Adsorption and degradation of neonicotinoid insecticides in agricultural soils

The adsorption and degradation of seven commercially available neonicotinoid insecticides in four types of agricultural soils from three states (Mississippi, Arkansas, and Tennessee) in the USA were studied. The adsorptions of all the neonicotinoids fit a linear isotherm. The adsorption distribution coefficients (Kd) were found to be below 2.0 L/kg for all the neonicotinoids in all the soils from Mississippi and Arkansas. Only in the Tennessee soil samples, the Kd ranged from 0.96 to 4.21 L/kg. These low values indicate a low affinity and high mobility of these insecticides in the soils. The soil organic carbon–water partitioning coefficient Koc ranged from 349 to 2569 L/kg. These Kd values showed strong positive correlations with organic carbon content of the soils. The calculated Gibbs energy change (ΔG) of these insecticides in all the soils ranged from − 14.6 to − 19.5 kJ/mol, indicating that physical process was dominant in the adsorptions. The degradations of all these neonicotinoids in the soils followed a first-order kinetics with half-lives ranging from 33 to 305 days. The order of the insecticides with decreasing degradation rate is as follows: clothianidin > thiamethoxam > imidacloprid > acetamiprid > dinotefuran > thiacloprid > nitenpyram. The moisture content, clay content, and cation exchange capacity showed positive effects on the degradation rate of all the neonicotinoids. The Groundwater Ubiquity Score (GUS) calculated from the adsorption distribution coefficient, organic content, and half-life indicates that, except for thiacloprid, all the neonicotinoids in all the soils are possible leachers, having potentials to permeate into and through groundwater zones.


Introduction
The neonicotinoid insecticides play a vital role in the control of different types of insects in the process of crop production and management. The use of neonicotinoids has been increasing in the last two decades since the first neonicotinoid imidacloprid was commercialized in 1991 (Mörtl et al. 2016). Neonicotinoids are registered for use on over 290 crops in more than 120 countries (Jeschke et al. 2011;Main et al. 2015). In developed countries, neonicotinoids are mainly used as seed dressing agents for many crops such as canola, sunflower, grains, beets, and potatoes (Tomizawa and Casida 2005;Goulson 2013). The amounts of neonicotinoids applied in the United Kingdom, one of few countries from which detailed records are available, rose from 3 tons in 1994 to 120 tons in 2016 (Goulson 2013;FERA 2017). Given the massive scale of use of neonicotinoids in both rural and urban areas, their impacts on environmental health and other non-target organisms have become a global concern (Main et al. 2015;Aseperi et al. 2020).
There are seven neonicotinoids commercially available on today's insecticide market: imidacloprid, thiamethoxam, clothianidin, dinotefuran, acetamiprid, thiacloprid, and nitenpyram (Li et al. 2018). It is estimated that, once applied through soil or seed dressing, only about 2-20% of the neonicotinoids are taken up by the crop and the rest will typically enter the soil (Goulson 2013;Sur and Stork 2003;Wood and Goulson 2017). The possible accumulation of the neonicotinoids in the environment has been found and reported. A study on 48 streams in the USA found that more than 50% of them had more than one neonicotinoid (Hladik and Kolpin 2015). In the Pearl River of Guangzhou, China, the insecticides of acetamiprid, thiamethoxam, imidacloprid, and clothianidin were detected in all 14 sampling sites with a concentration ranged 93-321 ng/L (Yi et al. 2019). In Sydney, Australia, more than 90% of 14 selected rivers were found to contain two or more neonicotinoids with concentrations of 0.06 to 4.5 μg/L (Sánchez-Bayo and Hyne 2014). Data on degradation and adsorption of the neonicotinoids in soils are critical for evaluating the fate and transport of these insecticides in soils and groundwater. Previous studies to obtain these data have been reported in many countries including Austria (Kah et al. 2018), India (Gupta et al. 2002(Gupta et al. , 2008, Africa (Dankyi et al. 2014(Dankyi et al. , 2018, China (Wu et al. 2012;Han et al. 2019), Spain (Rodríguez-Liébana et al. 2018), Canada , and the USA (Papiernik et al. 2006;Li et al. 2018). They concluded that a low soil adsorption (log K oc ) is found, which suggests a transport of these insecticides through irrigation and runoff. Meanwhile, Rodríguez-Liébana et al. (2018) and Papiernik et al. (2006) reported the adsorption of thiacloprid and imidacloprid could relate to the organic matters or pH in soils. A comprehensive study on the effects of soil properties including clay content, silt content, cation exchange capacity, organic carbon content, and pH on the adsorption of all seven insecticides is still missing.
In this study, four different agricultural soils with different physicochemical properties were collected in Mississippi, Arkansas, and Tennessee in the USA. The adsorption and degradation of all seven commercially used neonicotinoids were investigated. The effects of soil properties and conditions on adsorption and degradation of the neonicotinoids were scrutinized. The data are important for evaluating the environmental impacts of the use of these insecticides in the world.

Materials and methods
The neonicotinoids used in this study were obtained from Chem Service, Inc., and had a purity of 99.9%. The stock solutions (1000 mg/L) of each insecticides were prepared in methanol (HPLC grade). The physical-chemical properties of these neonicotinoids are summarized in Table 1.
The soils used in this study were collected from different agricultural sites in Mississippi, Arkansas, and Tennessee. Two were from Mississippi State University's Truck Crops Branch Experiment Station located in Crystal Springs (TCB) and Beaumont Horticulture Unit in Beaumont (BHU), Mississippi; one from Lon Mann Cotton Research Station of Arkansas Agricultural Experiment Station (CRS), Arkansas; and another from Tennessee State University's Agricultural Farm (TSU) in Nashville, Tennessee.
Soil samples were collected from three depths: top 30 cm, 30-60 cm, and 60-90 cm. The collected soils were air-dried in the laboratory with a room temperature 21 ± 2 °C for at least 5 days, then crushed and passed through a US standard No. 10 sieve (2 mm openings). The moisture contents of all air-dried soil samples were measured to be less than 2%. The characteristics of the selected soils at different depths were summarized in Table 2.

Adsorption tests
The batch adsorption tests of the neonicotinoids in four different soils were carried out. Aliquots of 10 mL of 0.01 M CaCl 2 solution with insecticides' concentrations from 0.05 to 10 mg/L were mixed with 2.0 g of air-dried soil of each type in 50 mL centrifuge tubes. The centrifuge tubes were agitated on a shaker at 200 rpm in the laboratory for 48 h. The samples were then centrifuged at 4500 rpm for 20 min and supernatants were taken and filtered through a 0.45-µm filter. The samples were analyzed using high performance liquid chromatography (HPLC).

Degradation test
Degradation tests were conducted in soils sampled from three different depths (0-30, 30-60, 60-90 cm) at four different sites (BHU, CRS, TCB, TSU) and at different moisture contents (air-dried, 20%, 40% w/w). For each moisture content, 30 degradation test samples were prepared for the soil sample from each depth and each site. In each test sample, 20 g soil was weighed into an aluminum tin, and mixed with 1 mL of 200 mg/L neonicotinoid stock solution so that the initial insecticide content in each soil sample was 10 mg/ kg. The sealed tins were cultured in a dark incubator at room temperature (21 ± 2 °C) for different degradation periods. Upon the completion of the degradation periods, triplicate tins were taken and the content of the insecticide in the soil in each tin was determined by extraction and analysis using HPLC as described in the "Extraction of soil samples for analysis" and the "HPLC analysis" sections.

Extraction of soil samples for analysis
For the degradation test samples, the soil in each tin was transferred to a centrifuge tube and 30 mL of water and acetonitrile (20: 80 v/v) was added to extract the neonicotinoid. This extraction solution was also used in the studies by Gupta et al. (2008), Wu et al. (2012), and Mörtl et al. (2016).
The mixture was agitated on the shaker for 30 min, then centrifuged at 4500 rpm for 20 min. Each soil sample was extracted three times following the same method to ensure a complete extraction. The extractants were then combined and transferred into a separatory funnel, and 5 mL of saturated saline and 20 mL of dichloromethane were added into the funnel to mix with the extractant. The organic fraction was separated in a round bottom flask and concentrated using a rotary evaporator (Heidolph G1, Germany). The evaporation residue was dissolved in methanol and transferred into a 2 mL amber vial for HPLC analysis.

HPLC analysis
The analysis of all the samples from the adsorption tests and the degradation tests was performed using HPLC (LC virtual Advisor, Shimadzu) equipped with the Diamonsil C 18 column (5 µm, 250 × 4.6 mm). The separation conditions in the HPLC test were as follows: flow rate 0.6 mL/min, column thermostat 25 °C, injection volume 10 µL, detection wavelength of UV absorption 254 nm, and mobile phase of 65% methanol and 35% water with 0.1% formic acid. The retention times under these separation conditions were 6.1 min for imidacloprid, 5.3 min for thiamethoxam, 8.0 min for thiacloprid, 5.6 for clothianidin, 6.7 min for acetamiprid, 4.2 min for nitenpyram, and 4.5 min for dinotefuran.

Adsorption
The linear isotherm (S = K d C) was found to well-fit the adsorption results of all the neonicotinoids in this study. This model presents the amount of neonicotinoids adsorbed by the soil (S) versus equilibrium concentration in the solution (C), and the slope of the model represents the adsorption distribution coefficients (K d ). Since the adsorptions of all the seven neonicotinoids in all the soil samples had the same isotherm pattern, only the graph for the adsorption equilibrium concentrations of imidacloprid is given (Fig. 1). The results based on the graphs for all the neonicotinoids can be found in Table 3. The coefficients of determination (R 2 ) and another soil adsorption coefficient (K oc ) are also provided in Table 3. The standard deviations for all the K d results in Table 3 are less than 3% of the mean values except clothianidin in TCB soil for which the standard deviations of K d range are 5-10% of the mean values. Table 3 summarizes the adsorption coefficients of all neonicotinoids from slopes of the fitting equation at different depth and soil sites. The K d values are below 2.0 L/kg for all the neonicotinoids in all the soils in Mississippi and Arkansas. Only in the Tennessee soil, the K d values ranged from 0.96 to 4.21 L/kg. In all the soils nitenpyram has the lowest K d and thiacloprid has the highest. The K d values in the top layers of all the soils are the highest. The K d values for nitenpyram in the top layers of soils range from 1.02 to 2.58, while those for thiacloprid 2.08 to 4.21. The K d values appear to be closely related to the OC in the soil -the higher the OC, the greater the K d value. It can be seen in Table 2 that the OC values in the soils decrease as the depth increases, so do the K d values as shown in Table 3. The highest and lowest K d values are found in the top layer soil at TSU and bottom layer at BHU respectively; the highest OC and the lowest OC are also found in these layers. Zhang et al. (2020) and Wang et al. (2020) reported the same phenomena on the adsorption of herbicide pyraclonil and pesticide exianliumi in soils. Meanwhile, the K oc value, which is obtained by normalizing the adsorption coefficient to the OC of the soil tends to be steady in TSU soil but varies significantly in BHU, CRS, and TCB soils. Papienik et al. (2006), Morrissey et al. (2015) and Mortl et al. (2016) reported the K oc values for imidacloprid, clothianidin, and thiamethoxam to be 82 to 43,000, 84 to 350, and 33 to 237 L kg −1 respectively. The K oc of clothianidin and thiamethoxam found in this study were higher than the reported values. This could be due to the low OC in the soils. Papiernik et al. (2006) and Li et al. (2018) reported that the value of K oc could vary by an order of magnitude, especially when OC is low. The relations between soil adsorption and the soil properties were analyzed using Kendall's Tau-b correlation coefficients, which are calculated using the IBM SPSS statistics 27. The detailed description of this software can be found in our previous work (Li et al. 2018). Table 4 summarizes the Kendall's Tau-b correlation coefficients calculated for K d and K oc for all neonicotinoids as they relate to the soil parameters. The results in Table 4 indicate that K d has a strong positive correlation with OC (p < 0.01), while the K oc has a strong negative correlation with OC (p < 0.01). Except for nitenpyram, all other neonicotinoids' K d values also have significant positive correlations with CEC of the soils (p < 0.05), and their K oc values have negative correlations with CEC (p < 0.05). However, neither K d nor K oc has significant correlation with clay content, silt content, or pH. Flores-Céspedes et al. (2002) and Aseperi et al. (2020) reported the adsorption of imidacloprid and also found that the K d is closely related to OC in the soils. However, Aseperi et al. (2020) reported that the adsorption of thiacloprid and thiamethoxam were not related to OC in the soils, which is opposite to the results from this study. The presence of the C-Cl bond in the insecticides represents an overall polar structure. The more polar a molecule is, the more likely it is to be close to charged surfaces, thus increasing the chance of van der Waals interactions (Spark and Swift 2002). The organic matter in the soils could have negatively charged adsorption surfaces. This could explain the strong relationship between adsorption and organic carbon content in this study.
To further investigate the mechanisms for the adsorption of the neonicotinoids in the soils, the Gibbs energy change (ΔG) were calculated using the equation ΔG = − RTlnK oc , where R (J/K·mol) is the gas constant, and T (K) is the absolute temperature. The Gibbs energy change indicates the degree of spontaneity of an adsorption process. The ΔG values can be used to describe the driving force of the adsorption. The more spontaneous adsorption process is, the higher absolute value of ΔG will be (Zhang et al. 2007). If the absolute value of ΔG is less than 40 kJ/mol, physical adsorption is dominant; if it is greater than 40 kJ/mol, chemical (irreversible) adsorption is dominant (Carter et al. 1995).
The calculated ΔG values of all the seven neonicotinoids in all four agricultural soils range from − 14.6 to − 19.5 kJ/ mol, indicating that the adsorption of all insecticides is mainly a physical process. The low ΔG values also indicates that the adsorption between the insecticides and soils is dominated by van der Waals force. Thus, the adsorption is relatively weak and reversible. This weak adsorption also indicates a high mobility of the insecticides in the soils.
The adsorption distribution coefficients (K d ) are relatively low for all the insecticides in all soils, indicating a low affinity of these insecticides in these soils. The calculated Gibbs energy change also confirms that the adsorption process is weak and reversible. This result suggests that the insecticides will be easily transported through irrigation or runoff. Han et al. (2019) also reported the part of the adsorption of     Tables 3 and  4, it can be found that the organic carbon content dominates the adsorption process.

Degradation
The degradation kinetics of all seven neonicotinoids in all four soils were well-fitted to the first order reaction model C s = C i e −kt , where C t and C 0 represent the concentration and the initial concentration of the neonicotinoids in the soil respectively, and k is the degradation rate constant. The root mean square deviations (RMSE) are used to indicate the quality of the fitting curves, and these values ranged from 0.2 to 0.26. Figure 2 presents the degradation data of imidacloprid in the four soils at three different depths in air-dried soils and the regression curves according to the first order reaction model. The degradation data of all other six neonicotinoids in the four soils and at three moisture contents (air dried, 20, and 40%) followed the same trend. The halflives of each neonicotinoid at different soil conditions can be calculated using the equation T 1/2 = 0.693/k. The standard deviations for all the results in Table 5 Table 5 summarizes the half-lives of all neonicotinoids in soils from four sampling sites and at three different depths and three moisture contents. It can be seen in the table that the half-lives of the neonicotinoids range from 33 to 305 days in all these soil samples. In general, the ranking of the half-lives is clothianidin > thiamethoxam > imidacloprid > acetamiprid > dinotefuran > thiacloprid > nitenpyram. The effects of the soil type on the half-lives are noticeable, but not as significant as the neonicotinoid type. Under the same moisture condition, the same neonicotinoid has the highest half-life in BHU soil and the lowest in TCB soil. Their half-lives in CRS and TSU soils have no significant  Fig. 2 Degradation of imidacloprid in BHU, CRS, TCB, and TSU soils at three depths in air-dried soils differences. Considering the physicochemical parameters of the soils (Table 2), the significant differences between BHU soil and TCB soil exist in clay content (3% for BHU and 20% for TCB) and CEC (4-7 cmol/kg for BHU and 11-12 cmol/kg for TCB). According to Das and Adhya (2015), clay content associated with iron oxides can increase the surface area and induce surface catalyzed hydrolysis. This could be a cause for the significant differences in half-lives of the neonicotinoids in BHU and TCB soils; the latter had higher clay content and in which the neonicotinoids had shorter half-lives. The half-lives of all the neonicotinoids also increase with the soil depth as shown in Table 5. The soil properties in Table 2 show that OC is the only parameter that decreases significantly with soil depth. Therefore, it is very likely that the OC in the soil facilitates the degradation of the neonicotinoids. Anhalt et al. (2008) reported a similar result that the degradation rate of imidacloprid in subsurface was much slower than that in the surface soil. The higher degradation rate of the insecticides in soils with higher OC could be attributed to the higher activities of soil bacteria. Tao and Yang (2011) and Ou et al. (2020) reported that microorganisms such as Pseudomonas and Actinomycetes are effective for degrading insecticides since these species use these chemicals as nutrients. The higher amounts of OC could modulate the microbial activities and then facilitate the degradation of these insecticides.
The data in Table 5 also reveal that the half-lives of all neonicotinoids decrease with the increase of soil moisture content, indicating that hydrolysis could play a key role in the degradation process. Similar results for clothianidin and thiamethoxam were also reported by Li et al. (2018) and Gupta et al. (2008). The biochemical degradation could be another route of the degradation of the insecticides. The microbial activity seems to be negligible in dry condition. With moisture content increases, the microbial activity also increases. However, the results in Table 5 show that, when the moisture content increased from 20 to 40%, there was no significant increase in the degradation rates. This indicates that, at least in this study, the hydrolysis was the main process contributing to the increase of degradation rates with moisture content because microbial activity would increase with the increase of moisture content, whereas hydrolysis could have reached the highest level when the moisture content reached about 20%.
The Groundwater Ubiquity Score (GUS) was used in this study to predict the leachability of a chemical into groundwater (Gustafson 1989). The GUS is calculated with the following equation: GUS = log 10 (T 1/2 ) × [4 -log 10 (K oc )]. According to this study, when the GUS value is great than 2.8, the chemical is defined as a leacher; when it is less than 1.8, a non-leacher; and when it is between 1.8 and 2.8, a possible leacher. The calculated GUS values of all the neonicotinoids in different soils at different depths and different moisture contents are presented in Table 6 and Fig. 3. In this study, the GUS values for imidacloprid was calculated to be between 2.0 and 2.9 in all the tested soils, clothianidin 2.0-2.9, thiacloprid 1.5-2.1, thiamethoxam 2.1-3.0, dinotefuran 1.8-2.7, acetamiprid 2.0-2.8, and nitenpyram 1.7-2.8. If an insecticide has low GUS value and short half-life in a soil, its transport in the soil will be limited because it is slow to leach and fast to degrade. This is the case for the neonicotinoids in TCB soil where they have the lowest GUS values and the shortest half-lives compared to other soils tested. All neonicotinoids in the tested soils except thiacloprid in BHU, CRS, and TCB soils and nitenpyram in TCB soil with 40% moisture content are possible leachers, having potentials to permeate through groundwater. TSU soil provides the most leachable environment for the neonicotinoids.

Conclusions
The adsorption of all the seven neonicotinoids in the four agricultural soils from three states (Mississippi, Arkansas, and Tennessee) in the USA followed a linear isotherm. The K d values were found to be below 2.0 L/kg for all the neonicotinoids in all the soils in Mississippi and Arkansas. Only in the Tennessee soil, the K d values ranged from 0.96 to 4.21 L/kg. In all the soils, nitenpyram had the lowest K d and thiacloprid the highest. The K d values for all insecticides are closely related to the OC in the soils-the higher the OC, the greater the K d value. The calculated ΔG values of all neonicotinoids in all soils ranged from − 14.6 to − 19.5 kJ/mol, indicating that the adsorption between insecticides and soils is mainly van der Waals force, which result in a weak and reversible adsorption process. The degradations of all insecticides in all tested soils followed the exponential decay model, and half-lives of the neonicotinoids ranged from 33 to 305 days. The correlation between half-life and soil property parameters indicates a positive effect of clay content and/or CEC on the degradation rates of the neonicotinoids. The Groundwater Ubiquity Scores calculated from the leachability index model indicate that all neonicotinoids, except for thiacloprid, in the tested soils are possible leachers, having the potential to permeate into or through groundwater zones.
Author contribution Yang Li was responsible for conducting the tests and drafting the manuscript. Yadong Li supervised the project, interpreted the data and revised/finalized the manuscript. Guihong Bi was responsible for testing the soil properties and revising the manuscript. Timothy Ward and Lin Li provided technical support for the analysis the samples and revised the manuscript.
Funding This work was supported by the National Science Foundation under Grant No. 1900451.
Data availability All data from this study can be obtained from the corresponding author.

Declarations
Ethics approval The authors in the study have read and approved the work and have given their consent to the submission and publication of the manuscript.

Consent to participate Not applicable.
Consent for publication Applicable.

Competing interests
The authors declare no competing interests.