whereas acetone and methanol show more orange/red emission. As soon as the solvent evaporates, a change in colour appears. For instance, the relatively nonpolar PE surrounded by NR in methanol appears red at first. When the solvent completely evaporates, a change from red towards yellow/green is observed within seconds. This process is shown in Table 2.

Testing live staining shows that for most plastics, the staining works quickly. This indicates that the incubation time is not a very important factor. In addition, PA shows a rather slow switch when NR in hexane (NRHX) was added, compared to NR in acetone (NRA). A more polar solvent could be more suitable for the relatively more polar plastics.

As described by Shim et al., NR may partition more easily from solvent to plastic, by using a non-polar solvent such as hexane [36]. However, plastics can be relatively polar as well. Therefore, a different protocol was tested. This staining protocol includes an aqueous phase, to which the dye is added. It is performed in an Eppendorf tube and in the case of hexane, two layers arise. Relatively non-polar MPs reside in the hexane phase, while relatively polar plastics reside in the aqueous phase. In addition, some plastics turn pink after being incubated for some time. Advantages of this protocol include that the solvent is less likely to evaporate, closed Eppendorf tubes enable less contamination and the presence of two phases could allow the best penetration of NR. The emission intensities were compared and show that a higher intensity is obtained when the staining is performed in Eppendorf tubes. A comparison of the two protocols and the differences in emission intensities of LDPE are shown in Table 3. The emission intensities show that a higher intensity is obtained when the staining is performed in Eppendorf tubes. Although the staining procedure worked quickly as shown by live staining, evaporation is less likely to occur in Eppendorf tubes. More NR is able to probe

Table 2: Colour switch observed during live staining of the relatively nonpolar UHMW PE with NR, concentration 10 µg/mL. Photos taken with Canon 550D, 4× magnification.

Start During During After

Methanol

Acetone

(19)

Page 19 of 48

MPs without evaporation, resulting in a higher fluorescence intensity. Therefore, staining in Eppendorf tubes is preferred over staining on slides, allowing less contamination and providing a higher emission intensity.

Mounting medium

Generally, a mounting medium is employed in microscopy to embed the specimen. This medium is chosen in such a way that the refractive index of the mounting medium matches with the particle(s) of interest. When

Table 4: The influence of different mounting media on PA stained with NRHX 10 µg/mL, 4× magnification, 100 ms exposure.

BF WIBA WIB WIG

Before After addition of chloral hydrate Before After addition of glycerol Before After addition of paraffin oil

Table 3: Differences in emission intensities of LDPE stained with NR 10 µg/mL on a slide compared to an Eppendorf tube, exposure 1000 ms, 4× magnification.

Staining on slide Staining in Eppendorf tube

WIBA WIB WIG WIBA WIB WIG

Methanol

Acetone

(20)

chosen correctly, the particle may appear more translucent, making other elements more visible. A cover glass is added to hold the specimen in place and protect it from dust. The use of different mounting media was tested after staining, namely glycerol, paraffin oil and chloral hydrate. An example of the effect of addition of these mounting media to stained PA is shown in Table 4. All three mounting media result in a lower fluorescence intensity and leakage of the dye from the particles. More specifically, a blue colour appears when chloral hydrate is added to stained PA, which is visible in the

after-picture in BF. In addition, PA appears to dissolve in chloral hydrate. The fluorescence of NR is almost completely quenched as well. Glycerol also causes decreased fluorescence intensity, although not as clearly. Green vesicles are observed in the after-picture under filter WIB. These are most likely hexane vesicles that were not evaporated yet, appearing between the PA particles. Finally, events occurring in paraffin oil were not captured well because of a too high magnification. Therefore, another picture was taken with a phone through the ocular, shown in Figure 4. It shows that NR leaks out of the particles, which is indicated by the green glow that is formed around the orange PA particles. The influence of the mounting medium is in agreement with the effect of hydrogen bonding, earlier described by Cser et al. [52] Since glycerol, paraffin oil and chloral hydrate all are capable of hydrogen bonding, they allow quenching of the NR stain. In addition, Cser et al. described that viscosity barely influences the NR fluorescence.

However, the mounting media cause NR to leak out. Although this may seem as a disadvantage, it may allow removal of the dye in case of interference for identification with FTIR. Pick et al. showed that glycerol keeps NR in solution, which may explain the “leaking out” that is observed here. [53] These experiments emphasize that mounting media should not be used after staining with NR, unless removal of the dye is desired. Due to the effects of the mounting media, the samples were observed when dry. To allow storage of the samples without chance of contamination or losing the sample, a cover glass is attached by applying nail polish as a glue. The polish is applied around the particles, following the contour of the cover glass. Each layer is allowed to dry on a heating plate, reducing the required drying time. Multiple layers are added to increase height, avoiding the chance of the cover glass breaking caused by some larger particles. Insufficient drying of the nail polish causes the polish to run out and touch the particles, resulting in NR leaking out. Based on these experiments, the protocol in Eppendorf tubes was determined to be best, followed by dry inspection of the sample. Finding the optimal solvent and concentration was therefore the next step.

Solvent and concentration

With the optimal staining protocol determined, experiments were performed with NR at two concentrations (1 and 10 µg/mL) for each solvent (methanol, acetone, hexane) to determine the best combination of concentration and solvent. The slides were visually inspected and photos were taken at 1000 ms exposure and lower if applicable. To allow uniform comparison, a scoring chart was designed to assign scores based on fluorescence intensity. This chart is shown in Table 5. The scores for all MPs were added for each filter and solvent, visually depicted in Figure 5.

Figure 4: Paraffin oil added to PA stained with NRHX. The green glow is NR leaking out of the orange PA particles. 4× magnification, photo taken with a phone.

(21)

Page 21 of 48

X

Figure 5 shows that lower emission intensities are obtained for 1 µg/mL of NR compared to 10 µg/mL. The higher scores obtained for 10 μg/mL are closer to reaching the maximum score of 84. In theory, a maximum score of 84 can be obtained for seven MPs in three filters, when the maximum score of 4 is assigned to each photo. Practice shows that this is not achievable, because filter WIBA has an emission upper limit at 550 nm, whereas the two other fluorescence filters do not have an upper limit to the emission spectrum. As a result, some MPs have a comparable lower fluorescence intensity in filter WIBA which is most likely related to their polarity. The relation between polarity and fluorescence intensity is more deeply discussed in 4.1.3. Another advantage of the concentration of 10 µg/mL is that some MPs turn pink under visible light after staining. This provides an easy visual inspection to find MPs with the bare eye. Because of these considerations, 10 µg/mL is preferred over 1 µg/mL. As for the solvent, acetone was regarded unsuitable because it deformed some MPs. Methanol and hexane obtained similar scores for seven MPs, however PMMA and PS were excluded

Table 5: Scoring chart for scoring the intensity of stained MPs

Intensity

1 2 3 4

Barely fluorescence visible.

Fluorescence visible, edges of particles are well

visible. However, white dots of high intensity are

missing.

Clear fluorescence visible, some white dots of high intensity visible.

Fluorescence is so intense that all particles are white

due to such a high intensity.

Figure 5: Visual representation of scoring of the emission intensity of seven MPs after staining with NR at two concentrations (1 and 10 µg/mL). Scores summed per solvent and per filter (WIBA, WIB and WIG). Staining performed in Eppendorf tubes. The dashed blue line indicates the maximum score of 84 that can be achieved.

16 14 14.5 19 16 20 24 14 22 27 25 26.5 26.5 25.5 24.5 28 28 28 Maximum score of 84 0 10 20 30 40 50 60 70 80 90

Methanol Acetone Hexane Methanol Acetone Hexane

1 μg/mL 10 μg/mL

Emission intensity of microplastics stained with Nile Red,

comparing 1 and 10 μg/mL

(22)

from Figure 5. Both show a higher intensity after staining with NR in hexane. Furthermore, some plastics were only able to reside in the non-polar hexane phase. Including the fact that hexane would be the safest chemical to work with, hexane was preferred as solvent for the NR stain.

Fluorescence filters

During the optimization, the MPs were visually inspected under bright field and the three fluorescence filters. As mentioned before, different fluorescence filters were applied in literature, as depicted in Appendix C: Overview fluorescence filters from literature. In some literature references, MPs are only visually inspected with green emission. The application of only green emission excludes more polar MPs from detection, or makes this more difficult. Switching to a different fluorescence filter that is already incorporated into the microscope is easily and quickly achieved. Therefore, all three filters should be applied to get as much information as possible.

4.1.3 Colours and intensities of emitted light by stained MPs

The optimization of the staining protocol resulted in a concentration of 10 μg/mL NR in hexane, performed in an aqueous environment in an Eppendorf tube. Mounting media are not applied and a cover glass is attached with nail polish. During the optimization, different colours and intensities of the emission light were observed under different circumstances. These observations are discussed more deeply.

Depending on the emission wavelengths, colours vary among the filters, where only green is observed in WIBA, yellow/green towards orange in WIB and orange towards red in WIG. Different colours within one filter can only be observed in WIB and WIG, having no upper limit for the emission spectrum compared to WIBA. Relatively non-polar plastics, such as LDPE, UHMW PE and PP, show a yellow/green colour using filter WIB. On the contrary, relatively polar plastics such as PMMA, PET and PA, show a yellow/orange colour using filter WIB. MPs capable of hydrogen bonding are yellow/orange in colour, while those who cannot are green/yellow. With higher polarity, a larger shift was observed towards darker orange or red in filter WIG. A summary of these photos is shown in Table 6, where the plastics are sorted on their dielectric constant, indicating the polarity of the MPs. The table is focusing on the colour of the different plastics in the different filters, meaning that photos from different experiments were chosen to reduce the interference caused by high intensity.

(23)

Page 23 of 48

In addition to the scoring chart for emission intensity (Table 5), an additional scoring chart was designed for colours. This chart is shown in Table 7. Together with the scoring chart from Table 5, the intensity and colours are visualized in Figure 6. This visualization is based on photos taken with the optimized staining method, allowing comparison under the exact same circumstances. The photos are shown in Appendix D: Microplastics stained in Eppendorf tubes.

Table 6: Different emission colours of reference MPs after staining with NRHX in Eppendorf tubes, focusing on the appearance of different colours. Magnification 4×.

Dielectric constant

Concentration (μg/mL)

Exposure (ms)

WIBA, WIB, WIG WIBA WIB WIG

LDPE 2.2-2.35 1 1000, 100, 100 UHMW PE 2.3 1 1000, 100, 100 PP 2.2-2.6 10 40, 40, 40 PS 2.56 1 1000, 100, 100 PET 3 1 1000, 1000, 100 PMMA 3 10 1000, 100, 100 PHB/PHV 3 1 1000, 1000, 1000 PVDC 3-6 1 1000, 1000, 100 PA 3.6 1 1000, 1000, 100

(24)

The visualization in Figure 6 shows clearly that increasing polarity results in a shift in colour towards the red part of the spectrum. Together with this shift, an increase in intensity is observed. These differences are also clearly visible in the case of cellulose and cellulose acetate. Cellulose acetate is less polar than cellulose. Hydroxyl groups are replaced by acetyl groups, which causes less interchain hydrogen bonding [54]. Therefore, different colours were expected after staining with NR. Photos of cellulose and cellulose acetate after staining with NR are shown in Table 8.

Table 7: Scoring chart for colours, applied to MPs.

Colours

1 2 3 4 5 6 7

Figure 6: Visualization of intensity and colours after staining with NRHX, 10 µg/mL, filters WIBA, WIB and WIG, 4× magnification. MPs are sorted on their dielectric constant. The position on the x-axis shows the colour as assigned according to the chart. The size shows the intensity of the fluorescence.

(25)

Page 25 of 48

The difference is clearly visible: the more non-polar cellulose acetate has higher fluorescence intensity in filter WIBA, and the colour differences in WIB and WIG are distinct. Cellulose acetate is yellow in WIB and orange in WIG, while cellulose is orange in WIB and red in WIG. This confirms the red shift caused by the increasing polarity. Given the obtained information from all tested MP standards, the broadest spectrum of colours is observed in filter WIB. A high emission intensity is observed and colours from yellow/green towards orange can be observed, allowing an indication of the polarity. The option to visualise the dielectric constant can be used as a first aid for identification of MPs.

With an optimized staining protocol, the next step involves sample pre-treatment.

4.2 Sample pre-treatment

Normally, a sample pre-treatment aims at extraction of the analyte from the matrix and may be followed by further purification. However, this is generally applied for analytes in solution. MPs exist as solid particles, requiring a different sample pre-treatment focused on the removal of the matrix instead of extraction of the analyte. As explained by van Raamsdonk et al., more complicated matrices require more severe treatments [6]. Since MPs are able to adhere to seaweed and are possibly found in seaweed, a simple extraction such as a density separation will not suffice for removal of the matrix.

Seaweed is mostly comprised of polysaccharides. They can be hydrolysed, allowing smaller saccharides to be soluble and therefore be removed from the sample. The WFSR procedure for heavy metal detection includes complete destruction of the seaweed matrix material by a microwave-assisted digestion. The sample

Table 8: Different colours observed after staining cellulose and cellulose acetate with NRA, Eppendorf tube, 10 µg/mL. 4× magnification, exposure of 1000 ms.

Cellulose and cellulose acetate

BF Fluo WIBA Fluo WIB Fluo WIG

Cellulose acetate

Cellulose

Summary of staining protocol

Staining is performed according to the following protocol. The sample is transferred to an Eppendorf tube, to which 0.75 mL demineralized water and then 0.25 mL of 10 µg/mL NR in hexane is added. The tube is shaken manually, followed by incubation for at least one hour. Next, the sample is transferred to a microscopic slide and allowed to dry on a heating plate. A cover glass is attached with nail polish to allow storage and prevent contamination. The sample is visually inspected using all three fluorescence filters (WIBA, WIB and WIG), where most information is obtained with filter WIB.

(26)

pre-treatment protocol was therefore initially designed with elements of the protocol for detection of heavy metals. The goal was to find a sample pre-treatment allowing to lose a sufficient amount of the matrix in weight, while the MPs remain unaffected. The microwave allows a faster pre-treatment compared to a 24 h process, for instance involving KOH as developed by Dehaut et al. [7].

The optimization involves effects on both portions of a sample: the matrix and the MPs. The matrix should be sufficiently affected by the treatment, or being dissolved mostly. In addition, remaining matrix should not interfere with the detection of MPs. Therefore, three post-treatment factors are considered for screening the optimal effect on the matrix: weight loss (%), autofluorescence and fluorescence after staining. The goal is to lose at least 67% of the weight of the matrix. Since fluorescence is used for detection of the MPs, the autofluorescence should not interfere with the detection and the fluorescence intensity of the seaweed after staining should be minimal. For screening the optimal effect on MPs, two post-treatment factors are considered: weight loss (%) of MPs and fluorescence after staining. The weight loss of MPs should be minimal, and the intensity of the fluorescence should be affected minimally. First, the possible reagents will be discussed and tested for their effects on both the matrix and the MPs. Three reagents were tested, which were sulphuric acid, nitric acid and hydrogen peroxide. Secondly, the effects of three parameters for the application of a microwave will be optimized: temperature, time and concentration. In the next section, both elements, the optimal reagent and the optimal settings of the parameters will be combined in a procedure for the detection of MPs in seaweed.

4.2.1 Method development – reagents

The first experiments were mainly focused on determining whether the microwave in combination with a reagent destroys the matrix sufficiently, without destroying the MPs. Therefore, the seaweed and MPs after pre-treatment were not filtered or weighed yet, but the results were interpreted only visually.

Sulphuric acid

A low concentration of sulphuric acid (H2SO4) has been used by Teh et al. for hydrolysis of red macroalgae

[55]. Allowing a low concentration to suffice, H2SO4 was the first reagent to be tested. To determine a working

range considering temperature and time, initial experiments were performed with seaweed, LDPE and PA. Half a gram of the materials was weighed and 15 mL of 0.1 M H2SO4 was added, while various temperatures

were tested for 10 min. The initial settings are shown in Table 9. Since the experiments with H2SO4 were the

first to be performed, seaweed and MPs were not filtered or weighed yet, but only visually inspected post-treatment.

LDPE melted at 100 °C and PA melted at 140 °C. As soon as melting was observed during experiments, higher temperatures were not tested. Visual inspection indicated degradation of the matrix: the particles appeared smaller and the matrix showed less autofluorescence. To determine whether the microwave with sulphuric acid provided sufficient effect, seaweed from the sample at 140 °C for 10 min was spiked with a small number of PA and stained with NR. The stained PA particles are visible despite of the matrix still being present, as can be seen in Figure 7.

Table 9: Initial settings for first tests with microwave. Testing was performed with 0.5 g sample and 15 mL of 0.1 M H2SO4.

Sample Temperature (°C) Time (min)

1 LDPE 100 10 2 PA 100 10 3 Seaweed 100 10 4 LDPE 120 10 5 PA 120 10 6 Seaweed 120 10 7 PA 140 10 8 Seaweed 140 10 9 Seaweed 160 10

(27)

Page 27 of 48

Although LDPE melted during the experiments, the question was whether this is an occurring problem for all MPs. Since the microwave could give promising results, a second set of experiments was performed to determine the effect of the microwave and H2SO4

on MPs and post-treatment staining with NR.

Effect of sulphuric acid on extended set of MP standards

All MP standards from Goodfellow were tested with 15 mL of 0.1 M H2SO4 at 100 °C for 10 min. The other MPs available, PS and PMMA

from Sigma-Aldrich, were not tested because they are available in dispersion and therefore difficult to weigh. Moreover, a lower quantity was available of these standards. LDPE and UHMW PE were additionally tested at 80 and 120 °C, respectively, to learn more about the effects of the MW treatments. After the experiments, LDPE

and PVDC had melted, where PVDC even showed a slight change in colour. As a result, LDPE, PVDC and PP (because of the 3 mm particles) had to be fragmented. All samples were finally stained with NR in all three solvents to get as much information as possible. The photos are shown in Appendix E: Microplastics stained after MW treatment with 0.1 M H2SO4. A decreased fluorescence intensity was observed for PET,

PVDC and PHB/PHV. In addition, orange dots were observed for LDPE in filter WIB. It was unknown whether these differences were caused by H2SO4 or the microwave. According to literature, H2SO4 is able to

change the surface of LDPE, as well as to introduce polar groups [56]. Wang et al. used FTIR to look for chemical changes on the surface of LDPE and found sulphonic groups after exposure to H2SO4. On the

contrary, Wang et al. described little effect caused by treatment with HNO3. To determine the effect of the

microwave, a set of experiments was initiated with a selection of MPs in water and in HNO3.

Comparison of effects of water, sulphuric acid and nitric acid on MP standards

To determine the cause of decreased fluorescence after staining of PET, PVDC and PHB/PHV pre-treated in microwave with 0.1 M sulphuric acid, experiments were performed with water and 0.1 M nitric acid with fixed parameters (100 °C, 10 min). Post-treatment, MPs were stained in all three solvents and photos were taken for all three fluorescence filters. The photos from testing with water are shown in Appendix F: Microplastics stained after MW treatment with water and the photos from testing with HNO3 are shown in

Appendix G: Microplastics stained after MW treatment with 0.1 M HNO3. Additionally, an overview of the

physical effects of MW treatment with H2SO4, HNO3 and water, indicated by melting of plastics, is shown in

Table 10. After treatment with water in the microwave, PVDC still showed similar morphological changes to pre-treatment with H2SO4. Its physical appearance changed from powder towards melted aggregates, which

appeared to be brittle and beige in colour. After staining with NR, a change in colour was still observed: with filter WIB, colours varying from green/yellow towards red were observed. The cause of these changes is yet unclear and out of the scope of this research. In contrast to PVDC, no visible change of PET and PHB/PHV was observed after pre-treatment: the plastics were still a powder and not aggregated. The final choice for the reagent with optimal effect remains inconclusive as based on the results from Table 10. Therefore, other parameters have to be considered.

Figure 7: Seaweed pre-treated with 0.1 M H2SO4 at 140 °C for 10 min,

post-treatment spiked with PA and stained with NRM. Filter WIB, 1000 ms exposure, 4× magnification.

(28)

X

The fluorescence response of MPs after treatment with two reagents and after NR staining was used for selecting the most suitable reagent. Initially a decreased fluorescence intensity was observed for some MPs after treatment with sulphuric acid. Therefore, it is very likely that pre-treatment with 0.1 M sulphuric acid caused decreased fluorescence intensity. For confirmation, the photos of all stained MPs after treatment with both nitric and sulphuric acid were compared and scores were assigned for each plastic, according to the scoring chart from Table 5. The scores were added for all MPs, for each solvent and filter, which is visualized in Figure 8. For comparison of emission intensities with and without MW treatment, Figure 8 can be compared to Figure 5. A higher fluorescence intensity after staining with NR is obtained after treatment with nitric acid, compared to sulphuric acid. The difference appears highest for acetone and hexane, although not extreme. Besides fluorescence intensity, other aspects for the decision of a reagent were considered as well. Most important was the safety of the method: HNO3 provides less risks than H2SO4 when combined with the

solvents used for NR staining. As a result, nitric acid was preferred over sulphuric acid.

Table 10: Effects of microwave treatment of microplastics expressed as presence of melted aggregates under different circumstances. Temperature varies from 80 to 140 °C, 0.1 M acid was applied and the samples were microwaved for 10 min. Blank means not executed.

H2SO4 HNO3 H2O

80 100 120 140 100 100

LDPE Present Present Present

UHMW PE Absent Absent Absent

PP Absent Absent

PVDC Present Present Present

PHB/PHV Absent Absent Absent

PET Absent Absent Absent

PA Absent Absent Present Absent

Figure 8: Visual representation of scoring of the emission intensity of seven MPs after MW treatment with sulphuric acid and nitric acid. Post-treatment, the MPs were stained with NR according to the optimized method. Scores were summer per solvent and per filter (WIBA, WIB and WIG). The dashed blue line indicates the maximum score of 84 that can be achieved.

16.5 16.5 16 17 18.75 18.5 24.25 25 24 25.5 26.25 26 25.25 24.5 25.75 26 27.75 27.75 Maximum score of 84 0 10 20 30 40 50 60 70 80 90

Methanol Acetone Hexane Methanol Acetone Hexane Sulphuric acid Nitric acid

Emission intensity of microplastics stained with Nile Red after

pre-treatment with sulphuric acid or nitric acid

(29)

Page 29 of 48

Seaweed is completely dissolved by a microwave-assisted digestion procedure according to a WFSR method for detection of heavy metals. The digestion is achieved by applying a combination of concentrated nitric acid and hydrogen peroxide with high temperatures. This heavy treatment was therefore expected to have a negative effect on MPs. To screen the effects, half a gram of LDPE was tested according to the procedure. As expected, the treatment was able to completely dissolve LDPE. This confirms a required balance between maximal effect on the matrix but minimal effect on the MPs. Nevertheless, hydrogen peroxide was an interesting reagent for further testing.

Hydrogen peroxide

Since hydrogen peroxide (H2O2) is applied in a WFSR protocol for microwave-assisted digestion of seaweed,

further experiments were focused on determining the effect of H2O2 on the matrix. However, it was unknown

what the effect of treatment with exclusively H2O2 on seaweed could be. Therefore, initial testing was

performed in Erlenmeyer flasks as follows. Half a gram of seaweed was added to four 200 mL Erlenmeyer flasks, to which 15 mL of 5, 10, 15 and 30% H2O2 was added. This was left to stand for about 24 h at room

temperature. However, no air bubbles or other visual effects were observed. To get more information, the samples were filtered and the residues were weighed and stained with NR. Blanks (not stained with NR) were included as well for determining the autofluorescence of the seaweed. Higher concentrations of H2O2 were

expected to result in larger weight loss. However, 30% H2O2 resulted in only 2% more weight loss compared

to 5% H2O2, indicating there was no relation between the concentration H2O2 and the obtained weight loss.

Since all four experiments resulted in only about 20% weight loss as presented in Table 11, H2O2 at room

temperature was considered to have insufficient effect on the matrix.

Besides weight loss of seaweed after pre-treatment with H2O2 at room temperature, an increase in fluorescence

intensity after staining was observed, especially for the fluorescence filters WIB and WIG. Additionally, the increase appeared highest for the more polar solvents methanol and acetone. The weight loss and change in fluorescence intensity indicate that at least some reaction is taking place, although no visual signs of a presumed destruction process were observed during the experiment. Even though a maximum weight loss of only 22% was obtained, the matrix turned more transparent, allowing easier detection of MPs. The absence of air bubbles in the solution allowed a new set of experiments to be performed, aiming to achieve more weight loss of seaweed in the microwave by applying low concentrations of H2O2 and a higher temperature.

To increase the effect of the pre-treatment on the matrix, a new set of experiments was performed in the microwave with a concentration of 5% H2O2 at 80 °C, below the boiling point of H2O2. The time varied (10,

15, 20 min) while other parameters remained the same (15 mL reagent, 0.5 g seaweed). Post-treatment, the samples were filtered, weighed and stained, including blank stains. Merely a maximum weight loss of 46% was obtained as shown in Table 11, despite of the increased temperature. Moreover, the fluorescence intensity increased after staining in methanol and acetone, especially for filter WIG. It is possible that a more polar compound is formed, given that filter WIG shows emission in the red spectrum and acetone and methanol are more polar, possibly allowing NR to penetrate the seaweed better. Although the weight loss of the matrix increased under the given circumstances, the effect was insufficient. To determine the effect on the matrix of an increasing concentration, more experiments were performed with 5, 10 and 15% H2O2. However, when a

sample is in the microwave, it cannot be visually inspected during the process. To prevent the volume to be increased too much by e.g. air bubbles, the temperature was decreased to 60 °C. Additionally, the effect of constant application of microwaves was determined by applying a gradient. Therefore, the temperature was first ramped to 40 °C and then allowed to reach 60 °C in 26 min. Other parameters were unchanged. Again, residues were weighed and stained, including blanks. All photos of the experiments are shown in Appendix H: Autofluorescence of seaweed after H2O2 treatment and Appendix I: Seaweed stained with NR after H2O2

treatment. No clear differences in weight loss were obtained, with a maximum of only 42%. Furthermore, increasing concentration did not result in increasing weight loss, which can be observed from the data in Table 11. The effect of a gradient was also unclear, since a gradient resulted in more weight loss when combined

(30)

with 5% H2O2, but resulted in decreased weight loss for 10 and 15% H2O2 compared to a treatment without a

gradient of temperature increase. Visual inspection indicated a similar increase in fluorescence intensity after staining with methanol and acetone in filter WIG as for the other experiments with H2O2. Up to this point the

effect of H2O2 on seaweed was still unclear, since no correlation between temperature, time, concentration of

H2O2 and weight loss of the matrix appeared to exist. As air bubbles are an indication of a reaction taking

place with H2O2, a final set of experiments was performed to determine if H2O2 has any effect on the matrix.

A last set of experiments was performed to determine whether H2O2 was suitable as a reagent, in a quick a

simple way. Therefore, Erlenmeyer flasks were placed on a heating plate, allowing to observe air bubbles more easily. The heating plate was adjusted to achieve a temperature of 40 °C in 15 mL of water to ensure sufficient heating of the samples. Concentrations of 5, 10, 15 and 30% H2O2 were tested, other parameters

remaining unchanged. Every 5 min the Erlenmeyer flasks were swirled and the total time was about 30 min. However, no visual indication such as air bubbles was observed. To ensure more sufficient heating and allow stirring of the samples, the experiments were repeated in glass tubes on a magnetic stirring plate in a heated water bath at 40 °C. After 30 min, still no effect was observed. Therefore, they were not weighed or stained. Considering the rigid structure of seaweed, it is possible that H2O2 is not able to hydrolyse the polysaccharides.

Therefore, H2O2 was considered unsuitable as reagent for sample pre-treatment and further testing was

performed with HNO3.

4.2.2 Method development – microwave procedure

With a suitable reagent determined, the next step was to establish the optimal temperature, time and concentration of the reagent. Other parameters remain unchanged (15 mL reagent, 0.5 g sample), allowing the best comparison between the experiments. To gather most information, each parameter is optimized along different levels: temperature at 80, 100 and 120° C; time at 10, 15 and 20 min; concentration HNO3 at 0, 0.1,

1 and 2 M. This results in thirty-six experiments to gain full information. However, to save time, it was decided to work with a fractional Design of Experiments (DoE). This design included every combination of different levels of every pair of two parameters. Twenty-four experiments were performed, of which the results will be discussed next, considering the weight loss (%) and the (auto)fluorescence of the matrix.

Table 11: Weight loss (%) of seaweed after treatment with H2O2 under various circumstances.

Gradient: MW program with temperature increase from 40 to 60 °C.

% H2O2 Temperature (°C) Time Weight loss (%) Notes

5 Room temperature 24 h 20 Erlenmeyer

5 80 10 min 38 Microwave

5 40-60 30 min 36 Microwave, gradient

10 40-60 30 min 28 Microwave, gradient