A Method to Replace NaCl as a Flotation Solution for Extracting MPs in Soil: A Case Study of the Jiaxing Agricultural Soil from China

Microplastics (MPs) have become an important global issue in recent years. However, MPs in the soil have received far less attention than water. Effective and nondestructive extraction of MPs is important for studying MPs in agricultural soils. This study uses different floatation solutions as experiments and uses MgCl2 as the floatation solution of the density extraction method. Five types of standard MPs (PE, PP, PS, PVC, and PET) are used as the objects of this experiment. The recovery of the two particle sizes was between 90.82% and 109.69%. The extracted standard MPs were then subjected to IR and Raman spectroscopic analysis, and the results showed that Raman spectroscopy was more suitable for the identification of the extracted MPs. Finally, this method collected and verified a vast number of soil samples and further analyzed the abundance and characteristics of the collected MPs.

China is a largely agricultural country with a large annual demand for plastic film mulch (Liu et al. 2014;Huang et al. 2020). The increasing annual use of mulch has led to plastic film residues being one of the main sources of MP contamination in agricultural soils Zhou et al. 2020). MPs in soil ecosystems may also cause various ecological risks (Li, et al. 2019b). Hazardous substances contained in plastics such as PBDEs, bisphenol-A, phthalates (e.g. diphthalate), and heavy metals are readily released into the environment (Wang et al. 2013;Zhou et al. 2019;Zhou et al. al. 2020). At the same time, MPs in the soil can also affect the detection of basic soil properties (WoongKim et al. 2021). MPs in the soil also adsorb various hydrophobic organic substances, such as pesticides, fungicides, PAHs, PCBs, and other pollutants, during degradation (He et al. 2018;Sarker et al. 2020). In addition, MPs are also enriched in soil organisms, causing negative growth effects on soil organisms and plants Sheng et al. 2021). Eventually, these enriched MPs can enter the human body through the transmission of the food chain. Residual MPs have been found in human feces, but the health risks posed by MPs to humans are still unknown due to limited available data (Yang et al. 2021;).
Current methods for the extraction of MPs from the soil are mainly density extraction methods (Han et al. 2019). MPs and soil are separated from each other in the flotation solution 96 Page 2 of 8 because of their different density characteristics. The lessdense MPs are suspended and then filtered or separated for extraction purposes (Besley et al. 2017). The prevailing flotation solution chosen for the density extraction method is a saturated solution of NaCl (Mai et al. 2018). However, the density of NaCl saturated solution is only 1.2 g/cm 3 , which makes it impossible to separate denser MPs from the soil (Zhang et al. 2020a, b). Therefore, previous studies have also investigated other large-density flotation solutions, such as zinc chloride (ZnCl 2 ), calcium chloride (CaCl 2 ), sodium bromide (NaBr), sodium iodide (NaI), etc. However, there are also problems associated with these methods to a greater or lesser extent. For example, both zinc chloride and sodium iodide are relatively toxic to the environment and cannot be discharged directly into the environment (Kumar et al. 2020). Calcium chloride substances containing divalent Ca 2+ ions can agglomerate organic material and thus interfere with subsequent identification (Scheurer and Bigalke 2018). In addition, these methods are relatively expensive. Therefore, nontoxic, low-cost, high-density, and efficient methods for the extraction of MPs are needed.
Once isolated from soil, MPs are typically analyzed using microscopy to count putative MPs and determine their color, shape, and sizes, with data reported as particles per volume or mass. When done manually, this approach is prone to both counting and misidentification errors. For identification, increase use of automated techniques, such as FTIR and Raman imaging are being pursued (Vandermeersch et al. 2015;Wiesheu et al. 2016). Such chemical imaging identifies MPs based on the vibrational energy of specific functional groups. These methods work particularly well on the μm scale and have the advantage of being non-destructive.
In view of the limitations of previous methods for the extraction of MPs, it is necessary to develop a new method. The objectives of the current study were to (1) develop and evaluate MgCl 2 as a new method for extraction MPs from soil samples, (2) optimize extraction and detection procedures such as extended suspension times and use Raman analysis that minimize loss of microplastics and maximize their accurate identification, and (3) demonstrate the feasibility of the method by examining microplastic pollution in agricultural soils in Jiaxing of China. We focused on five types of plastics commonly found in environments (PE, PP, PS, PET, and PVC) and validated the method by spiking with known amounts of microplastics and evaluating recoveries. Herein we describe the method and present the abundance and characteristics of the collected agricultural soil MPs in Jiaxing.

Preparation of Chemicals and Materials
Reagent-grade chemicals NaCl, MgCl 2 , KCl, NaHCO 3 , KH 2 PO 4 , and H 2 O 2 (35%) were purchased from XILONG SCIENTIFIC, and their basic properties are presented in Table S1. Saturated solutions were prepared by dissolving the five crystals in excess in distilled water. Particles of various sizes of uncolored polyethylene (PE), polypropylene (PP), polystyrene (PS), polyvinyl chloride (PVC), and polyethylene terephthalate (PET) plastic particles (3 mm pellets and 80 µm powder) were obtained from Taobao. These particles are the five most common MPs on the market. Pure, plastic particle-free sand was provided by the Zhejiang Academy of Agricultural Sciences. The sand was sieved and panned multiple times to ensure no microplastics are present. Real samples were collected from agriculture in Pinghu, Jiaxing ( Fig. 1), at coordinates of 121.055894°/30.702205°. Surface samples of the top 3 cm measuring 0.25 m 2 square were collected with a stainlesssteel tool and stored. The main characteristics of the soil were as follows: pH: 6.46, organic matter: 25.70 g-kg −1 , organic phosphorus content: 68.72 mg/L, and fast-acting potassium concentration: 305 mg/L. During the collection process, large objects such as stones, organisms, and plants were removed. The soil was dried in an oven at 60°C for 48 h and then sieved to remove any large particles over 5 mm. Clean sand was used for spiking experiments, and real soil samples were used to determine their abundance and characteristics.

Recovery Experiments for the Flotation Solution
As the MPs selected were of large and small particle sizes, the recovery experiments and the calculation of recovery rates were slightly different for the two particle sizes, as detailed below.
Large particle size (3 mm pellets): Pure plastic-free sand (50.000 g) was weighed and respectively mixed with 3 mm diameter PE, PP, PET, PVC, and PS MPs (50 particles in total). Only 10 particles of each type of MPs. They were transferred into five clean glass beakers, and then 200 ml of the saturated solutions are added. The mixture was sonicated for two minutes and stirred with a glass rod for ten minutes and then left to stand for 8 h, after which the supernatant was separated and filtered through Whatman glass fiber filter paper (0.45 μm, GF/A)) and then this repeated three times. After extractions and filtration, the recovered material was calculated by counting the particles. Three replicate samples were then tested.
Small particle size (80 µm powder): Pure plastic-free sand (50.000 g) was weighed and respectively mixed with 0.5 g of PE, PS, PP, PVC, and PET MPs with a particle size of 80 µm and transferred into a clean glass beaker. Besides 50.000 g pure plastic-free sand was used as a control. This was then added to 200 mL of saturated salt solution. The mixture was sonicated for two minutes, stirred with a glass rod for ten minutes, and then left to stand for 8 h, after which the supernatant was separated and filtered through Whatman glass fiber filter paper (0.45 μm, GF/A), with this operation being repeated three times. The filtered particles were thoroughly washed with deionized water. After these filtrations and rinses, the filter paper is placed in an oven to dry (24 h, 60°C). Finally, the recoveries were calculated by weighing the filter paper. Three replicate samples were then tested. The percentage recovery of MPs spiked in artificially spiked soil can be calculated using Eq. (1). .
Equation 1 where P is the recovery of the MP sample (%); m 1 is the mass of impurities extracted from the unspiked soil sample (g); m 2 is the mass of trace plastic and impurities extracted from the manually spiked soil sample (g); and m 0 is the spiked mass of MPs (g).
In previous literature, a settling time of approximately five hours for the flotation solution can make the mixed sand/ salt solutions clear (Besley et al. 2017). As shown in Fig.  S1, during the experiment, we found that the soil did not settle below the solution very quickly. After 8 h of settling, the flotation solution changed color from yellow to colorless and then stabilized. Therefore, in subsequent experiments, we used a suspension time of 8 h. Subsequently, to remove organic matter, the filtered MP was placed in 30% hydrogen peroxide and ablated at 60°C for 48 h until the solution was elucidated (Nuelle et al. 2014;Tagg et al. 2015). Finally, the suspended MPs were re-filtered and dried.
We also placed the MPs in a scanning electron microscope (TM-3000, Hitachi, Japan) to see if the surface of the MPs had been damaged. Finally, we examined the MPs using FTIR (Nexus 8700, Thermo, USA) and Raman spectroscopy (Instrument: LabRAM HR Evolution, Detector: SIN-EM FIVIS) to determine whether the extraction impacted the spectra and the ability to identify plastic.

Qualitative and Quantitative Analysis of the Extracted Agricultural Soil Samples
The developed method was used for extracting areal agricultural soil samples to demonstrate its usefulness. MP was extracted from three 1 kg Jiaxing agricultural soil samples using the procedure in step 2.2. The extracted MP was collected using a glass fiber filter paper, which was deposited in a clean glass dish and then dried in an oven. To measure and calculate the water content, each freshly collected soil sample (separated from the sample at the same time as the 1 kg fraction described above) was weighed to 20 g and stored in a precleaned aluminum box, then covered with aluminum foil and allowed to air-dry at room temperature for more than 14 days till a constant weight was achieved(The temperature is 25 degrees and the humidity is 68%). The water content was then calculated as the mass loss after air-drying (Schmugge et al. 1980). The dry weight of the collected soil samples was obtained from the water content, and the abundance of soil MPs was calculated from the dry weight (Zhou et al. 2020). Use the ×25-×100 magnification range of a stereomicroscope (Axio Lab. A1, ZEISS, Germany) to scan the residue for all possible microplastics, the smallest observable particle size is about 30 μm. The MPs were observed and counted and their characteristics, including size and shape, were recorded using a digital camera. The collected MPs were identified by matching using a Raman spectrometer and identified from a database of spectra. The polymer type was confirmed when the similarity was above 60%. The parameters of the Raman spectrometer were set as follows: Before analysis, the Raman instrument was properly calibrated with silicon wafers. The settings used in Raman spectrum measurement were as follows: range (cm −1 ) 50-4000, AE level (cnts) 50,000, Acq. time(s) 8, accumulations 2, RTD time(s) 1. Measurements were performed with a 785 nm laser and a 50 × objective lens (Olympus, Japan).

Quality Control and Assurance
The collected samples of agricultural soil from Jiaxing are transferred by iron spoons to iron basins. Then they are dried by using an oven. All samples for MP analysis should not come in contact with any plastic material. Nitrile gloves and cotton lab coats were worn during the experiments to minimize the contamination of MPs. Instruments used for analysis were rinsed 3 times with ultrapure water and covered with tin foil before and after use. Blank controls were performed simultaneously with each analytical experiment to avoid contamination by airborne particles or experimental reagents. All experiments were performed in triplicate, and three replicate procedural blanks were performed in each experiment.

Data Analysis
Spectral library analysis of the MPs was performed using LabSpec6 software. All results on MP abundance were expressed as the number of MP particles per dry mass of soil (number (N) kg −1 ). Data analysis was performed using WPS office 2019 and OriginPro 8.5.

Evaluation and Selection of the Extraction Scheme
The extraction recoveries of the five salt solutions for the five large MPs (3 mm pellets) are shown in Fig. S2. From  Fig. S2, it can be seen that all five salt solutions showed good extraction results for the three less-dense MPs (PP, PS, and PE). However, in the extraction of PVC and PET, the extraction efficiency of all salt solutions except MgCl 2 was low. The extraction recovery of saturated NaCl, which is widely used for extraction, was 0%. The first extraction of MgCl 2 was able to recover most of the MP, but not all of the MPs were completely recovered. Therefore, more extractions were necessary. After three extractions, the low-density MPs (including PE, PP, and PS) were completely recovered by MgCl 2 , while the recovery of the other two MPs reached more than 91%. These results suggest that MgCl 2 saturated solution has a significant advantage over other salt solutions for the extraction of MPs. Therefore, MgCl 2 was selected as the most suitable flotation solution for the extraction in the subsequent small particle size experiments. The recoveries of MgCl 2 for the two different particle sizes of MPs are shown in Fig. S3. The results show that MgCl 2 also has a good extraction effect for small MPs. Therefore, we used MgCl 2 -saturated solutions as flotation solutions to extract MPs from agricultural soils.

Apparent Spectral Changes of MPs After Extraction
The effect of the extraction process on color and shape is also important, as these features provide insight into the origin of the MPs and the presence of additives. We did not observe significant differences in color between MPs extracted by MgCl 2 . Scanning electron micrographs showed that the surface morphology of 5 MPs hardly changed before and after treatment (Fig. 2). However, we found many tiny particles attached to the surface of the treated MPs. Figure  S4 shows the Fourier-transform infrared spectra of common polymers and MgCl 2 before and after processing. We can see from the infrared spectrum that there is a strong peak at approximately 1700 cm −1 for MgCl 2 and a strong outgoing peak near 3500 cm −1 due to the strong water absorption of MgCl 2 (Kuptsov Zhizhin 1998). We can see that the spectral signal of the microplastics is masked by the MgCl 2 and water signals. This phenomenon demonstrates that there are many tiny particles on the surface of the treated microplastics that are MgCl 2 crystals. It also shows that the method cannot identify the extracted MPs by FTIR spectroscopy. In addition, Raman spectroscopy is also commonly used to identify changes in the surface of polymers (Zuccarello et al. 2019;Dong al. 2020). Therefore, we also examined the MPs using Raman spectroscopy after sample extraction. As seen in Fig. S5, the extracted MPs can be identified accurately (similarity > 86.5%). Therefore, the Raman technique was chosen as the method for identifying MPs in the actual sample analysis that follows.

Abundance and Characteristics of the Soil Samples
In this study, agricultural soil in Jiaxing, China, was selected to extract MPs to verify the practicality of the method. A procedure blank was also set up during this extraction. We did not observe any suspected MP fragments extracted in the blank filter paper, a result that indicates a low level of laboratory contamination. The MPs found suggested a variety of shapes and colors. As shown in Fig. S6, we classified them into four categories based on the relative frequency of shapes: fibrous, thin film, and other (granules and foam). Of these, film-based MPs were found most frequently at 40.43%, followed by the fragment category at 28.72% and fibers at a relatively small percentage of 17.02% (Fig. 3a). We also found a small number of pellets and foams. We also observed many colors of microplastics. Their frequent occurrence is in black, transparent, white, red, and yellow colors. The vast majority of the MPs were black, transparent, and white (97.88%), while other colors accounted for less than 3% (Fig. 3b). The obtained fragments were in different colors, which refers that the particles originated from a variety of sources.
We further investigated the types of MPs in agricultural soils using Raman spectroscopy. In previous literature, PE and PP were the most common types found in agricultural soils (Yu et al. 2021), and similar results were found in this study. Raman spectra are shown in Fig. S7. The screened fragments were analyzed and compared. As identified by Raman spectroscopy, the most abundant of MPs agricultural soils were PE and PP, with 37.23% and 28.72%, respectively (Fig. S8). PE and PP are abundantly used in mulch, food packaging products, grocery, and garbage bags, etc. (Laura  Wang et al. 2022) We also found many PUs, with a percentage of 12.77%. Polyurea (PU) is often used as a shell material for pesticide microcapsules due to its easy production method and stable nature ). In addition, polysulfone (PSF), polystyrene butadiene (SBR), and acrylonitrile butadiene styrene (ABS) have all been found. PSF is commonly used in film preparation . SBR and ABS are also materials commonly used in industry, for example, as shells for electronic devices, and have also been found in soil (Chai et al. 2020;Scheurer and Bigalke 2018). In 99% of the samples, the MPs were smaller than 1 mm in size (Table 1), which had a significant impact on the identification analysis. MPs (1-5 mm) are gradually broken down into smaller MPs (< 1 mm) through various physicochemical and biological interactions. Therefore, smaller MPs will gradually increase. However, smaller MPs pose higher ecological risks because of their larger specific surface area. They are more likely to attract and release heavy metals, and organic pollutants from the environment. There is therefore a need to further investigate the contamination of MPs in agriculture local regions and assess their impact on agricultural ecology. These objectives will be investigated in future work. This study aimed to provide an effective method for extracting MPs from agricultural soils and was validated with actual agricultural samples.

Conclusions
In this study, we developed and validated a method for the separation and extraction of soil MPs based on MgCl 2 density solutions, providing a valuable tool for the separation and extraction of MPs of various densities from the soil.
(1) The best flotation solution was selected from five flotation solutions by subjecting to extended suspension time and multiple extractions. (2)MgCl 2 maintained a recovery of more than 91% for the five different particle size MPs with high separation and extraction efficiency. (3) Raman spectroscopy is more suitable for identifying MPs in this method than FTIR. (4) The feasibility of using this method was verified by applying it to actual agricultural soil samples. (5) The film PE MPs were found most frequently in the Jiaxing agricultural soil. It aspect confirms the abuse of agricultural mulch. It is suggested that the government should further standardize the use of agricultural mulching film, to establish a reasonable mechanism for ecological environmental protection.
Overall, the method has good application prospects and demonstrates great promise for extracting and analyzing MPs from soils. It is noteworthy that the method has not been applied to other types of soils. We will add more soil samples to increase the applicability of the method in the future.