Comparison of pre-treatment methods and high-density liquids to optimize the extraction of microplastics from natural marine sediments


Sediment sampling

Ten sediment cores were collected at a station in Askeröfjorden, N58°5′21′′ E11°48′6″ (center), outside Stenungsund, Swedish west coast, on October 24, 2018, at the using a Gemini corer. The sediment was characterized according to the recommendations of the Swedish Geotechnical Society42 as a sandy, silty gyttja-clay, bioturbation and benthic macrofauna species have been identified, with the highest abundance of the species Amphiura spp. and Arctica islandica. The top 2 cm of all carrots were pooled, homogenized in a 10 L stainless steel pot with a stainless steel spoon, and transferred to 15 glass containers with glass lids, each with a volume of 200 mL ( about 314 g) (Table 1). All samples were stored at 8°C until further analysis.

Table 1 Microparticles identified in blank samples and sediment samples treated with inorganic chemicals, enzymes and without treatment for the two density separation solutions, ZnCl2 and NaCl.

Sediment treatment

Sediment samples were weighed before a 5 mL subsample was collected and dried at 105°C for 24 h for water content analysis. Sediment samples (n=15) were divided into three groups, one group was treated with inorganic chemicals (n=4), one with enzymes (n=5) and the third remained untreated (n = 6). Treatment with inorganic chemicals is based on a protocol developed by Strand and Tairova 201622 but slightly modified as it consisted of a mixture of 0.67 mol/L NaClO, 0.45 mol/L KOH and 0.022 mol/L Na4P2Oseven. About 400 ml of the chemical treatment (twice the volume of the sediment) was added to glass vessels with glass lids along with the sediment. The samples were incubated at room temperature for 1 h on an oscillating shaker table at 160 rpm. An additional rinsing step was necessary after the chemical treatment to lower the pH and avoid the precipitation of zinc hydroxides. Subsequently, the sediment was washed by adding Milli-Q water followed by vigorous shaking and centrifugation for 30 min at 1000 rpm. This washing procedure was repeated three times. The supernatant was removed between replicates and new Milli-Q water was added. The supernatant was filtered through a 50 μm filter, which was saved for further analysis. After centrifugation, the sediments were transferred back to the glass containers. Porcine pancreatic enzymes have been used for the enzymatic digestion of organic matter43. A buffered enzyme working solution was prepared by dissolving one capsule of the pharmaceutical enzyme (Creon 40,000, Abbott Laboratories GmbH, Germany, Mylan) for 10 ml of Tris hydrochloride solution (Trizma, pH 8.0, 1 M , filtered at 0.2 μm, Sigma-Aldrich, T3038, USA). Complete dissolution was achieved by gentle heating (30°C) and shaking at 125 rpm for 30 min on a heated incubation shaker (New Brunswick Scientific, Innova 40). Ten ml of the working solution was added per g wet weight (WW) of sediment to a prewashed glass jar. The sediment-enzyme solution was mixed and the pH was assessed using a pH indicator stick. The pH was adjusted to 8.0 by adding more Tris hydrochloride solution if necessary. The filled glass vessels were capped and placed under vigorous shaking (150 rpm) at 37.5°C overnight on the incubation shaker.

The level of degradation was quantified by TOC analysis using an elemental analyzer coupled to an isotope ratio mass spectrometer (20–22, Secron Ltd., Crewe, UK) before and after treatment of degradation.

Separation and filtration by density

To compare the efficiency of two widely used density separation solutions, NaCl (density 1.2 g/cm3) and ZnCl2 (1.8 g/cm3), the pretreated samples were divided into two groups subsequently treated with ZnCl2 (n=7) or NaCl (n=8). Density separation was performed using the Kristineberg Microplastic Sediment Separator (KMSS). This divider was designed in-house based on the Munich plastic divider19but it is smaller, both in height and width, with a steeper incline on the riser with a glass cylinder above the sediment pan to monitor sedimentation (Fig. 1).

Figure 1

The Kristineberg Microplastic Sediment Separator (KMSS) designed after the Munich Plastic Separator (Imhof et al., 2012). Part 1, sediment vessel with rotor and lower valve, part 2, glass cylinder, part 3 riser and upper part separation chamber with ball valve and filter holder.

After pre-treatment, the sediment samples were moved to the lower sediment container of the KMSS, the riser portion of the tower was mounted, and the rotor, positioned at the bottom of the tower, was started. The saline solution was introduced through the lower valve and filled to 85% of the tower volume. The rotor was stopped three hours later and the sediment allowed to settle. After 12 h of settling, the split chamber was mounted above the riser and the tower was filled with the density separation fluid, ZnCl2 or NaCl. When filled, the ball valve was closed, the liquid level dropped, and the dividing chamber was removed. Then the dividing chamber was inverted and the solution was filtered. Sediment exposed to pretreatment with inorganic chemicals and separated with ZnCl2 was filtered through 300 μm polyamide (PA) mesh and the remaining overhead solution was collected, and a second separation with ZnCl2 in a glass beaker was carried out due to the high mineral content. After second density separation (shaking and settling for 24 h), the solution was filtered through a 100 μm mesh PA filter. For enzymatic and untreated samples as well as all samples extracted with NaCl, only one separation was performed because fewer mineral particles were extracted compared to pretreatment with inorganic chemicals and separation with ZnCl2. The solution was filtered through two PA filters with 300 μm and 100 μm mesh. After filtration, the filters were rinsed with Milli-Q water to remove salt crystals.


All equipment used during sampling was cleaned and rinsed in the laboratory with Milli-Q water before being dried in a plastic-free fume hood. For the storage of sediment samples, glass containers with glass lids were used, and a stainless steel pot and a stainless steel spoon were used for sediment homogenization. All the samples were stored in their glass bottles until they were processed in the laboratory. All lab equipment and surfaces were cleaned prior to lab work, and only cotton lab coats and clothing were worn. ZnCl2 was filtered through three membranes 10 μm, 5 μm and 1 μm. The NaCl was filtered through 10 µm.

Six sediment-free blank samples were processed, one for each combination of pretreatment and density separation solution. All blanks were treated as their corresponding sediment sample from bottle cleaning prior to sampling, in the field, during storage, processing, separation and analysis.


Analysis workflows followed current consensus guidelines, published or available as working drafts from Regional Seas Conventions and the European Commission10,14,32. All filters were first visually inspected with a stereomicroscope (Leica M205 C 80–160× Wetzlar, Germany). Subsequently, all particles visually identified as suspected to be anthropogenic according to Karlsson et al. 202044 were transferred with tweezers to different aluminum oxide filters with a pore size of 200 nm (Whatman Anodisk 25). One filter for each treatment and size fraction, i.e. particles collected on the 300 μm mesh were moved to an aluminum oxide filter and particles from the 100 μm were moved to a other. The particles collected from the centrifugation solution were also transferred to a filter. Two filters were used if there were more suspicious particles than could fit on one Anodisk filter. The entire surface of the Anodisk filters with particles was imaged with an optical microscope (Zeiss, AxioImager). All particles (collected from the 300 μm filters, 100 μm filters and from the spin solution, 50 μm) were characterized according to their visual appearance, according to Karlsson et al. 202044 where their 2D shape, 3D shape, solidity, color and classic visual identification were noted. Subsequently, all particles from the 300 μm filter and all particles from eight randomly selected 100 μm samples were chemically identified by Raman microscopy (Witec, alpha 300R) using a 532 nm laser and of a network of 600 g/mm. Laser power was selected based on polymer, signal intensity, particle size, and magnification. Spectra were measured in a wavenumber range of 200 to 3500 cm−1 and were compared to our internal library for identification (HQI minimum 75, majorities greater than 80). The library consists of spectra we have obtained from known plastic particles, including weathered particles, and spectra from the RUFF45 and ST Japan spectral databases. For comparison, 300 µm filters from two randomly selected samples were also chemically identified by Fourier transform infrared spectroscopy, FTIR (Thermo Scientific Nicolet iN10) using transmission mode (256 scans, resolution 4 cm−1spectral range 4000–675 cm−1, detector cooled by liquid nitrogen and correlated to 256 background scans). Particles in the supernatant of the centrifugation solution (all inorganic pretreatments and one of the pretreated enzymes) were first characterized visually before being analyzed by Raman microscopy. The remaining particles on the PA filter (300 μm, 100 μm, and 50 μm of the spin solution) from two random samples, i.e. particles that were not visually identified as anthropogenic particles, were were analyzed by Raman microscopy to identify false negatives.


Comments are closed.