Relationship between biochars’ porosity and adsorption of three neutral herbicides from water

Pesticides contribute a lot to water pollution in many regions around the world, and adsorption is still a cost-efficient method for removing water pollutants. Being a class of cheap adsorbents produced by pyrolysis of various biomass resources, biochars have exhibited their adsorption capacity for many kinds of water pollutants including both heavy metals and organics (Chen et al. 2011; Hyder et al. 2014), among which the adsorption of herbicides is of particular significance (Cao et al. 2009; Sun et al. 2011; Uchimiya et al. 2012). Biochar additions to soil could be beneficial to decrease the runoff and leaching risks of herbicides for surface water and groundwater, so that the major source of water pollution by herbicides could be controlled (Kookana 2010; Lü et al. 2012).

The adsorption of herbicides is significantly related to the surface properties of biochars, which are then varied with heat treatment techniques and feedstocks used for preparing biochars (Graber et al. 2012; Li et al. 2013). Generally, the biochars obtained at relatively high heat treatment temperatures (HTTs, e.g. >450 °C) have developed porosity and high surface area, and surface adsorption would dominate the adsorption of organic herbicides on these porous biochars (Mukherjee et al. 2011; Wang et al. 2013). However, a distinct herbicide adsorption gap between two biochars with similar surface area (Brunauer–Emmett–Teller (BET) specific surface area ≈ 400 m²/g) was observed in our preliminary study, which drives us to probe the potential influence of biochars’ porosity on adsorption. As a large portion of biochars’ surface area consists of micropores (Li et al. 2015), most of which would be too small to be accessible for the nano-scale herbicide molecules, it may be hypothesized that the mesopores of biochars would play a critical role in adsorption of herbicides. For the purpose of verifying this hypothesis, the surface characteristics and porosity of six biochars produced from different feedstocks and at relatively high HTTs (450–800 °C) were examined, and adsorption of three neutral herbicides on these biochars was measured in this study.

Six biochars were produced via heat treatment of alkali lignin (Aldrich) powder, softwood (Pinus radiata) and hardwood (Populus euramericana) shavings, respectively, in nitrogen atmosphere in a box furnace (interior dimensions: 30 cm long × 20 cm wide × 20 cm high), or under oxygen-limited condition in a muffle furnace at different HTTs (see Text S1 in the Supplementary material for details, available with the online version of this paper) (Li et al. 2015). The specific feedstocks and heat treatment conditions for the six biochars are listed in Table 1, and they are referred to hereafter as L800n, S800n, S600n, S650, S450 and H550 biochars, respectively, where the prefixes ‘L’, ‘S’ and ‘H’ refer to different feedstock of lignin, softwood and hardwood, respectively; the numbers represent the HTT used in degrees Celsius, while the suffix ‘n’ indicates nitrogen atmosphere. Those without suffix were obtained under oxygen-limited condition.

Three herbicides, metolachlor, atrazine and isoproturon, were used in this study to evaluate the adsorption capability of biochars. All herbicides have often been used in previous studies to evaluate the influence of biochars on environmental fate of pesticides. Their fundamental physical and chemical properties are listed in Table S1 in the Supplementary material (available with the online version of this paper). The adsorption isotherms of herbicides on biochar samples were determined by batch equilibration (for 48 h at 25 °C) of the solid biochars in aqueous solutions of various initial herbicide concentrations (C₀, μmol/L). Background solution contained 0.01 mol/L CaCl₂ to maintain constant ionic strength. The initial herbicide concentrations varied with the saturated concentration (C_s, μmol/L) of different herbicides, namely 300–1,400 μmol/L for metolachlor (C₀/C_s = 0.15–0.75), 20–120 μmol/L for atrazine (C₀/C_s = 0.15–0.85), and 50–250 μmol/L for isoproturon (C₀/C_s = 0.15–0.80). The volume of herbicide solution used for adsorption experiments was 30 mL for metolachlor, and 50 mL for both atrazine and isoproturon, respectively. The mass of biochar sample added to each herbicide solution was selected to obtain 20%–80% uptake of initially added sorbate after equilibrium. The initial pH of the mixed suspension was adjusted to 6.5 ± 0.1 with droplets of HCl or NaOH solutions, and maintained at 6.0–7.0 during the adsorption experiment. At the pH used in the adsorption study, all three herbicides will exist dominantly in the form of neutral molecules.

The analysis of herbicide concentration in the equilibrium solutions (C_e, μmol/L) was conducted in an LC-20A high-performance liquid chromatography system (Shimadzu) equipped with an ultraviolet detector. The wavelength for herbicide analysis was 218, 225 and 239 nm for metolachlor, atrazine and isoproturon, respectively. The amount adsorbed (Q_e, μmol/g) was calculated from the difference in concentration between the initial (C₀) and the equilibrium (C_e) solutions. Blanks without herbicide and duplicates of each adsorption point were used in every experiment series.

Surface characteristics of various biochars are presented in Table 2. All biochars, except the S450 biochar produced at relatively lower temperature, have high surface area (>300 m²/g) determined by the BET method, and also have developed micropores (SA_micro > 200 m²/g, and PV_micro > ≈0.1 cm³/g). The higher surface areas of biochars produced at higher HTTs are consistent with many previous reports (Zhang et al. 2011; Trigo et al. 2016), but little research has been reported previously to determine the variety of biochar porosity. In this study, the S800n biochar produced at higher HTT (800 °C) has much higher surface areas (SA_BET, SA_micro and SA_meso) and pore volumes (PV_micro and PV_meso) than the S600n biochar derived from the same feedstock. The three biochars (S650, H550 and S450) obtained under oxygen-limited condition also exhibit the same trend; namely higher HTT results in the biochar with more developed porosity (higher surface areas and pore volumes). Despite this common changing trend, the apparent variety of meso-porosity (pore diameter of 1.7–50 nm) was observed on the biochars made from different feedstocks and under different pyrolysis conditions. First, the lignin biochar (L800n) exhibited more developed mesopores than the softwood biochar (S800n) produced under the same pyrolysis condition. The SA_meso and PV_meso of the L800n biochar were 3.2 times and 5.7 times of those of the S800n biochar, although the BET surface area of the two biochars is almost equal. The reason could be related to the inorganic residuals (particularly alkalis) in lignin as the by-product of pulping process, and these inorganic compounds could activate the erosion of pore wall of lignin char and result in the pore expansion (Gonzalez-Serrano et al. 1997; Raymundo-Piñero et al. 2005; Adinata et al. 2007). Second, heat treatment of wood feedstock under oxygen-limited condition also results in the more developed mesopores than that in nitrogen atmosphere. The SA_meso and PV_meso of the S650 biochar are 4.5 and 5.7 times those of the S800n biochar, although the latter was produced at a higher HTT. This uncommon enhancement of biochars’ meso-porosity can only be attributed to the oxygen-containing atmosphere for producing the S650 biochar, which was characterized with higher oxygen content and higher O/C ratio than the S800n biochar (Table 1). The limited quantity of oxygen in heat treatment atmosphere would favor the slight burning of the pore wall, which led to the connection or expansion of some micropores into mesopores. The hardwood biochar (H550) obtained under oxygen-limited condition also has a higher SA_meso and PV_meso than the softwood S600n biochar, although both biochars have the BET surface area close to 300 m²/g. Pyrolysis of softwood at lower HTT of 450 °C resulted in a biochar of less developed porosity, which could be related to the incomplete carbonization of biomass feedstock (Chen et al. 2008; Keiluweit et al. 2010). In conclusion, the L800n and S650 biochars have the more developed mesopores according to the values of SA_meso/SA_micro and PV_meso/PV_micro in Table 2, and the mesopore development should benefit from the alkalis in feedstock for the L800n biochar and from the oxygen-containing atmosphere in heat treatment for the latter one (S650).

As can be seen, the higher contribution of surface adsorption to the total sorption was observed on the biochars made at relatively high HTT (L800n and S800n), corresponding to the higher carbonization degree as indicated by lower H/C ratios. According to the Q_A/Q_e values in Figure 5, surface adsorption played a dominant role in most cases for sorption of herbicides on the biochars, except for the S450 biochar produced at relatively lower HTT (Chen et al. 2008; Keiluweit et al. 2010). For this low temperature biochar, the partition in un-carbonized fraction would also contribute a lot to the total sorption (Chen et al. 2008; Zheng et al. 2010).

The adsorption data of herbicides to the biochars are fitted to the Langmuir model well, with all correlation coefficients R² > 0.90 (Table S2 in the Supplementary material). According to the regression results with both MAP and Langmuir models, we can conclude that the herbicide adsorption on the biochars used in this study is mainly attributed to the surface adsorption. And the adsorption of organic pollutants on biochars is dependent on the surface area of biochars according to many previous reports (Graber et al. 2012; Zhang et al. 2013; Trigo et al. 2016). In this context, the surface area-normalized adsorption based respectively on the BET surface area (Q₀/SA_BET) and mesopore area (Q₀/SA_meso) was calculated to investigate the influence of surface area of biochars on herbicide adsorption, with results shown in Table 3. When the same herbicide (e.g. metolachlor or atrazine) is considered, the apparent deviation of Q₀/SA_BET values for different biochars can be observed. For example, the Q₀/SA_BET values obtained for the L800n biochar were 6 to 10 times of those for the S600n biochar. In contrast, the Q₀/SA_meso values are much closer to each other for different biochars for adsorption of the same herbicide, and correlation of Q₀ with both SA_meso and PV_meso showed that Q₀ values were in direct proportion to the two mesopore parameters (see Figure S1 in the Supplementary material, available with the online version of this paper). So it is reasonable to propose that the adsorption capacity is dependent on the size of biochars’ mesopores (1.7–50 nm). The surface area-normalized adsorption obtained with the MAP model (Q_A/SA_BET vs. Q_A/SA_meso in Table 3) also implies that it is more appropriate to use the mesopore area to evaluate adsorption capacity of biochars, because a difference by an order of magnitude was observed between the large and small Q_A/SA_BET values for different biochars (e.g. 0.619 for L800n vs. 0.0386 for S600n) if the BET surface area (SA_BET) was used for comparing the adsorption capacity. Recently, Xiao & Pignatello (2015) found that the mesopores of hardwood biochars are critical for adsorption of triazine herbicides, and Li et al. (2015) proposed that the adsorption of herbicides acetochlor and metribuzin was determined by the biochars’ mesopore area when they investigated the role of some soil minerals in changing the herbicide adsorption on biochars. The results in this study are in good agreement with these published observations on the dominant role of mesopores in adsorption of organic herbicides.

The reason for the dominant role of mesopores in herbicide adsorption could be related to the molecular dimension of herbicides, which makes them inaccessible to some micropores, because almost all the herbicides have a molecular weight >200 g/mol, corresponding to an estimated molecular length >1.0 nm (Chen et al. 2004). Pelekani & Snoeyink (1999) pointed out that the pore size of carbons should be large enough for adsorption to occur, and Nguyen et al. (2007) proposed that the micropores with nominal width less than 1.0 nm in wood chars were not significant for adsorption of 1,2,4-trichlorobenzene with average dimension of 0.5 nm. Recently, Dong et al. (2014) suggested that the minimum pore diameter for adsorption of organic pollutants with molecular weight >200 g/mol should be 1.7 nm for activated carbons. Therefore, it may be concluded that a significant fraction of micropores (pore diameter <1.7 nm) in biochars would be inaccessible for herbicide molecules (Li et al. 2015).

The surface adsorption of three neutral herbicides on the biochars is dominated by the mesopores of biochars. The results implied that large BET surface area of biochars did not guarantee good adsorption for organic pollutants, particularly for organic compounds with relatively large molecular dimension. And the meso-porosity of biochars should be analyzed before evaluating the biochars’ potential for adsorption of nano-scale organic pollutants. In addition, an oxygen-containing atmosphere for heat treatment and alkalis in feedstock would favor the expansion of micropores into mesopores, which will be useful for producing biochars with high adsorption capacity for organic pollutants.

Funding was provided by the National Natural Science Foundation of China (41271475) and Natural Science Foundation of Zhejiang Province, China (LY16B070003).