Elsevier

Chemosphere

Volume 225, June 2019, Pages 810-819
Chemosphere

Introducing a soil universal model method (SUMM) and its application for qualitative and quantitative determination of poly(ethylene), poly(styrene), poly(vinyl chloride) and poly(ethylene terephthalate) microplastics in a model soil

https://doi.org/10.1016/j.chemosphere.2019.03.078Get rights and content

Highlights

  • Relationships of soil components represent soil universal model indicators.

  • Soil universal models can replace soil standards.

  • Deviations from soil universal model indicators indicates soil contamination.

  • Indicators are promising for qualitative and quantitative determination of studied μP (microplastics) except poly (ethylene).

Abstract

Methods for analysis of microplastic in soils are still being developed. In this study, we evaluated the potential of a soil universal model method (SUMM) based on thermogravimetry (TGA) for the identification and quantification of microplastics in standard loamy sand. Blank and spiked soils (with amounts of one of four microplastic types) were analyzed by TGA. For each sample, thermal mass losses (TML) in 10 °C intervals were extracted and used for further analysis. To explain and demonstrate the principles of SUMM, two scenarios were discussed. The first refers to a rare situation in which an uncontaminated blank of investigated soil is available and TML of spiked and blank soils are subtracted. The results showed that the investigated microplastics degraded in characteristic temperature areas and differences between spiked and blank soils were proportional to the microplastics concentrations. The second scenario reflects the more common situation where the blank is not available and needs to be replaced by the previously developed interrelationships representing soil universal models. The models were consequently subtracted from measured TML. Sparse principal component analysis (sPCA) identified 8 of 14 modeled differences between measured TMLs and the universal model as meaningful for microplastics discrimination. Calibrating various microplastics concentrations with the first principal component extracted from sPCA resulted in linear fits and limits of detection in between environmentally relevant microplastics concentrations. Even if such an approach using calculated standards still has limitations, the SUMM shows a certain potential for a fast pre-screening method for analysis of microplastics in soils.

Introduction

Microplastics, i.e. degraded debris or intentionally produced synthetic polymer particles of 100 nm to 5 mm size (Cincinelli et al., 2017), have recently become a recognized threat to aquatic environments including sediments and living organisms (Rochman, 2018). Currently, soil scientists are actively discussing also the potential pollution of soils by microplastics (Rillig et al., 2017). Microplastics might be introduced to the soil through air transport of particles and fibers, by unseparated microplastics via wastewater treatment plant (WWTP) sludge, littering or tire wear from roadwash (Bläsing and Amelung, 2018) and compost application (Hurley and Nizzetto, 2018). Nevertheless, the primary source of plastics in soil is probably plastic mulch (Steinmetz et al., 2016). Presence of at least macroplastic parts (from ca. mm to cm dimensions) is already confirmed in soil after the agricultural season with the use of plastic mulch (Liu et al., 2014, Ramos et al., 2015) or even in home gardens (Duis and Coors, 2016, van der Wal et al., 2011), therefore, plastic particles are considered as a marker of urban soils and Technosols (Rillig et al., 2017) and agricultural soil is now considered as a part of “Plastic Cycle model” (Horton and Dixon, 2018). Through all of this, it is estimated that 63–430 thousand tons of microplastics are annually added to European soils (Hurley and Nizzetto, 2018). Macro- and microplastics are presumed to have various effects on soil physicochemical properties. They may accumulate in the soil food web (Rillig et al., 2017), function as sorbent for pesticides (Ramos et al., 2015) or microorganisms (Kirstein et al., 2016), and influence soil aggregation and quality (Rillig et al., 2017), thereby, they are now recognized as a global environmental issue (Sintim and Flury, 2017).

Therefore, the acute call for rapid, high-throughput, accessible and reliable analytical methods for microplastics in the soil matrix has been made in several recent publications (Bläsing and Amelung, 2018, Horton and Dixon, 2018, Rillig et al., 2017, Silva et al., 2018), but this seems to be a serious challenge. Light microscopy, a commonly used method, requires a labor-intensive preconcentration procedure, otherwise their result may be false positives and misinterpretations (Lachenmeier et al., 2015, Woodall et al., 2015). Spectroscopic-microscopic techniques help to distinguish particles or fibers of synthetic origin from natural structures in the samples, i.e. plastics from cellulose or silicate materials. Yet, such techniques have been used for easier accessible and more homogeneous atmospheric or aqueous matrices, not for soil, and present more-or-less only qualitative results (Comnea-Stancu et al., 2016; Fischer et al., 2015, Ioakeimidis et al., 2016) Analyzing soils with such techniques again requires complex matrix separation methods and microplastics pre-concentration as well as time-consuming measurements and data evaluation, however a first automated insight was published by Primpke et al. (2017). In Raman microscopy, soil organic matter (SOM) tends to hide microplastics and show artifacts induced by SOM autofluorescence (Bläsing and Amelung, 2018). In contrast, analytical techniques based on sample pyrolysis have the potential to overcome the drawbacks of microscopics methods, both in terms of lowering sample preparation efforts and enabling semi-quantitative to quantitative analyses. The thermal stability and typical pyrolysis products of many polymers have already been characterized, e.g. (Dimitrov et al., 2013, Duemichen et al., 2014), but some of them can be identical to pyrolysis products of soil organic matter (Schulten and Leinweber, 1999). SOM is a mixture of a number of molecules such as fragments of parental biomass including phenolic and alkylaromatic compounds, alkanes, alkenes and carboxylates. First insight of a quantitative application of this approach were published by Dümichen et al. (2015). The authors applied a two-step procedure (1) pyrolyzing a Berlin urban soil spiked with poly (ethylene) (PE) microplastics via thermogravimetry (TGA) and capturing pyrolysis products followed by (2) thermal desorption-gas chromatography-mass spectrometry (TED-GC-MS) and later applied the same setup for so far only qualitative assessment of PET microplastics (Dümichen et al., 2017). Curie-Point pyrolysis-gas chromatography-mass spectrometry (Py-GC/MS) was proven as a method to analyze microplastics quantitatively down to 5 μg amounts, however this required in some cases enzymatic and chemical digestion and also density separation (Fischer and Scholz-Böttcher, 2017). An interesting, however not completely specific, quantitative approach was presented by Wang et al. (2017) for PET and poly(carbonate) (PC), utilizing chemical depolymerization of the polymers and following analysis of the depolymerization products (such as bisphenol A) via liquid chromatography–tandem mass spectrometry (LC–MS/MS). Finally in another approach, PET microplastics were quantitatively characterized by means of thermogravimetry-mass spectrometry (TGA-MS) in a standard loamy sand matrix (organic C content of 1.61 ± 0.15%) (David et al., 2018). When only thermal analysis methods were applied, the differential scanning calorimetry was so far able to distinguish between particles of PE and poly(propylene) (PP) only, moreover well separated by pretreatment from other types of polymers and matrices as well (Majewsky et al., 2016).

In contrast to analysis of microplastics in aquatic systems, where water can be relatively easily removed, the analysis in soil is more challenging. In particular, microplastics may be dispersed within the soil structure and either extraction procedures or sophisticated methods are needed for their analysis (Rillig et al., 2017). In this paper, we address this issue and introduce a simple approach that is able to qualitatively and quantitatively determine the microplastics in soil without sample pre-separations or pre-treatments. In this case, three basic scenarios exist. In scenario (i), no blank soil is needed and microplastics are determined directly in soil samples by available analytical instruments. The instruments may be the hyphenated techniques consisting of pyrolysis, which products are detected by mass spectrometry, e.g. TGA-MS (David et al., 2018) or TED-GC-MS (Dümichen et al., 2015). The use of this approach is restricted when the pyrolysis products and/or degradation temperatures of SOM and microplastics are identical. The scenario (i) is not discussed further in this paper. In scenario (ii), a blank soil is needed for the analysis and is available for comparison (a blank in this paper refers to the same soil that is under investigation, but without microplastics). In scenario (iii), such a blank soil is needed, but is unavailable. Here, we address scenarios (ii) and (iii).

In this work, we evaluated two approaches to analyze the microplastics in soils based solely on TGA analysis. TGA is a thermoanalytical method in which the mass of an investigated sample is monitored in dependency on time or temperature while the temperature is programmed (Rotaru et al., 2008). As a result, TGA fractionates soils based on thermal (carried out under inert atmosphere) or thermo-oxidative (carried out under a stream of air or oxygen) stability or degradability (White, 2019). TGA, alone or in combination with other techniques (e.g. MS), is gradually receiving attention in soil analysis as a method for determination of free and bound water (Wang et al., 2011), thermolabile and thermostable fractions (Peikert et al., 2015), mineral composition and content (Karathanasis and Hajek, 1982), presence of contaminants (Siewert and Kučerík, 2015), indicator of soil biological activity (Peltre et al., 2013) and others. Earlier results revealed a close connection between mass losses determined in 10 °C steps (TML, thermal mass losses) and content of soil organic carbon (SOC), total organic nitrogen (TON) and clay content (Siewert, 2004). These TMLs were shown to correlate with total SOM content (large TML [LTML] between 110 and 550 °C) (Kucerik et al., 2016) and SOM thermo-oxidative fractions (LTML between 200 and 300; 300–450 a 450–550 °C and their combinations) (Kučerík et al., 2018). Mutual correlations between TML and soil properties and LTML were found and verified by analysis of hundreds of soil samples of various types collected globally from surface horizons. Therefore, these correlations and extracted equations represent a soil universal model representing a universal soil standard. The soil universal model method (SUMM) proposed in this work is based on the determination of the difference between the modelling relationships between TML/LTML and measured values. More specifically, the method is based on the determination of the deviation between soil universal model indicators (reported further below), which are calculated from previously determined relationships (Siewert, 2004, Kučerík et al., 2018) and indicators, where they have been measured directly. The difference between calculated (or better predicted) and measured mass losses is indicative for the type and content of a contaminant. The use of simple soil universal model relationships has already been tested for identification of extraneous organic compounds and particles relatively untransformed via soil processes such as pyrogenic carbon (Kucerik et al., 2016) or organic amendments such as straw or manure (Tokarski et al., 2018).

The aim of this work is to introduce an advanced SUMM and demonstrate its application for microplastics identification and quantification in order to test the limits of this approach. For better understanding, we address the principles of scenarios (ii) and (iii) and analyze a model soil spiked by four types of polymers.

Section snippets

Model sample preparation

For the preparation of all samples, a standard loamy sand (Type 2.2, LUFA, Speyer RP, Germany) with 1.61 ± 0.15% organic C content and a dominant particle size of 0.02–0.63 mm was used. The soil was shaken and vortexed (mid speed, 3 min, Mini Analog Vortex; VWR, Radnor PA, U.S.A.) in glass vials with different standards of polymers with varying particle sizes: poly(ethylene terephthalate) (PET) cryo-milled (150–400 μm dominant particle size) microplastics from dust arisen from PET bottle

TGA records of spiked soil

TMLs were extracted and plotted on dependency of temperature for all microplastics at all concentrations. Fig. 1 shows the TML of blank LUFA soil and soil samples enriched with PET, PE, PVC, and PS at concentrations between 0.5 and 5 wt%. Fig. 1 shows significant changes in TGA plots after addition of microplastics in the temperature area between 30 and 550 °C. These changes were polymer-type-dependent in terms of temperatures intervals and rate of degradation, which were then further used for

Conclusion

In this work we performed a first thermoanalytical and statistical setup introducing the SUMM used for the determination of microplastic load in soil without any further detection techniques (which need to be used e.g. for scenario (i)) and without any analyte pre-separation). As could be seen, already the differences between TML of spiked and blank soils produced reasonable calibration curves (R2 = 0.84–0.96; scenario (ii)). The SUMM confirmed the viability of principles of scenario (iii) when

Declaration

The authors declare no competing interests. All authors have contributed to the paper and given approval to the final version of the manuscript.

Acknowledgement

This work was financially supported by project FCH-S-19-5971 of the Ministry of Education, Youth and Sports of the Czech Republic. JŠ thanks to the Deutsche Bundesstiftung Umwelt, Germany, Project #AZ 30017/718, for the financial support of her internship in Dresden.

References (44)

  • B. Peikert et al.

    Characterization of topsoils subjected to poorly controlled olive oil mill wastewater pollution in West Bank and Israel

    Agric. Ecosyst. Environ.

    (2015)
  • L. Ramos et al.

    Polyethylene film incorporation into the horticultural soil of small periurban production units in Argentina

    Sci. Total Environ.

    (2015)
  • A.B. Silva et al.

    Microplastics in the environment: challenges in analytical chemistry - a review

    Anal. Chim. Acta

    (2018)
  • Z. Steinmetz et al.

    Plastic mulching in agriculture. Trading short-term agronomic benefits for long-term soil degradation?

    Sci. Total Environ.

    (2016)
  • L.C. Woodall et al.

    Using a forensic science approach to minimize environmental contamination and to identify microfibres in marine sediments

    Mar. Pollut. Bull.

    (2015)
  • J. Xu et al.

    Investigation on the thermal degradation of flexible poly(vinyl chloride) filled with ferrites as flame retardant and smoke suppressant using TGA–FTIR and TGA–MS

    Polym. Degrad. Stabil.

    (2013)
  • I.R. Comnea-Stancu et al.

    On the identification of rayon/viscose as a major fraction of microplastics in the marine environment: discrimination between natural and man-made cellulosic fibers by Fourier transform infrared spectroscopy

    Appl. Spectrosc.

    (2016)
  • J. David et al.

    Quantitative analysis of poly(ethylene terephthalate) microplastics in soil via thermogravimetry–mass spectrometry

    Anal. Chem.

    (2018)
  • Deutsches Institut für Normung

    Chemical Analysis - Decision Limit, Detection Limit and Determination Limit under Repeatability Conditions - Terms, Methods, Evaluation

    (2008)
  • K. Duis et al.

    Microplastics in the aquatic and terrestrial environment: sources (with a specific focus on personal care products), fate and effects

    Environ. Sci. Eur.

    (2016)
  • P. Filzmoser et al.

    Review of sparse methods in regression and classification with application to chemometrics

    J. Chemom.

    (2011)
  • D. Fischer et al.

    Identification of microplastics in the marine environment by Raman microspectroscopy and imaging

    Am. Lab.

    (2015)
  • Cited by (43)

    View all citing articles on Scopus
    View full text