Development of a new test design to investigate the degradation of pesticides in soil under sunlight conditions

Abstract

Pesticides applied to soil surface are subject to photodegradation if the parent molecule is sensitive to UV-light absorption. Photodegradation studies are therefore mandatory for the registration of plant protection products to provide data on the degradation rate and on the nature of photoproducts formed. In general, sunlight is simulated in these studies with xenon lamps, e.g., a Suntest® device. Surface application on very thin soil layers followed by direct irradiation is common practice, but the control of the boundary conditions, i.e. soil temperature and moisture, to maintain the structure and viability of the soil is challenging. A homogeneous and stable soil microclimate is crucial to compare the degradation data of the test item from the irradiated soil samples to the dark controls as well as to the results from the aerobic soil metabolism study. After trying different scale-up test systems with the UV-sensitive herbicide imazamox as comparative test item, a new soil photolysis test system was developed which is manageable in the laboratory and enables a more favorable management of the boundary conditions, especially with regard to the soil moisture and temperature. For this, the solar simulator SolarConstant® 1200, equipped with metal halide lamps Radium HRI-TS 1000W/D/S/PRO, was installed by Atlas Ametek (Germany) in a temperature controllable walk-in incubation chamber with aluminum racks and reflectors to minimize diffuse light and to maintain a homogenous temperature of 22(± 1)°C within the irradiated soil. Borosilicate glass vessels with an inner diameter of 10 cm and a maximum height of 9 cm, covered by quartz glass, were used for the incubation of the applied soil under light. Contrary to the imazamox degradation half-lives obtained with the Suntest® test system, where an unusual slower degradation was observed under light compared to the dark controls, the results from the new SolarConstant® study design showed the expected faster degradation under light. Hence, it can be concluded that the experimental boundary conditions of the new test system are more suitable to maintain the viablity of the irradiated soil. Since no adjustments of the soil water content were needed, compared to daily water adjustments for thin soil layers incubated under a Suntest®, drying–wetting cycles are eliminated and microbial-induced soil processes are maintained.

Introduction

The widespread application of pesticides demands numerous studies concerning their biological, chemical, and photochemical degradation in the environment. Photochemical transformation can be an important process for pesticides in the atmosphere and for those in water or on surfaces such as leaves, vegetation, and soils. In principle, chemicals present in the environment can undergo either direct phototransformation through an excited state or indirect phototransformation by reacting with another chemical in an excited state.

In this context, it is required to conduct photochemical experiments aimed at the determination of phototransformation rate constants and half-lives for the pesticides at environmentally realistic concentration levels and at representative wavelengths on the one hand and for the identification of all photoproducts on the other.

Some national guidelines [1,2,3] as well as EU regulation requirements [4] and SETAC procedures [5] were the basis for a draft OECD guideline proposal on the phototransformation of chemicals on soil surfaces [6], which has never been finalized and agreed upon. The SETAC procedures and the OECD draft proposal recommend the use of xenon light at a test temperature of 20 °C (± 3 °C) and of viable, air-dried soils which can form thin soil layers, i.e. silty loam, clay loam, and which have been used within the aerobic soil metabolism study according to OECD guideline 307 [7]. The OECD draft proposal recommends for the incubation under light a soil layer thickness of 2 mm. Furthermore, the air-dried soil can be stored at ambient temperature as microbial activity is of little relevance when studying physico-chemical processes on the soil surface [6].

However, to investigate the abiotic and biotic fate of pesticides under simulated sunlight conditions, it is important to manage the water content of the light exposed soils as close as possible to their natural state, in a moisture- and temperature-controlled environment [8,9,10]. In general, a viable soil should neither be too wet nor too dry to maintain adequate aeration and nutrition of soil microflora, i.e. moisture contents recommended for optimal microbial growth range from 40 to 60% maximum water holding capacity [7].

The most recent guidance on photodegradation on soil is the US-EPA guideline OPPTS 835.2410 [11]. The guideline does not specify a certain lamp to be used, but the soil samples should be exposed to a spectrum of light providing or simulating expected use conditions. If an artificial light source is used, its intensity, wavelength distribution, and the length of the exposure should be comparable to sunlight. The soil type should be equal to one of the soil types that are used in the aerobic soil metabolism study and taken from the same location. Soil temperature should be held constant (± 1 °C) between 18 °C and 30 °C.

In general, soil photolysis studies which were conducted at BASF SE for the registration of pesticides have been accepted worldwide by the authorities, without any objections. This implies that for all registration studies which were conducted in the past the same soil as used in the aerobic soil metabolism was applied, adjusted to 40–60% max. water holding capacity and with a soil layer thickness of 1 cm. A droplet application was performed onto the soil surface and samples were incubated for 15 days under continuous irradiation using the Suntest® xenon light, adjusted to 3 mW cm⁻² in the 290–400 nm range, comparable to the light conditions at a summer midday at BASF SE, Limburgerhof, Germany.

However, from time to time a significant slower degradation of a UV-sensitive test substance in the Suntest® set-up was observed compared to the dark controls. Semi-sterile conditions on the top of the thin soil layer under the xenon lamp could be the reason but should then be a general issue in all studies which is not the case. Erratic external stress factors coming from the boundary conditions of test system, especially regarding to soil moisture, seem to be more obvious. Soil drying and the necessary rewetting to the initial soil moisture is a general problem in Suntest®-studies with thin soil layers that may affect microbial activity, soil–water interactions and soil microstructural stability [12].

Hence, an approved test design was developed in several steps to improve the control on the boundary conditions within the experimental set-up.

Materials and methods

For the experiments given in the following, the herbicide imazamox ((RS)-2-(4-isopropyl-4-methyl-5-oxo-2-imidazolin-2-yl)-5-methoxymethylnicotinic acid; CAS No. 114311–32-9) was applied as ¹⁴C-radiolabelled test item (chemical purity > 98%, radiochemical purity > 95%), either labeled with the ¹⁴C-label placed in the imidazolinone ring at the (5-C) carbonyl carbon [imidazolinon-5-¹⁴C] or at the 3-C of the pyridine ring [pyridine-3-¹⁴C]. Imazamox belongs to the imidazolinone class of herbicides and is used for the control of certain annual grass and broad leafed weeds in different crops. Imazamox is susceptible to photochemical degradation, e.g., in water (λ_max 191 nm: ε 27791 l mol⁻¹ cm⁻¹; λ_max 267 nm: ε 6152 l mol⁻¹ cm⁻¹; λ_max 290 nm: ε 3057 l mol⁻¹ cm⁻¹) without any pH-dependency of the UV adsorption [13].

Different test designs were compared which are described in the following. First, the standard soil photolysis test design in an Atlas Suntest® with xenon light used in previous registration studies, second, an exploring scale-up experiment in a phytotron to investigate the influence of soil volume and a more realistic soil water content on the degradation of the test item, and third, a new test system, i.e. soil photolysis in an incubation chamber equipped with SolarConstant®, developed on the experiences and results of the scale-up experiments in the phytotron. An overview of the experimental parameters is given in Table 1; further information is provided in the supplementary material (Tables S1–S3).

Table 1 Experimental parameters of the photolysis studies

Soil photolysis in an Atlas Suntest® with xenon light

A Suntest® study was conducted by McCall and Blood (2013) [14] and evaluated within the EU-registration of imazamox [15], relevant study details are cited herein.

Photolysis was carried out in an Atlas Suntest® CPS Plus unit with a xenon lamp equipped with filters to mimic sunlight (wavelengths < 290 nm were filtered out). The light intensity between 300–800 nm was measured using a LI-COR model LI-1800 Spectroradiometer at the beginning and end of the study. The average light intensity was 52.9 mW cm⁻², comparable to the natural sunlight intensity in spring at the test facility site in Raleigh, North Carolina.

The Suntest® unit (Figure S1—supplement) accommodated a specially designed jacketed stainless steel tray. Smaller stainless steel trays (9 cm × 4.2 cm, L x W) containing 20 g dry weight of soil with a depth of 1 cm were placed inside the jacketed tray. After placing the soil trays inside the jacketed tray, it was covered with a quartz glass plate and secured with clamps. The temperature of the soil was maintained at 20 ± 2 °C by circulating coolant through the bottom jacket of the steel tray. The jacketed steel tray also contained three holes. Two holes were used as inlet and outlet vents for the collection of volatiles and the third hole allowed insertion of a temperature probe to monitor the temperature of the soil during the study. The temperature probe was inserted into a control sample which contained untreated soil. The soil temperature was recorded each day during the photolysis experiments. Premoistened air was supplied to the irradiated samples, air exiting the tray was drawn through 0.5 N sodium hydroxide to capture volatile radioactive products that were produced during photolysis.

The dark control soil samples were also weighed into stainless steel trays and treated with the radioactive test substance. The treated samples were placed in a jacketed metal tray identical to the one used for the treated samples, sealed and placed in an incubator that was then kept in the dark at 20 ± 2 °C.

The soil moisture of the irradiated and dark control samples was adjusted to approximately 50% of the maximum water holding capacity prior to application of the test material to the soil and maintained throughout the experimental period by pipetting deionized water to the soil surface on daily basis, if necessary.

[Imidazolinon-5-¹⁴C]-imazamox was applied to the soil at an application rate equivalent to a field application rate of 50 g ai/ha; [pyridine-3-¹⁴C]-imazamox was applied at a lower application rate, i.e. 3 g/ha.

The total soil amount of the irradiated and the dark control samples were sampled concurrently at 0, 1, 3, 7, 11 and 15 days after treatment. Volatile residues were collected and analyzed at the soil sampling intervals. Subsets of each sample were extracted, the extracts were pooled, concentrated on a rotary evaporator and subjected to reverse phase HPLC analysis with radiodetection (Table S2—supplement).

Soil photolysis in scale-up experiments in a phytotron

Scale-up experiments were conducted to achieve more realistic scenarios regarding temperature and soil water content contrary to the thin soil layers in a Suntest® apparatus. Two systems were used, i.e. square Euro container (Auer GmbH, Germany) filled with 20 kg soil and circular Duran glasses filled with 3.5 kg soil, adjusted to 40% and 50% maximum water holding capacity by thoroughly mixing the soil with the needed amount of water using a concrete mixer. The test systems were placed on weighing cells (Figure S2—supplement) and incubated in a temperature-controlled phytotron (Weiss Technik GmbH, Germany), simulating sunlight conditions with metal halide lamps. Since the handling of the soil water management was the main focus of this study, the boxes were not covered by quartz glass plates and a material balance was not established. Furthermore, no dark controls were set up.

[Pyridine-3-¹⁴C]-imazamox was evenly distributed onto the soil surface by droplet application, with an application rate of 75 g/ha. The weight of each container was continuously monitored and lost amount of water was immediately adjusted to the initial total weight by adding water to the soil surface. Sampling was conducted 7 and 15 days after treatment. The total amount of soil from each Euro box was homogenized in a portable concrete mixer and divided into 3 kg subsets. The total amount of soil from each Duran glasses was homogenized in stainless steel bowl using a handheld electric mixer. For further analysis soil aliquots were extracted with organic solvents and analyzed by radio-HPLC (Table S2—supplement).

Soil photolysis in an incubation chamber equipped with SolarConstant® for solar simulation

Since the open scale-up test systems tested in the phytotron are not suitable for a routine laboratory test system that has to fulfill guideline requirements, the next step was to downscale the size and establish a smaller system which can be handled in the laboratory, enables a mass balance and—most importantly—ensures realistic and stable boundary conditions, i.e. with regard to light conditions and soil water and temperature management.

For this purpose, smaller test vessels were designed and placed in a temperature controllable walk-in incubation chamber (Conviron, UK) (Figure S3—supplement) and irradiated with a SolarConstant® 1200 solar simulator (Atlas Ametek, Germany; https://www.atlas-mts.com/products/custom-solar-simulation/mhg-solar-simulation) (Figure S4—supplement) at a light intensity corresponding to the UV-A radiation of a clear summer day in Limburgerhof, Germany.

Irradiation was performed continuously by a HRI-TS Metal halide lamp, installed about 90 cm above the test vessels. Filtered metal halide global lighting (MHG) is a mature technology used since the 1980s by Atlas and others for larger solar simulators inside environmental test chambers. MHG lamps are filled with a mixture of mercury, halides and rare earth elements which vaporize during ignition, generating a plasma. After a few minutes, the plasma reaches thermal equilibrium and emits a spectrum similar to global solar radiation when properly filtered. To achieve the required uniform irradiance on the test object, special reflector designs are needed and the number of required SolarConstants® modules as well as their ideal positioning inside the chamber have to be defined. MHG technology exhibits two advantages over xenon light. First, MHG are high efficiency sources and provide a higher output for electrical input and are better suited to large area illumination. Second, MHG emits lower amounts of infrared light as compared to xenon, making thermal management easier. As with xenon, optical filters fine-tune the spectral irradiance to provide glass-filtered indoor solar radiation. Two lamps each are installed in small cabins within the walk-in chamber. Each cabin, in which all test vessels for a study are placed, has a table surface of 120 × 70 cm (width x depth), the lamp housings including the reflectors (30 cm × 28 cm × 28 cm; width x height x depth) are installed in a height of 80 cm to the test vessels, the distance between the lamps is 60 cm to ensure the required light intensity for the test systems of 3 mW cm⁻² (UV-A). The light intensity was controlled weekdays with a radiometer (Opsytec Dr. Gröbel, Germany). The spectrum of the applied lamp is given as supplementary information (Figure S5 and S6).

The test vessels are made of borosilicate glass with an inner diameter of 10 cm and a total height of 9 cm and were filled with about 450 g moist soil (equivalent to 400 g dry soil) to a height of 4 cm and a surface area of about 78 cm² (Fig. 1). Optional, to reduce the volume of soil extraction solvent, the amount of soil can be reduced to a minimum amount of 200 g with no effect on the soil water management when the height of the vessels is reduced, but the surface area is kept. The day 0 samples were directly prepared in centrifuge tubes.

Fig. 1

New test design: borosilicate glass vessel with an inner diameter of 10 cm and a max. height of 9 cm, covered by a quartz glass cover. A thin heating wire is installed under the quartz glass cover to prevent water condensation

The soil moisture was adjusted to pF 2.0–2.5, corresponding to 50% of the maximum water holding capacity, by adding the required amount of water and mixing with a hand-mixer. The soil was pressed very gently to create a uniform surface before treatment. The test vessels were weighed at each sampling to readjust eventual losses of water and to maintain the initial water content as constant as possible, however, no adjustments were needed during the study period (Table 2).

Table 2 Water adjustment in % of soil wet weight (% w/w), range of all test vessels

The amount of imazamox to be applied on the soil surface was calculated based on a field application rate of 75 g/ha and a soil surface area of 78.54 cm². The application was performed with a Gilson Microman pipette dropwise and uniformly onto the soil surface. Samples of day 0 were weighed in the centrifuge cups and the application solution was directly added.

The temperature inside the test vessels was measured with a probe placed inside an additional test vessel filled with untreated soil. The chamber temperature was set to 15—17 °C to maintain the temperature of 22 ± 1 °C in the soil. The test vessels for the dark controls were identical and placed in an incubator at 22 ± 1 °C.

Each vessel was closed airtight by a quartz glass cover. A thin heating wire was installed under the quartz glass cover to prevent water condensation. Each vessel was continuously aerated with moistened, CO₂-depleted air via an air inlet. In order to trap potentially evolving volatiles, the emergent air from the air outlet was passed through three different trapping solutions located between test vessel and pump, i.e. ethylene glycol, 0.5 M sulfuric acid and 0.5 M sodium hydroxide (Fig. 2).

Fig. 2

Each vessel has two adapters flanged in opposite direction for incoming air and as output for volatile degradation products, which were trapped in appropriate absorbers

The sampling dates were 0, 1, 3, 7, 10 and 15 days after treatment (DAT). Two test vessels were taken at every sampling time from each photolysis test system and the dark control.

At each sampling date the entire amount of the soil sample was filled into a centrifuge tube and consecutively extracted three times with 400 mL methanol and three times with 400 mL methanol/water (1:1, v/v) (Table S2—supplement). After each extraction step, soil and extract were separated by centrifugation and the extracts were analyzed by liquid scintillation counting on the total amount of radioactive material. The combined soil extracts of each sample were concentrated and analyzed by radio-HPLC. At each sampling time (except for day 0) the volatile trapping solutions were transferred into volumetric flasks, the volume was adjusted to 50 mL with distilled water and aliquots were analyzed by liquid scintillation counting. In order to determine the amount of non-extractable residues (NER) after solvent extraction, five aliquots of each soil sample (about 0.7—1.2 g each) were combusted in an analytical oxidizer Robox 120C (Zinsser, Germany). The evolved ¹⁴CO₂ from each combusted aliquot was trapped in a scintillator solution and measured by liquid scintillation counting.

Kinetic evaluation

Kinetic evaluations of the Suntest® and SolarConstant® experiments were performed considering the methods and criteria described in the FOCUS Guidance Document [16]. The kinetic models single first-order (SFO), double first-order in parallel (DFOP) and first-order multi-compartment (FOMC) were tested in order to identify a description of the kinetics. The appropriateness of a distinct kinetic model to describe the degradation was tested with the following checks recommended by FOCUS:

• Visual assessment of goodness-of-fit.

• Estimation of the error percentage at which the χ2 test (Chi-square test) is passed.

• T-test to evaluate whether estimated degradation parameters differ from zero.

The visual fit was categorized as follows:

• Poor fit = the fit does not follow the pattern of the measured residues, not acceptable to derive degradation endpoints.

• Acceptable fit = the fit mainly follows the pattern of the measured residues with small deviations, acceptable to derive degradation endpoints on a case-by-case basis.

• Good fit = the fit follows the pattern of the measured residues well, residuals are randomly scattered around zero, acceptable to derive degradation endpoints.

A kinetic model is considered appropriate if the visual fit is good or acceptable, the χ2 error value is ideally < 15%. The estimated degradation parameters should be significantly different from zero, i.e. the t-test for the degradation parameters is passed at 5% error level. The software package KinGUI (version 2.2014.224.1704) was used for parameter fitting [17, 18].

Results and discussion

Soil photolysis in an Atlas Suntest® with xenon light

The results of a Suntest® study [14] that was evaluated within the EU-registration of imazamox [15] are cited herein.

The imidazolinon-5-¹⁴C-labeled study showed that the parent decreased from 91.5% total applied radioactivity (TAR) at 0 DAT to 60.5% TAR at 15 DAT in the irradiated samples. In the dark controls a similar decreasing trend was observed in going from 91.5% TAR at 0 DAT to 51.7%TAR at 15 DAT. The pyridine-3-¹⁴C-label similarly saw a moderate decrease in parent from 93.2% TAR at 0 DAT to 73.0% TAR at 15 DAT. In the dark control samples, parent decreased from 93.2% TAR to 64.2% TAR at 15 DAT. Results are given in the supplementary material (Table S5). One degradation product was observed and identified as the diacid derivative of the parent molecule, i.e. 2-(4-isopropyl-4-methyl-5-oxo-2-imidazolin-2-yl)pyridine-3,5-dicarboxylic acid in both irradiated and dark control samples. The levels of this product steadily increased throughout the study. The degradation patterns observed between the two labels were similar except that a greater amount of radiolabeled CO₂ was generated from the imidazolinon-5-¹⁴C-labeled compound.

Soil photolysis in scale-up experiments in a phytotron

In contrast to the results of the Suntest®-study, the scale-up experiments to investigate and improve the soil water management showed a faster degradation of imazamox under simulated sunlight, with values of 46% TAR (Euro box with 40% max. WHC) resp. 40% TAR (Euro box with 50% max. WHC) and 42% TAR (Duran with 40% max. WHC) resp. 34% TAR (Duran with 50% max. WHC) after 15 days of continuous irradiation (Table S3—supplement), less than the values obtained in the Suntest®-study (Table S4—supplement). The maintenance of the initial soil moisture was much easier to handle in the scale-up experiments and considerably less amounts of water were needed for rewetting (Table 2) due to the higher amount of soil and the better soil water distribution than in the thin layer of the Suntest®-study. Less moisture dynamics in the form of drying and rewetting support the organization of soil pore space and the microstructural stability of the soil [12]. Hence, the soil boundary conditions for testing pesticides in the laboratory should be as realistic as possible to ensure a stable and viable soil microclimate.

Soil photolysis in an incubation chamber equipped with SolarConstant® for solar simulation

The mean material balance values for the photolysis and dark control samples ranged from 94.8% to 102.6% TAR (Table S5 and S6—supplement). Non-extractable residues reached a maximum of 54.8% TAR in the photolysis test and of 30.0% TAR in the dark control after 15 days. Mineralization to carbon dioxide was higher under irradiated conditions, with 6.6% TAR in the photolysis test and 0.1% TAR in the dark control after 15 days.

The combined methanol and methanol/water (1/1, v/v) extracts of each sampling point were analyzed by radio-HPLC. Imazamox degraded from 89.7% TAR at day 0 to 13.3% TAR in the photolysis test and to 55.4% TAR in the dark control after 15 days (Table S7—supplement).

No addition of water was needed to maintain the humidity of the soil (Table 2), i.e. the soil moisture dynamics were unchanged in the course of the incubation and the basic soil properties and microbial-induced soil processes remained stable.

Kinetic comparison of imazamox degradation in the Suntest® and SolarConstant® test systems

Kinetic fits were derived for both test systems, light conditions (irradiated resp. dark control) and radiolabels. In general, degradation half-lives of imazamox in non-irradiated soils can vary from 2 days to more than 120 days, depending on the soil pH, with faster degradation in neutral soils than in acidic soils [19].

The visual evaluation of residuals (Figure S7—supplement) showed an acceptable or good fit for all experiments using a SFO kinetic model. The χ2 tests resulted in errors less than or equal to 5% and t-tests for the model parameters were significant (Table 3). Therefore, SFO kinetic model is appropriate for all experiments. Other kinetic models (FOMC, DFOP) were tested as well, but resulted in a lower performance and are not presented.

Table 3 Statistic summary of SFO kinetic fits of observed residues in the soil extract of test trials with Suntest® and SolarConstant® devices under light and dark conditions

Degradation varied between devices and conditions as illustrated in Fig. 3. The Suntest® shows an unusual faster degradation under dark conditions (DT₅₀ = 15.9 days and 18.6 days) compared to light conditions (DT₅₀ = 23.1 and 40.7 days). This observation was already discussed within the most recent EU-evaluation of imazamox and it was deduced that photolysis in soil is not expected to be significant [15].

Fig. 3

Comparison of SFO kinetic fits of observed residues in the soil extract of test trials with Suntest® and SolarConstant® devices under light and dark conditions

However, imazamox degraded much faster under irradiated conditions in the SolarConstant® test system (DT₅₀ = 5.4 days) compared to the dark controls (DT₅₀ = 18.6 days), as expected for a UV-sensitive compound. Hence, it can be concluded that photolysis enhances the degradation of imazamox in soil if the experimental settings ensure stable soil moisture conditions during UV-light exposure.

Conclusion

The installation of the soil photolysis test systems in a temperature-controlled incubation chamber enables a more favorable management of the boundary conditions, especially with regard to the soil water and temperature. Compared to the thin soil layers used in the Suntest®, soil drying and rewetting is minimized and, consequently, artificial stress factors on soil structure and microbial activity caused by drying-wetting cycles [12] are eliminated.

This is even more important considering EFSA’s new guidance on soil phototransformation products in groundwater [20], which have to be considered in the lower tier approach of the groundwater exposure assessment. Currently, laboratory soil photolysis studies are designed as route studies in order to get information about the transformation process and products. Since there is currently no harmonized finalized OECD guideline, laboratory soil photolysis studies are considered not to be suitable in order to derive kinetic parameters for modeling purposes, because the test conditions are rather artificial. If a new or revised OECD guideline is established, this statement might be revised, i.e. soil photolysis studies might be more suitable for a kinetic evaluation than data from field dissipation studies with changing irradiances. The herein presented new test design constitutes an important input to this discussion.