Standard atrazine (purity 99.2%) was obtained from Shenyang Research Institute of Chemical Industry (Shenyang, China). HPLC-grade acetonitrile, formic acid, and sodium azide (NaN₃) were purchased from Sigma-Aldrich (Steinheim, Germany). All analytical reagents including calcium chloride (CaCl₂), sodium chloride (NaCl), anhydrous magnesium sulfate (MgSO₄), sodium hydroxide (NaOH), and hydrochloric acid (HCl) were obtained from Sinopharm Chemical Reagent Co., Ltd (Beijing, China). MMT was purchased from Shanghai Macklin Biochemical Co., Ltd (Shanghai, China). Ultrapure water was prepared using a Milli-Q reagent water system (Bedford, MA, United States).

Peanut shells were collected from the Shandong Province of China, washed with tap water to remove the surface dust and soil, dried at 70°C, and then crushed by using a high-speed pulverizer and passed through an 18-mesh sieve. A stable MMT suspension was prepared by adding MMT powder to 100 ml ultrapure water and mixed with ultrasound for 30 min. Then, 20 g peanut shells were immersed in MMT suspensions, stirred by magnetic stirrer for 2 h, and dried in oven at 105°C for 12 h. The impregnated samples were placed in a quartz boat of a tubular furnace (Zhonghuan Experimental Furnace Co., Ltd., Tianjin, China), pyrolyzed in a continuous flow of N₂, heated at a rate of 5°C/min, and maintained for 1 h when the pyrolysis temperature reached 600°C. The produced biochar samples were crushed, passed through an 80-mesh sieve, stored in a brown ground glass bottle, and sealed in a dryer. In total, five types of MMT–biochar composites were prepared by changing the mass ratio of MMT to BC (i.e., 20, 25, 30, 40, and 50%), and the resulting products were designated as 20% MMT/BC, 25% MMT/BC, 30% MMT/BC, 40% MMT/BC, and 50% MMT/BC, respectively. At the same time, peanut shells were treated with the same method to prepare BC as control. Unless otherwise specified, the MMT/BC mentioned later is 25% MMT/BC. By changing the pyrolysis temperature, MMT/BC under different temperature conditions was prepared, and the pyrolysis temperature was set to 400, 500, and 600 °C, respectively.

Elemental analysis using surface structure, scanning electron microscope (SEM), Fourier transform infrared spectroscopy (FTIR), and X-ray photoelectron spectrometer (XPS) was based on our previous analysis methods (Wang et al., 2020). Transmission electron microscopy (TEM) (FEI, Thermo Fisher Scientific, United States) was used to observe the microstructure of the biochar. Zeta potential of biochar was determined using a Zetasizer Nano ZS90 (Malvern Instruments, Malvern, United Kingdom). The pH value of the solution was measured by using a pH meter (FiveEasy Plus, Mettler Toledo, United States).

Two kinds of agricultural soil without detectable atrazine were collected from Jilin (44°23′52″N, 125°10′2″E) and Shandong provinces (36°12′43″N, 116°41′29″E) of China, denoted as S-DB and S-SD, respectively. The soils were taken at the top 10 cm and screened by using a 2 mm sieve for sorption and degradation experiments. The physicochemical properties of the S-DB showed it to be sandy loam (65.4% sand, 15.1% silt, and 19.5% clay) with pH 7.83, 41.0 g/kg organic matter, 1.15% organic carbon, and 17.2 cmol (+)/kg CEC. The S-SD also belonged to sandy loam (57.9% sand, 17.3% silt, and 24.8% clay) with pH 7.24, 29.2 g/kg organic matter, 0.39% organic carbon, and 21.4 cmol (+)/kg cation exchange capacity (CEC).

To determine the effect of MMT dosage (0, 20, 25, 30, 40, and 50%) and pyrolysis temperature (400, 500, and 600°C) on the removal of atrazine, 100 mg biochar and 40 ml of 10 mg/L atrazine aqueous solution (contained 0.01 mol/L CaCl₂ and 200 mg/L NaN₃) were introduced into 50 ml brown sample bottles. The bottles were then sealed and shaken in the dark on a reciprocating shaker at 180 rpm and 25 ± 1°C for 72 h. The pH of the mixtures was adjusted to 7.0 ± 0.2 using 0.1 M HCl/NaOH and monitored during and after sorption. The same conditions were used to test the effects of initial pH in the range of 2–9 on MMT/BC sorption. After the experiments, the bottles were put to rest for 10 min to make the biochar samples sink into the bottom by gravity to achieve solid–liquid separation, and the supernatant was filtered using a 0.22 μm membrane filter. The filtered solutions were then stored at −20°C until analysis.

The sorption kinetic and isotherm of atrazine were carried out on BC and MMT/BC using the same conditions and procedures, as described before. The kinetic tests were conducted with an initial atrazine concentration of 10 mg/L, 150 mg of BC, or 50 mg of MMT/BC per 40 ml atrazine solution. Duplicate bottles were sampled at predetermined times (1, 2, 4, 6, 8, 10, 12, 24, 48, and 72 h) and analyzed for atrazine concentration in the solution. In the sorption isotherm experiment, 150 mg BC or 50 mg MMT/BC was mixed with 40 ml atrazine solution with the concentration of 0.25–30 mg/L for 72 h.

Under the same conditions, the solutions without biochar were set for the control test. The loss of atrazine was negligible, including hydrolysis and sorption on the bottle wall. Based on the initial and final concentration of atrazine in solution and the dosage of biochar, the sorption capacity of atrazine on biochar was calculated.

The sorption isotherms of atrazine on biochar and biochar–soil mixtures were determined by three repeated batch sorption experiments under eight initial concentrations of atrazine (0.25–30 mg/L). A 5 g aliquot of soil with or without 50 mg BC or MMT/BC was weighed into a 50 ml centrifuge tube. Then, 25 ml of the background solution containing 0.01 M CaCl₂ and 200 mg/L NaN₃ was added. According to the initial concentration, a certain amount of atrazine stock solution was added to the centrifuge tube, the cover was tightened, and then shaken for 72 h at 180 rpm and 25 ± 1°C in the dark. At the end of the experiment, the tubes were centrifuged at 4,000 rpm for 10 min, and 1.5 ml of supernatant was filtered into the injection vial through 0.22 μm membrane filter. The filtered solutions were stored at −20°C until analysis.

Before the degradation experiments, a part of the collected soil was pre-cultured with a moisture content of 60% of the water holding capacity in the dark incubator (25 ± 1°C) for 2 weeks, and the other part was sterilized at 121°C for 30 min in an autoclave (Lead-Tech Scientific Instruments Co., Ltd, Shanghai, China). MMT/BC was chosen to investigate the impacts on atrazine degradation in the two soils. First, an aliquot of 20 g soil (dry weight) or the biochar–soil mixture (4 g of MMT/BC) was weighed into a 500 ml glass bottle, and then 100 μL of the atrazine acetone stock solution (10,000 mg/L) was added into the bottle. The soil samples were kept for 10 min until the acetone volatilized completely and then vortexed for 5 min. Next, another 180 g soil was weighed into the bottle, vortexed for 30 min, and mixed well with the spiked soil; the final concentration of atrazine in soil was 5 mg/kg (dry weight). After that, 15 g of spiked soil was weighted into 50 ml pre-cleaned glass bottle, and the soil moisture content was adjusted with ultrapure water to 60% of the saturated soil moisture content. In the same operation, the sterilized ultrapure water containing 200 mg/L NaN₃ was added into the sterilized soil to inhibit the growth of microorganisms. The bottles were stoppered with cotton and cultured in the dark in an incubator under constant temperature (25 ± 1°C) and humidity (70%). The water content was monitored regularly during soil incubation to maintain the initial state of soil moisture. Triplicate sample bottles of each treatment were removed after predetermined incubation times (0, 1, 3, 5, 7, 14, 21, 30, 40, 60, 90, and 120 days), and the soil samples were freeze-dried and stored at −20°C until extraction. Details of the soil sample treatment are shown in Supplementary Text S1.

Different mathematical models including first-order, second-order, and Elovich models were used to fit the kinetics data (Inyang et al., 2014; Yao et al., 2014; Liu et al., 2015): First-order:   q t = q e ( 1 − e − k 1 t ) , (1) Second-order :   q t = k 2 q e 2 t 1 + k 2 q e 2 t , (2) Elovich:   q t = 1 β ln ( αβt + 1 ) , (3) where q_t and q_e represent the amount of atrazine removed at time t and at equilibrium, respectively (mg/kg); k₁ and k₂ are the sorption rate constants (1/h); α represents the initial sorption rate (mg/kg); and β represents the desorption constant (kg/mg).

Three models (Freundlich, Langmuir, and dual-mode) were used to fit the isotherm experimental data (Uchimiya et al., 2010; Zhang et al., 2011; Suo et al., 2019): Freundlich:   q e = K f C e n , (4) Langmuir:   q e = q m KC e 1 + KC e , (5) Dual-mode:   q e = bQC e 1 + bC e + K D C e , (6) where q_e represents the atrazine concentration on the solid-phase at equilibrium (mg/kg), C_e represents the solution concentration of atrazine at equilibrium (mg/L), K_f [(mg/kg)/(mg/L)ⁿ] and K (L/mg) represent affinity coefficient, n represents the Freundlich linearity constant, b represents the affinity constant (L/mg), K_D represents the partition domain coefficient (L/kg), and q_m and Q represents the maximum capacity (mg/kg).

The first-order reaction kinetic model was used to fit the data of atrazine degradation kinetics in soil and biochar–soil mixtures (Zhang et al., 2018): C t = C 0 e − kt , (7) where C₀ is the initial concentration of atrazine (mg/kg), C_t is the concentration of atrazine (mg/kg) at sampling time t (d), k is the degradation rate constant (d⁻¹), and the half-life (t_1/2) was ln2/k.

The experimental data were fitted by Origin 9.1. The statistical analyses were analyzed by SPSS 25.0 one-way analysis of variance (ANOVA).

The properties of biochar varied greatly between BC and MMT/BC. The elemental composition (C, N, H, and O) of BC and MMT/BC, presented in Table 1, suggested that the addition of MMT reduced the content of C, N, H, and O, especially the carbon content, from 87.35% to 50.91%. The surface area of MMT/BC was slightly decreased from 262.86 to 233.16 m²/g, which may be related to the pore coating or plugging caused by the existence of MMT minerals (Yao et al., 2014). The total pore volume and the average pore diameter of the MMT/BC were increased from 0.014 to 0.066 cm³/g and 4.44–12.02 nm in comparison with BC. This is probably because MMT is layered silicate mineral, which contains mineral elements with small surface areas and abundant transitional pores (Li et al., 2015; Li et al., 2017a; Yao et al., 2014). The N₂ adsorption–desorption isotherms and pore size distribution of MMT/BC and BC are shown in Supplementary Figure S1. As described by the International Union of Pure and Applied Chemistry (IUPAC), the isotherms of BC evolved to type I, which shows the characteristic of a microporous material (Jing et al., 2014). The N₂ isotherms of MMT/BC resemble those of type Ⅱ, and the observed hysteresis curves of MMT/BC is of type H4, indicating the narrow cracks and pores in the sorbent materials (Jing et al., 2014). In the pore size distribution curve (Supplementary Figure S1B), BC and MMT/BC exhibited heterogenous pore structures, and mesopore (2–50 nm) was the main pore structure.

The SEM images of BC and MMT/BC are shown in Figures 1A,B. Surface and pore filling mechanisms may be responsible for the removal of atrazine as the pores were observed on biochar surfaces. The SEM images of MMT/BC clearly showed that MMT particles adhered to the biochar surface and completely coated the pure biochar, changing the surface morphologies and providing more sorption sites (Liang et al., 2019; Premarathna et al., 2019). Also, the MMT did not block the pores of biochar, which ensures the accessible to the adsorbate molecules (Premarathna et al., 2019). TEM imaging of the MMT/BC (Figure 1D) showed that the biochar surface was widely covered by the layered structures compared with BC (Figure 1C). In addition, it is reported that the structure is a common clay structural morphology (Tyagi et al., 2006; Zhou et al., 2009).

As shown in FTIR (Figure 2), the spectrum at ∼3,400 cm⁻¹ showed stretching vibrations of O-H, ∼1,580 cm⁻¹ was related to stretching vibration of C=C of the aromatic ring (Chen et al., 2008; Keiluweit et al., 2010). The band observed around 1,050 cm⁻¹ was related to the Si-O functional groups on the MMT/BC surface (Tyagi et al., 2006; Chen et al., 2017). The bands appearing below 800 cm⁻¹ of MMT/BC could be attributed to Si-O stretching, Si-O-Mg bending, Si-O-Al bending, and Si-O-Si bending (Zhou et al., 2009; Chen et al., 2017).

Raman spectra (Supplementary Figure S2) showed two prominent peaks at ∼1,350 cm⁻¹ and ∼1,600 cm⁻¹, representing D-band originated from sp³ hybridization with disordered mode and G-band induced by crystalline graphitic/sp² carbon stretching vibrations with the tangential mode, respectively (Ferrari and Robertson, 2000; Akhavan, 2010). The higher ratio of I _D /I _G peak intensity means higher defect concentration and increased functional groups on the sorbents surface. The I _D /I _G decrease from 1.04 for BC to 1.00 for MMT/BC, which was not a significant change.

Zeta potentials of MMT/BC are depicted in Supplementary Figure S3. In the pH range between 2 and 9, the MMT/BC resulted in a zeta potential drop of 35 mV. The MMT/BC surface was negatively charged at pH values greater than the point of zero charge pH (pH_pzc, 2.7) and was positively charged at pH values lower than pH_pzc (Nandi et al., 2009).

The full XPS spectra representing the chemical composition and crystalline states are displayed in Figures 3A,E. For MMT/BC, the six peaks at 285.0, 400.0, 532.6, 74.0, 103.2, and 1,304.3 eV were attributed to the element of C, N, O, Al, Si, and Mg on the surface, respectively, while there are no obvious peaks corresponds to Al, Si, and Mg on the surface of BC. The high-resolution C 1s spectra (Figures 3B,F) could be deconvoluted into four peaks: C-C/C-H (∼284.8 eV), C-O (∼286.4 eV), C=O (∼288.0 eV), and O-C=O (∼289.5 eV) (Gao et al., 2018; Liu et al., 2015). The O 1s spectra (Figures 3C,G) could be attributed into three groups: C=O (∼531.6 eV), -OH (∼532.5 eV), and C-O (∼534.1 eV) (Lyu et al., 2017b; Zhou et al., 2007). The N 1s spectra of BC (Figure 3D) can be deconvoluted into two peaks at 398.7 and 400.6 eV, which can be assigned to the pyridinic N (C-N=C) and pyrrolic N (C-N). A new peak corresponding to graphitic N can be found at 402.0 eV for MMT/BC (Figure 3H) (Jansen and Van Bekkum, 1995; Lyu et al., 2017a; Mueller et al., 2015). Supplementary Table S1 listed their relative percentages on BC and MMT/BC. For C 1s, the C-O ratio of biochar increased from 8.74% to 12.09% after MMT was added, and the C=O ratio increased from 4.52% to 4.98%, indicating the participation of oxygen-containing functional groups (Liang et al., 2019). For O 1s, the -OH ratio of biochar increased from 42.47% to 67.22% after MMT was added, and the effect of MMT/BC on atrazine had hydrogen bond involvement.

The sorption isotherms (Figure 6) were fitted using the Freundlich model (Eq. 4), and the parameters of the atrazine in two agricultural soils and biochar–soil mixtures are shown in Table 3. The Freundlich model can well describe the sorption data of atrazine (R ² > 0.98). It was well documented that organic matter affects the sorption potential of soils, and higher organic matter content leads to more contaminants sorption (Fenoll et al., 2011). The organic matter contents of S-DB and S-SD were 41.0 and 29.2 g/kg, respectively, so the K_f of S-DB [19.01 (mg/kg)/(mg/L)ⁿ] was higher than that of S-SD [16.71 (mg/kg)/(mg/L)ⁿ]. The addition of biochar increased the sorption affinity of atrazine in soil, and the effect of MMT/BC is more obvious. For S-DB, the addition of BC increased K_f to 31.08 (mg/kg)/(mg/L)ⁿ and the addition of MMT/BC increased to 57.35 (mg/kg)/(mg/L)ⁿ; for S-SD, the addition of BC increased K_f to 36.32 (mg/kg)/(mg/L)ⁿ and the addition of MMT/BC increased to 60.48 (mg/kg)/(mg/L)ⁿ. The n values for biochar-amended soils were less than that of unamended soils, which indicated that the degree of isotherm non-linearity increased after biochar amendment, especially MMT/BC. Due to the non-linearity of the isotherms, the distribution coefficients (K_d) of three equilibrium concentrations (C_e = 0.05, 0.5, and 5 mg/L) were calculated (Table 3). Similar to K_f, the K_d of the soil was also increased with biochar amendment, and the K_d values of all the soils and biochar–soil mixtures decreased with increasing atrazine concentration. Previous studies have shown that water-soluble organic matters could be sorbed by biochar and compete with the added organic matter for limited sorption site or pores of biochar, thus reducing the sorption capacity of biochar (Cornelissen et al., 2005; Koelmans et al., 2009). However, organic substances may diffuse through the humic layer into biochar micropores overtime (Pignatello et al., 2006). Obviously, in the current research, atrazine may reach the micropores of biochar through the humic layer, thus improving the sorption capacity of atrazine.

The degradation parameters of atrazine in soils and biochar–soil mixtures for 120 days under sterilized and non-sterilized conditions are shown in Table 4. The degradation curves of atrazine in non-sterilized S-DB and S-SD were fitted well with the first-order reaction kinetics model (R ² > 0.96), and the half-lives were 29.1 and 40.5 days, respectively. The degradation of atrazine in S-DB was faster than that in S-SD; the main reason could be the high content of organic matter in S-DB, which can serve as an energy source to stimulate microbial activities and accelerate the degradation of pesticides (Gavrilescu, 2005; Briceño et al., 2007). However, in sterilized soils, the half-lives increased to 341.5 and 521.2 days, indicating that soil microbial was the main reason for atrazine degradation. After MMT/BC addition, the half-lives of atrazine in the non-sterilized soils increased (from 29.1 to 97.8 days for S-DB and 40.5 days to 126.7 for S-SD) and the degradation amount decreased (from 89.39% to 60.54% for S-DB and 81.69%–59.13% for S-SD). Previous studies have shown that biochar application in soil could reduce the biodegradation of pesticides (such as atrazine, acetochlor, diuron, and acetamiprid) by enhancing its sorption to the soil (Yang et al., 2006; Loganathan et al., 2009; Yu et al., 2011; Li et al., 2018) because the sorbed pesticides can only be biodegraded after being desorbed by the soil and diffused into the soil solution (Yu et al., 2011). The strong sorption of atrazine by MMT/BC made atrazine difficult to desorb, and atrazine was not easy to be degraded by microorganism; therefore, the degradation of atrazine was delayed. For sterilization treatment, the addition of MMT/BC accelerated the degradation of atrazine, indicating that biochar promoted the chemical degradation. It has been proved that biochar can also affect the chemical degradation of pesticides; Zhang et al. (2018) reported that the active groups on the biochar mineral surface played an important role in the chemical degradation of thiacloprid in biochar–soil mixtures. However, the mechanism of atrazine chemical degradation by MMT/BC remains to be further studied.