Biochar Preparation

Rice husks, which are common and easily available agricultural residues in South China, were obtained from the rural areas of Chengdu, Sichuan Province, as biomass feedstock. According to our previous studies, the most effective RHB (RHB600) was pyrolyzed at 600 °C with a heating rate of 20 °C/min using a muffle furnace. The detailed biochar preparation process is presented in our previous study [16].

PG and PG Leachate Collection

The PG and PG leachate in this study were collected from one of the largest phosphate fertilizer factories in Shifang, China, which is one of the major areas of phosphate production. The leachate samples were collected daily at randomly selected times for 30 days. The PG leachate samples were filtered through 0.45- μm syringe filters to remove large suspended solids and stored at 4°C for further analysis.

The electrical conductance (EC) and pH of the leachate samples were measured by an EC and pH meter. The total phosphorus (TP) and aqueous F concentrations in the samples were determined using the Mo-Sb colorimetric method and the ion-selective electrode method, respectively [17-18]. The Pb, Zn, and Ca concentrations were measured via atomic absorption spectrometry (AAS) [19]. The parameters of the PG leachate samples were as follows: pH = 3.82 ± 0.65, EC = 3.74 ± 0.78 ms/cm, TP = 236.52 ± 68.05 mg/L, F = 79.64 ± 22.05, Ca = 426.65 ± 72.88 mg/L, Zn = 2.84 ± 0.57 mg/L, and Pb = 0.21 ± 0.04 mg/L.

Batch Experiments

Adsorption batch experiments were performed by mixing 50 ml of PG leachate with different solid/ liquid ratios (e.g., 1:2, 1:4, 1:8, 1:10, and 1:20) with 3 g of RHB, according to our previous study [16], to measure the isotherms of the adsorption of P and F from the PG leachates onto the RHB. The mixture was placed in a 120-ml centrifuge tube and immediately shaken at room temperature and 250 rpm for 24 h. All batch experiments were conducted in triplicate. At the end of the experiments, the supernatant solutions were centrifuged at 4500 rpm for 10 min, filtered through 0.45-μm syringe filters, and stored at 4 °C for analysis. Additional adsorption experiments were conducted with RHB via scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM-EDX) to explore the adsorption mechanism of the biochar. The biochar samples (1 g) were mixed with real PG leachate solutions (50 ml), which were shaken at 250 rpm for 24 h. After equilibration, biochar was obtained and dried in an oven at 105 °C for 24 h.

Column Experiment

A fixed-bed column experiment was conducted to measure the breakthrough curve (time-concentration profile) to determine the dynamic adsorption capacity of the adsorbent [20]. In this study, a Plexiglas column was used as a fixed-bed column and packed with RHB to evaluate the ability of RHB to adsorb P and F from real PG leachate. Different amounts (0, 12.65, 31, 62, and 93 g) of adsorbent (bulk density = 0.40 ± 0.003 g/cm³) were packed in the columns at different bed heights (0, 2, 5, 10, and 15 cm) to determine the effects of the columns on the P and F removal. Different flow rates (1, 1.5, and 2 ml/min) were also used to identify the effect of the flow velocity on the dynamic removal of P and F. For each column, biochar was wet-packed with a moisture content of 5% (w/w) as an interlayer in a column, which was 300 mm in height, 45 mm in diameter, and 15.60 cm² in cross-sectional area. The top and bottom of each column were packed with washed quartz sand to sandwich and immobilize the biochar layer, which distributed the flow over the cross-sectional area. In addition, 75-μm nylon meshes were used to distribute the flow at the inlet of the column and stabilize the column content at the outlet of the column. A peristaltic pump was used to control the real PG leachate solutions and maintain a uniform flow rate of 0.5 ml/min through the columns from top to bottom. Effluent samples were collected from the bottom of the column during the experiments. All column experiments were conducted in duplicate. Six mathematical models, the Thomas, Yoon-Nelson, Adams-Bohart, Clark, modified dose‒response, and bed depth service time (BDST) models were used to study the characteristics of the P and F experimental breakthrough curves under different conditions [21].

Sorption Isotherms

To evaluate the isothermal adsorption characteristics of the adsorbate–adsorbent interactions, experimental data were fitted by the Langmuir and the Freundlich models by varying the initial P and F concentrations from real PG leachates with different solid/liquid ratios. Fig. 1 shows the isotherms of the P and F adsorption from PG leachate on RHB, and Table 1 shows the corresponding fitting parameters. Langmuir model and Freundlich model showed different adsorption behaviors. The Langmuir model is commonly used to elucidate monolayer adsorption behaviors such as precipitation, whereas the Freundlich model is more applicable for multilayer adsorption behaviors such as electrostatic attraction [22]. For P, the Langmuir model (R² = 0.98) clearly fits the experimental data better than the Freundlich model (R² = 0.90). Therefore, the adsorption process of P from the PG leachate on RHB is mainly controlled by monolayer adsorption behaviors. The Langmuir maximum adsorption capacity (Q_max) of RHB for P from PG leachates was 3838 mg/kg. Although the Q_max value was lower than that reported in previous studies for many modified biochars with P in laboratory aqueous solutions, RHB could better remove P from PG leachates than other pristine biochars in removing P from aqueous solutions. These results indicate that RHB is more suitable for treating P from PG leachates than engineered biochars because of its low cost and lack of secondary pollution, specifically for large-area treatment. Conversely, the Freundlich model better described the adsorption of F from PG leachate on RHB (R² = 0.92) than the Langmuir model (R² = 0.90), which suggests that the F adsorption process on RHB is mainly dominated by multilayer adsorption. The Freundlich K_f value [446 (mg^1-nLⁿkg^-1)] for the F-isotherm model was larger than that for the P-isotherm model, which reflects the high F affinity of RHB.

SEM-EDX Analysis

Fig. 2 shows the SEM-associated EDX elemental dot maps of representative P and F from the PG leachates before and after adsorption by biochar. The khaki, dark red, bright red, cyan, transparent green, yellow, emerald green, and blue dots in the elemental maps of P- and F-loaded biochar samples denote the concentrations of C, O, P, F, Ca, Al, Mg, and S, respectively. The EDX elemental mapping results show that C, O, Al, and Mg were the major elements in the biochar structure. However, the intensity and brightness of O on the biochar surface in the elemental maps increased after adsorption due to the phosphate adsorption on the biochar [23]. Importantly, the intensities of P, F, Ca, and Al in the elemental maps increased after the adsorption, which confirms that P, F, and metals were absorbed by the biochar. The brightness and distribution outlines of the bright red and cyan dots in the elemental maps of P and F were associated with those of the khaki (C), transparent green (Ca), yellow (Al), and emerald green (Mg) dots in the maps, which indicates a correlation among the adsorption of P and F and the basic elements C, Ca, Al, and Mg. The elemental map data may indicate that the process of P and F adsorption can be divided into two steps: first, metal ions (Ca, Al, and Mg) from the PG leachates were absorbed on the biochar through electrostatic attraction and complexation [24]; second, the absorbed metal ions acted as bridges to precipitate P and F from the PG leachates, which made relatively insoluble Ca/Mg-P, Ca/Mg-F, or Al-F compounds immobilize P and F in the biochar [25-26]. However, C- and O-containing surface functional groups could also contribute to the adsorption of P and F from leachates onto the biochar surface through ion exchange, electrostatic interactions, or surface precipitation [27], although the correlation between P and O in the elemental maps was not significant due to the impact of endogenous phosphorus in biochar. In addition, our elemental dot maps revealed the presence of S in the biochar, and the intensity and brightness of S did not obviously change before and after adsorption, which indicates that sulfate ions did not significantly affect the adsorption of P and F from PG leachates by the biochar.

Fig. 1.

Adsorption isotherms of RHB for (a) P and (b) and (c) F adsorption from PG leachate. The results are shown as averages and standard deviations (n=3). The lines represent the model fittings.

Fig. 2.

Scanning electron microscopy (SEM) images and energy-dispersive X-ray spectroscopy (EDX) elemental dot maps of RHB before (a) and after (b) adsorption. Elemental dot maps of C (khaki), O (dark red), P (bright red), F (cyan), Ca (transparent green), Al (yellow), Mg (emerald green) and S (blue) dots, which denote the C, O, P, F, Ca, Al, Mg and S concentrations, are mainly associated with the adsorption of P and F.

Fig. 3.

Breakthrough curves of P [(a), (b), (c) and (d)] and F [(e), (f), (g) and (h)] from PG leachate with an RHB-packed column at different dosages (bed heights): (a), (e) 12.65 g; (b), (f) 31 g; (c), (g) 62 g; (d), (h) 93 g. The breakthrough curves were fitted with five different breakthrough models. The experimental data (solid grey circles) were simulated with various models.

Fixed-bed Column Adsorption of P and F from PG Leachates

Effects of the Biochar Dosage

Fig. 3 shows the effects of different biochar dosages on the column adsorption of P and F from PG leachates. Four biochar dosages (12.65, 31, 62, and 93 g) were used to establish the column adsorbent layer. The results revealed much lower P and F concentrations in the effluents of all four biochar columns than in the control (i.e., a column filled with only sand), which indicates immediate P and F breakthrough, i.e., RHB in the fixed-bed columns can positively remove P and F from PG leachates. However, the breakthrough curves of P and F had considerably different profiles. The slope of each breakthrough curve of P markedly changed from high to low during the adsorption process, possibly because a transition to an interparticle control mechanism caused the diffusion in macro/micropores or surface reaction kinetics [28]. In addition, the slopes of the P and F breakthrough curves decreased with increasing biochar dosage due to a greater mass transfer zone. The contact time between PG leachates and RHB increased with increasing adsorbent dosage (bed height), which increased the P and F removal and decreased the slope of the breakthrough curves. Moreover, the surface area in the packed layer in the column increased with increasing adsorbent bed height, which contributed to the formation of adsorption sites on RHB to adsorb P and F.

The performance of RHB in fixed-bed columns can simulate real-life applications in P and F removal for PG leachate treatment. The breakthrough curve models (Thomas, Yoon-Nelson, Adams-Bohart, Clark, and the modified dose‒response models) were used for the practical design of fixed-bed filters. All experimental data were fitted by these models, and Table 2 lists the corresponding fitting parameters. The model fitting results show that although the experimental breakthrough curves were well fitted by all models (except the Adams-Bohart model), overall, all model simulations better fit the F experimental breakthrough curves than the P experimental breakthrough curves. The Adams-Bohart model had the worst fit to the experimental curves because it was only applied to the initial phase (approximately 50% concentration) of the breakthrough curves [21]. Hence, the performance of the F breakthrough curve at a bed height of 15 cm fitted by the Adams-Bohart model was very good (R² = 0.98; σ% = 3.12). For P, the Yoon-Nelson model exhibited the best fit because of the higher R² value (R² > 0.95) with average σ% among all models. The values of k_YN and λ of the Yoon-Nelson model decreased and increased with increasing bed height, respectively, because the increased contact time increased the adsorbate breakthrough time. Similarly, the parameter values of the Thomas model and the modified dose‒response model decreased with increasing bed height, which indicates that external and internal diffusion is not the limiting step [29]. In contrast, the Clark model exhibited the best fit to the F breakthrough curve at different adsorbent dosages with an average R² = 0.99 and the lowest σ%. The rate of mass transfer decreased with increasing bed height (dosage) because of the increased interactions between RHB adsorbent and F from the PG leachates with increasing adsorbent dosage, which reduced the mass transfer rate [30].

In addition, the BDST model was used to evaluate the effect of dosage on P and F removal by RHB in fixed-bed columns. The BDST model was applied to the breakthrough point at effluent and influent concentration ratios of C_e/C₀ = 0.6, 0.7, and 0.8, as shown in Fig. 4. The results indicate that the contact time at a certain effluent and influent concentration ratio increased with increasing bed height. Table 2 lists the best-fit model parameters. For P and F, the dynamic adsorption capacities obtained via BDST model fitting increased with increasing C_e/C₀ because P and F occupied more RHB active sites at higher C_e/C₀ values, which improved the adsorption capacity [31]. Moreover, according to the fitting results of q₀ (mg/L), RHB could remove more F than P under identical conditions. Therefore, it is feasible to use the BDST model to predict the operation of a fixed bed in a certain range of conditions, which has guiding significance for selecting the conditions for the practical application of RHB treatment of PG leachates on a large scale.

Fig. 4.

BDST model linear regression between the bed height (cm) [P (a) and F (b)] and the inverse of the linear velocity (min/cm) [P (c) and F (b)] and service time (min) at breakthrough points of C_e/C₀ = 0.6, 0.7 and 0.8.

Effect of the Flow Rate

Fig. 5 shows the effects of different flow rates (0.5, 1, 1.5, and 2 ml/min) on the dynamic removal of P and F from PG leachates using a column with an RHB dosage of 12.65 g (i.e., a bed height of 2 cm). In all flow rate experiments, the biochar column significantly affected the removal of P and F from the PG leachates, which indicates the effectiveness of RHB in treating P and F from real PG leachates. However, increasing the flow rate shortened the breakthrough time by reducing the contact time between P and F in the PG leachates and RHB in the column. At the lowest flow rate, the longest breakthrough times were 361 and 1113 min for P and F, respectively. In contrast, at the highest flow rate, the shortest breakthrough times were 63 and 285 min. These results indicate that a high flow rate can shorten the residence time of PG in the column and limit the exposure of P and F from PG leachates to RHB [32].

The five breakthrough curve models also fit the experimental data on the dynamic adsorption of P and F from PG leachates well. The models better fitted the flow rate data with higher R² and lower σ% values than the dosage data as shown in Table 2. Furthermore, the dynamic adsorption capacity q_m of the Thomas model and the modified dose‒response models gradually decreased for both P and F with increasing flow rate. As previously mentioned, this reduction is likely attributed to the decreased residence time of the PG leachates and limited contact time. However, with increasing flow rate, k_YN and λ of the Yoon-Nelson model increased and decreased, respectively, contrary to the increasing trend of dosage. Similarly, the excellent fitting quality of the Clark model for F might result from the batch adsorption process following the Freundlich model and a high degree of mass transfer during the entire process in the column [33]. The corresponding breakthrough points were fitted by the BDST model in Fig. 4, and the model well-fitted the linear relationship between the service time and the reverse of the linear velocity (Table 2; R² > 0.98 for P and F). Consistently, the fitting results of q₀ increased with increasing C_e/C₀, which confirms that RHB was successfully applied as a packing material to remove P and F from PG leachates.

Fig. 5.

Breakthrough curves of P [(a), (b), (c) and (d)] and F [(e), (f), (g) and (h)] from PG leachate with an RHB-packed column at a dosage of 12.65 g and different flow rates: (a), (e) 0.5 ml/min; (b), (f) 1 ml/min; (c), (g) 1.5 ml/min; (d), (h) 2 ml/min. The breakthrough curves were fitted with five different breakthrough models. The experimental data (solid grey circles) were simulated with various models.