Assessing the Mass Concentration of Microplastics and Nanoplastics in Wastewater Treatment Plants by Pyrolysis Gas Chromatography–Mass Spectrometry

The level of microplastics (MPs) in wastewater treatment plants (WWTPs) has been well evaluated by the particle number, while the mass concentration of MPs and especially nanoplastics (NPs) remains unclear. In this study, pyrolysis gas chromatography–mass spectrometry was used to determine the mass concentrations of MPs and NPs with different size ranges (0.01–1, 1–50, and 50–1000 μm) across the whole treatment schemes in two WWTPs. The mass concentrations of total MPs and NPs decreased from 26.23 and 11.28 μg/L in the influent to 1.75 and 0.71 μg/L in the effluent, with removal rates of 93.3 and 93.7% in plants A and B, respectively. The proportions of NPs (0.01–1 μm) were 12.0–17.9 and 5.6–19.5% in plants A and B, respectively, and the removal efficiency of NPs was lower than that of MPs (>1 μm). Based on annual wastewater effluent discharge, it is estimated that about 0.321 and 0.052 tons of MPs and NPs were released into the river each year. Overall, this study investigated the mass concentration of MPs and NPs with a wide size range of 0.01–1000 μm in wastewater, which provided valuable information regarding the pollution level and distribution characteristics of MPs, especially NPs, in WWTPs.


S1
Text S1. Determination of standard curves of selected plastics.
Direct weighting of certain quantities of the polymers was impossible because of uncertainties of the weighting of small quantities when preparing the lower calibration levels. The preparation of different calibration concentrations from a stock solution by serial dilution was conducted. Due to the poor solubility of several polymers like PE and PP in the most common solvents, the mixture of dichloromethane and methanol was used to disperse seven plastic polymers. Different plastic polymers have different densities, thus dichloromethane and methanol were mixed in certain proportions yielding a solution with a density close to that of these polymers (Table S2). 100 mg polymer powder was weighed and dissolved in 10 mL mixed organic solvent to obtain a plastic dispersion with a concentration of 10 g/L. The stock solution was continuously diluted to obtain 2−1000 mg/L (2,4,20,40,200,400, and 1000 mg/L) plastic dispersion.
The 50 μL standard plastic solution was transferred to an 80 μL pyrolysis cup and dried at 60 ℃ in a drying oven to obtain different calibration concentrations of different plastic polymers (Table S4).  (Table S4). The relative standard deviations (RSDs) of the quantitative ion peak area with 5 replicates for each standard sample were used to evaluate the precision of this method measured by Py-GC/MS.
The collected samples were filtered with 50 μm and 1 μm stainless steel meshes on a glass filtration unit. Next, the filtrate collected in a water tank was injected into a cross-flow ultrafiltration system with a molecular weight cutoff of 100 KDa. The clean water was discharged and the wastewater was injected into the tank again to achieve the concentration of raw water. After approximately 2 h, water sample was filtered and concentrated into a volume of approximately 200 mL. The final retentate was concentrated approximately 250 times. Additionally, cross-flow ultrafiltration can remove most of the dissolved organic matter because the molecular weight of dissolved organic matter in the aquatic environment is almost below 100 KDa. [1][2][3][4] In order to avoid background contamination and cross-contamination, the ultrafiltration cartridge was injected with ultrapure water and cleaned for 10 min prior to water sample concentration, and replaced when concentrating another water sample.

Text S3. Selection of indicator ions of different plastic polymers
As benzene (m/z 78) shows the highest peak intensity and sensitivity while other components have much low sensitivity, benzene was selected as an indicator of PVC. 5,6 For PMMA, methyl methacrylate (m/z 100) was selected as an indicator compound which is specific and high-sensitivity. 5-7 2, 4-Dimethyl-1-heptene (m/z 126) was specific and considered as an indicator ion for PP, 8 and the molecular ion m/z 43 with a high response was selected as the quantification ion. For PS, styrene monomer shows a high peak intensity but it may be produced from environmental matrixes like chitin and albumin, so the specific styrene trimer (5-hexene1, 3, 5-triyltribenzene, m/z 312) was considered as an indicator compound and the high-response molecular ion m/z 91 was selected as the quantification ion of PS. 5, 7, 9 1-decene was selected as the indicator ion for PE as it is the most representative pyrolysis component with high sensitivity. 6,10 For PET, benzoic acid (m/z 105) with high peak intensity and sensitivity was selected as an indicator component. 5, 10, 11 ε-caprolactam (m/z 113) is a specific pyrolysis product with high sensitivity, so it was considered as an indicator ion for PA. 12,13 Quantification of MPs and NPs by Py-GC/MS is based on the indirect determination by analyzing their pyrolysis products, but these products also be produced from natural matters present in water samples. The selectivity of the indicator compounds was tested by analyzing several selected organic materials including wood, leaf, fish, humic acid, S4 and black carbon (Table S6). 5,10 The indicator ions for PMMA, PP, PE, PS, PET and PA were not affected by these natural materials. 5,10 It cannot confirm that the pretreatment can completely remove the interference for PVC because benzene was abundant in most natural matters. 5,6,10 Thus the quantification of PVC in the water samples was potentially subject to minor bias from natural materials.

Text S4. Extraction and recovery efficiency of MPs and NPs.
In order to determine the sample process efficiency, an extraction test was performed using PET, PS and PP with high, medium and low density.  Figure S2. Total ion chromatogram pyrograms and the mass spectra of the characteristic pyrolysis products of six selected plastics. S10 Figure S3. Chromatograms of selected natural polymers. S11 Figure S4. Chromatogram of a representative blank sample. Figure S5. Chromatograms of representative samples with different size range. S12 S13 Figure S6. Results of similarity analysis the characteristic peaks of a representative sample. S14