Dieter
Hennecke
*a,
Mike
Kruse
a,
Joana
Bräutigam
a,
Boris
Meisterjahn
a,
Judith
Klein
a,
Daniela
Claßen
be,
Stefan
Trapp
c,
Matthias
Kästner
d,
Andreas Libonati
Brock
c and
Andreas
Schäffer
*e
aFraunhofer Institute for Molecular Biology and Applied Ecology, 57392 Schmallenberg, Germany. E-mail: dieter.hennecke@ime.fraunhofer.de
bGerman Environment Agency, Section Chemicals, 06844 Dessau-Roßlau, Germany
cTechnical University of Denmark, Department of Environmental Engineering, 2800 Kgs. Lyngby, Denmark
dUFZ, Helmholtz Centre for Environmental Research, Department Environmental Biotechnology, 04318 Leipzig, Germany
eInstitute for Environmental Research, RWTH Aachen University, 52074 Aachen, Germany. E-mail: andreas.schaeffer@bio5.rwth-aachen.de
First published on 20th January 2023
Most, if not all, chemicals, biocides, pharmaceuticals and pesticides are known to produce non-extractable residues (NER) in solid environmental media like soil and sediment during degradation testing to various extents. Since it has been found that parent substances and relevant metabolites can be contained and potentially released from NER there is currently much debate on how to include NER in the environmental persistence assessment. Using radioactive or stable isotope labelled test substances, three types of NER can be experimentally discriminated: entrapped, sorbed or heavily sorbed (type I) having the potential to be released from the matrix. Type II NER, i.e. residues covalently bound to organic matter in soils or sediments, are being considered to have very low remobilisation potential. Type III NER (bioNER) are formed after degradation of the xenobiotic chemical and incorporation into natural biomolecules (anabolism) like amino acids and other biomass compounds, and are, thus, of no environmental concern. Silylation has been suggested as a methodology to differentiate types I and II NER but concern has been addressed that this procedure is not suitable for routine analysis, e.g. in the context of studies for authorisation and registration of chemicals. Here, we describe a readily applicable and reproducible experimental procedure to apply this method for the analysis of NER derived from bromoxynil, sulfadiazine and isoproturon, respectively. This method is able to distinguish between heavily sorbed and covalently bound residues of chemicals, biocides, pharmaceuticals and pesticides in soils and to subsequently identify residues of the parent substance entrapped in type I NER.
Environmental significanceThe manuscript describes a procedure to distinguish the fraction of bound residues without environmental concern from the fraction with a potential concern. It enables chemical regulation to identify chemicals generating bound residues with significant fractions of environmental concern, which was not possible so far as appropriate methods are lacking. In order to set appropriate limits for use and release of those chemicals to the environment, regulation needs stand-ard procedures to determine reliable data in laboratory routine testing. The publication of applicable and reproducible laboratory methods for determining relevant parameters is a first step towards standardizing these methods. |
According to the IUPAC definition,2 NER in soils are defined as species originating from chemicals, that remain unextracted by methods which do not significantly change the chemical nature of these residues or the structure of the respective matrix.3 Although guidance documents4–7 describe some principles on the assessment of NER, clear definitions for the extraction procedure of the environmental matrix to obtain NER and, especially, of techniques to characterize NER and NER-subtypes are missing or vague.
Regulatory views on NER formation differ considerably, either defining them as degraded residues of no environmental concern as in the case of pesticides,8,9 or as potentially bioavailable and non-degraded residues (“parent substance”) in the regulation of general industrial chemicals4,6,10 if it has not been demonstrated that the chemical is completely degraded or irreversibly bound. In this respect the terminology of NER is used as either safe sink or as hidden hazard.11
In most investigations, NER have been only quantified by use of isotope labelled test substances. For quantification, the solid sample containing NER is combusted and the formed isotope labelled carbon dioxide can be trapped and measured. In the case of radiolabelled substances this is by liquid scintillation counting (LSC).12 In the case of stable isotope labelling this is conducted by use of compound specific high-resolution mass spectrometry (HR-MS).13 NER have been characterized using spectrometric techniques after labelling the molecule with suitable stable isotopes, e.g.13C or 15N for corresponding NMR analysis14–17 or HR-MS18–24 or 14C-labelling combined with LC-MS.25,26 However, in the case of stable isotopes this is often at elevated concentrations of the test substances and this might have led to significantly decreased degradation.27,28
Regarding the persistence assessment of chemicals under REACH, it has been suggested29 to consider unknown total NER as remobilisable parent or transformation products, if no characterization or additional information has been provided. Different subtypes have been proposed,39 type I NER defined as sorbed and physically entrapped parent substance and transformation products. Type II NER is defined as parent or metabolites covalently bound to the matrix. Type III bioNER comprising natural compounds like amino acids, nucleic acids or phospholipids formed by microbial anabolism and assimilation of the carbon from the environmental pollutant and eventually incorporated into the soil organic matter or humic fraction of the respective matrix.29 If type III/bioNER and type II NER have been identified, these fractions should be considered as 'safe sink',29 since (i) biomolecules are of no environmental concern and (ii) covalently bound NER have a very low remobilisation potential due to stable chemical bonds,30,31 unless evidence for remobilisation has been provided. Whether or not type I NER have to be considered in the persistence assessment of chemicals, is currently under debate32 and the European Chemicals Agency seems in favour of.33
A standardized and reliable technology is needed to clearly differentiate the type I and II fractions and to unroll the safe sink vs. hidden hazard presumption of NER in environmental risk assessment. Silylation of soils and sediments is a proposed method to differentiate types I and II.29 Alternatively, an EDTA based extraction method has been proposed.41 These authors admitted that silylation may provide a clear differentiation of entrapped and covalently bound residues. Since there have been strong reservations about the silylation method in routine laboratory work, the methods have only been compared occasionally. However, as shown here, silylation is a fit for purpose method for characterisation of NER.
Silylation with trimethylchlorosilane (TMCS) leads to the exchange of protons in soil/sediment functional groups like –OH, –COOH, –NH2 with trimethylsilyl residues, thereby hindering the formation of hydrogen bridges. This results in the fragmentation of the humic matter associates29 and release of entrapped, heavily sorbed residues (type I NER). If feasible, the silylation extract containing the released residues can be analyzed for the presence of the parent molecule and degradation products. Examples for both extremes have been published recently (see discussion). However, in the framework of regulatory degradation studies concern about the proposed methodology has been addressed that this procedure is not easily manageable for routine analysis. In the present paper we describe a standardized, readily applicable, and reproducible procedure to employ silylation for analysis of NER in soil fate and turnover studies, in order to distinguish type I and type II NER derived from chemicals, biocides, pharmaceuticals and pesticides. It should serve as basis for future studies to compare and evaluate the results of silylation and EDTA-extraction to decide on the best procedure for type I NER determination.
Data shown were obtained within the scope of a project funded by the German Environment Agency (UBA), which was set up in order to investigate the formation and characterization of NER of the herbicides bromoxynil and isoproturon and the antibiotic sulfadiazine by use of the silylation procedure.
During reaction, the samples were stirred at room temperature at 100–200 rpm to maintain a homogeneous suspension. After three hours a further 10 mL of TMCS and 1.5 g NaOH were added to each sample. For addition of the reagents, plugs were opened only for a short time. Slight pressing of the gas bags helped to maintain the inert atmosphere in the reaction flasks during the addition. Then, the plugs were secured again with a clamp and the samples were further stirred overnight at room temperature. The gas bags were emptied directly after the reaction was finished to prevent corrosion. At the slightest suspicion of a leak, the bags were replaced with new ones. In total 42 samples (5 biotic and 2 sterile samples in duplicate per substance) were silylated in duplicate which corresponds to 84 individual silylation results.
All extracts were initially analysed by liquid scintillation (LSC) counting for the total extracted radioactivity (Hidex 600 SL, Hidex, Turku, Finland). For further work-up the silylation extract was transferred into a 50 mL Sarstedt vial and concentrated to about 15 mL of extract volume by a gentle stream of nitrogen. The additional extracts of the silylation residues in the cases of sulfadiazine and isoproturon (see above) were added to the corresponding silylation extracts and concentrated again by a gentle stream of nitrogen to around 15 mL final of extract volume. The resulting extract was filled up with methanol to 20 mL total volume and an aliquot was analysed by LSC. The recovery of radioactivity during concentration was mostly between 90% and 100%. The bromoxynil silylation extract could be directly analyzed by radio-TLC without any further treatment, as no extraction of the silylation residue was needed because bromoxynil is soluble enough in chloroform. The radio-TLC application volumes were adjusted in a way that the spots contained about 2.5 Bq to 10 Bq of extracted radioactivity, respectively. Exposure time was 7 d. The TLC conditions used are stated in ESI Table S4.† All sample runs on radio-TLC were converted into chromatograms and integrated using Aida Version 3.44 software. An example is shown in Fig. 4. For quality control purpose in this study we analysed the silylation residue for remaining radioactivity by combustion analysis in order to establish a mass balance for the silylation procedure.
In addition, soil degradation experiments with 13C-labelled sulfadiazine, bromoxynil and isoproturon have been performed. The chemical analysis from 13C-silylation extracts was performed by substance-specific LC-MS analysis for extracted parent. The data were compared with the parent substance recovered in the 14C- experiments. Details on LC-MS analysis are stated in ESI (Tables S5 and Table S6†).
substance | Sampling day | In % of the applied radioactivity | |||
---|---|---|---|---|---|
Extractable | Mineralised | NER | Mass balance | ||
a One replicate not entirely sterile, data from sterile replicate only. | |||||
Bromoxynil | 0 | 108.3 | — | 3.4 | 111.8 |
1 | 95.3 | 0.3 | 8.6 | 104.3 | |
2 | 87.7 | 0.7 | 14.1 | 102.5 | |
7 | 68.0 | 3.9 | 30.1 | 101.9 | |
10 | 57.7 | 6.0 | 41.2 | 104.9 | |
14 | 45.0 | 8.5 | 49.4 | 102.9 | |
27 | 17.2 | 16.1 | 71.1 | 104.4 | |
62 | 7.5 | 22.0 | 70.3 | 99.8 | |
120 | 4.9 | 28.8 | 65.5 | 99.2 | |
Bromoxynil sterile samples | 14 | 102.1 | — | 7.8 | 109.9 |
119a | 99.6 | — | 6.3 | 108.9 | |
Sulfadiazine | 0 | 95.9 | — | 4.1 | 100.0 |
1 | 85.0 | 0.0 | 16.8 | 101.8 | |
2 | 73.3 | 0.1 | 25.5 | 98.9 | |
3 | 66.3 | 0.2 | 36.5 | 103.0 | |
7 | 53.1 | 0.3 | 43.4 | 96.8 | |
10 | 44.9 | 0.5 | 56.9 | 102.3 | |
14 | 33.0 | 0.6 | 68.6 | 102.2 | |
28 | 20.7 | 0.9 | 72.5 | 94.2 | |
58 | 12.7 | 1.0 | 92.5 | 106.1 | |
120 | 6.0 | 1.7 | 82.8 | 90.5 | |
Sulfadiazine sterile samples | 14 | 59.2 | — | 44.4 | 103.5 |
120 | 18.0 | — | 81.5 | 99.5 | |
Isoproturon | 0 | 100.3 | — | 1.2 | 101.5 |
1 | 90.5 | 0.2 | 1.9 | 92.6 | |
2 | 94.3 | 0.3 | 2.8 | 97.3 | |
3 | 91.1 | 0.4 | 4.0 | 95.4 | |
7 | 89.9 | 0.7 | 4.8 | 95.4 | |
11 | 86.0 | 1.0 | 6.2 | 93.2 | |
14 | 83.2 | 1.6 | 8.7 | 93.5 | |
29 | 71.9 | 3.5 | 17.4 | 92.7 | |
58 | 55.0 | 9.2 | 32.5 | 96.7 | |
98 | 30.2 | 18.1 | 41.7 | 90.0 | |
120 | 21.2 | 17.0 | 54.3 | 92.5 | |
Isoproturon sterile samples | 15 | 97.7 | — | 2.8 | 100.4 |
120 | 92.5 | — | 3.2 | 95.7 |
To analyse the reproducibility of the method, the recoveries of radioactivity of both replicate samples are compared using the Mann–Whitney U test (also known as Wilcoxon rank-sum test). Calculation was performed using the R function wilcox.test.35 There is no significant difference between both replicate samples (p-value 0.78 > 0.05). Thus, the two distributions are stochastically equivalent and reproducibility of the method is confirmed. In addition the variation of both replicate samples was similar (Levene test, F(1, 82) = 0.014, p = 0.91 > 0.05).36 The data used for this analysis and results of the statistical evaluation are shown in Tables S7 and S8 in the ESI.†
For all three test substances stability tests in spiked soil were performed and high recoveries of radioactivity were proven, i.e., for bromoxynil between 93% and 95% and for isoproturon 103%. The silylation extract from sulfadiazine spiked soil in the stability test contained only 13.1% of the spiked radioactivity, but in the methanol extract up to 83% were additionally recovered so that the mass balance could be accepted with a total recovery of 92–96%. The chemical analysis of the extracts by radio-TLC showed different results (Fig. S4–S6†). While bromoxynil was stable against silylation (100% parent in the recovered radioactivity, Fig. S4†), isoproturon showed some losses (82.5% of the recovered radioactivity identified as parent, Fig. S6†) and sulfadiazine degraded significantly with recoveries of 34.1% and 40.8% parent in the recovered radioactivity (Fig. S5†).
For the NER-containing extracted soil samples from the OECD 307 soil degradation test, significant amounts of radioactivity were released from all samples by silylation (Fig. 2).
Fig. 2 Release of radioactivity and of parent equivalents by silylation of thoroughly extracted soil, normalized to the respective total amounts of NER after 120 days of incubation. Absolute NER amounts see Table 1. |
As shown in Fig. 2, between 17% and 25% of the total unextractable radioactivity were released by silylation from the soil matrix. The remobilized fraction may include also biogenic NER and other degradation products besides the parent compounds. Chemical analysis by radio-TLC proved that only minor amounts of 1% to 3% of the total NER contained the parent test substances, which was also confirmed by LC-MS analyses (Table S6†). As an example, the TLC analysis of silylation extracts of soil incubated with bromoxynil is shown in Fig. 3 and 4.
Fig. 3 Radio-TLC analysis of silylation extracts from 7 d and 14 d replicates of bromoxynil incubated soil. The track in the middle (“standard”) represents the analytical bromoxynil standard. |
Fig. 4 Radioanalytical thin-layer chromatogram of the silylation extract of soil incubated with bromoxynil for 7 days. The chromatogram resembles the left track of the TLC plate picture shown in Fig. 3 (track “7 d 1–1”). Peak number 4 represents the parent compound bromoxynil. Percentages shown are the respective percentages of the individual peaks compared to the total peak area. |
The LC-MS analysis of the 13C-silylation extracts were very comparable to the results of radio-TLC from the 14C-silylation extracts. In no case significant amounts of parent test substance were detected in the silylation extracts (for details see Table S6 in the ESI†). However, the sulfadiazine data may not be as reliable due to the observed sensitivity of sulfadiazine against the silylation procedure.
Chemical analysis of the silylation extract is essential for a realistic assessment of the hazard of the released parent substance from NER over time that may occur under natural conditions. For two test substances the environmental risk from NER release is considered to be low. Sulfadiazine results are not reliable due to the instability of the substance against the silylation procedure. In such cases, an alternative method for NER characterization like EDTA extraction41 can be used, but unfortunately, sulfadiazine proved to be not stable under such conditions either (data not shown). However, for other compounds the released parent amounts can be significant, for instance in the case of bisphenol S. Bisphenol S (a substitute of bisphenol A) forms high amounts of NER (45% of the applied amount) in soil; about half of this amount has been shown to comprise type I NER and about one third type II NER. Chemical analysis of the silylation extract representing type I NER showed the dominant presence of the parent substance.37 Regarding the persistence assessment of bisphenol S, the NER fraction does considerably increase its degradation half-life, if the parent substance released by silylation has to be added to the amounts released by extraction with organic solvents in the context of degradation experiments. On the other hand, pendimethalin, forming about one third of the applied amount as NER, was shown to predominantly belong to the type II fraction with only low potential of remobilization under natural conditions and thus being of little environmental concern.38
Radio-TLC has proven to be a suitable analytical method for quantification of 14C-labelled parent test substances. The LC-MS analyses of the 13C-samples proved that silylation extracts are also appropriate for LC-MS analysis. No matrix effects of the silylation extract matrix were observed in LC-MS analyses. This is important if silylation is considered to become part of the routine characterisation of NER. Though NER are determined with isotope-labelled test substances only, LC-MS can serve as confirmatory analytical method.
The final aim of NER research was to provide a unified approach for NER characterisation and quantification to be incorporated in the persistence and environmental hazard assessment guidelines for REACH chemicals and biocides, human and veterinary pharmaceuticals, and pesticides, irrespective of the different regulatory frameworks. For this purpose, we here present a reliable methodology to quantify type I and type II NER and to identify remobilisable residues.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2va00314g |
This journal is © The Royal Society of Chemistry 2023 |