Abstract
The waste pomelo peel was pyrolyzed at 400 °C to prepare biochar and used as adsorbent to remove norfloxacin (NOR) from simulated wastewater. The adsorption conditions of norfloxacin by biochar were optimized by response surface methodology (RSM). On the basis of single-factor experiment, the adsorption conditions of biochar dosage, solution pH and reaction temperature were optimized by Box-Behnken Design (BBD), and the quadratic polynomial regression model of response value Y1 (NOR removal efficiency) and Y2 (NOR adsorption capacity) were obtained respectively. The results show that the two models are reasonable and reliable. The influence of single factor was as follows: solution pH > biochar dosage > reaction temperature. The interaction between biochar dosage and solution pH was very significant. The optimal adsorption conditions after optimization were as follows: biochar dosage = 0.5 g/L, solution pH = 3, and reaction temperature = 45 °C. The Y1 and Y2 obtained in the verification experiment were 75.68% and 3.0272 mg/g, respectively, which were only 2.38% and 0.0242 mg/g different from the theoretical predicted values of the model. Therefore, the theoretical model constructed by response surface methodology can be used to optimize the adsorption conditions of norfloxacin in water.
HIGHLIGHTS
Constructed a reliable response surface model for biochar adsorption of norfloxacin in water.
The interaction of adsorption conditions and its influence on the adsorption of norfloxacin by biochar have been explored clearly.
The constructed response surface model can be used to adjust the adsorption conditions of biochar to adsorb norfloxacin in actual water.
INTRODUCTION
Antibiotics are widely used in human clinical and animal husbandry to treat infections, saving millions of lives (Cabello et al. 2016; Zhou et al. 2016). Fluoroquinolones are the most widely used antibiotics (Duan et al. 2019), showing good antibacterial activity against both gram-negative and gram-positive bacteria (Chen et al. 2013). The excessive use of norfloxacin in animal husbandry, sewage discharge from pharmaceutical factories and hospitals, incomplete metabolism of norfloxacin in animals and other factors has resulted in the widespread presence of norfloxacin in surface water (Jia et al. 2012; Van et al. 2014; Prutthiwanasan et al. 2016), with concentrations up to ng/L and even mg/L (Mir et al. 2017). In addition, the non-biodegradability and ecotoxicity of norfloxacin in water induced the generation of some drug-resistant bacteria, which threatened human health (Jojoa et al. 2017). Therefore, it is necessary to find an effective method to remove norfloxacin from water. The common methods to remove norfloxacin from water include photocatalysis, electrochemical degradation, biodegradation, Fenton reaction, advanced oxidation process, etc. The adsorption method has been widely used because of its high removal efficiency, low cost and easy operation (Wang et al. 2017).
Yang et al. studied the adsorption behavior of norfloxacin on porous resin and carbon nanotubes. It was found that the kinetic model and adsorption isothermal model of norfloxacin adsorption by the two materials could be well described by the pseudo-second-order kinetic equation and Langmuir equation respectively (Yang et al. 2012). Yan et al. studied the adsorption behavior of norfloxacin on barley straw and pretreated barley straw respectively, and studied the potential energy distribution of the adsorption process based on the Dubinin-Astakhov model. The results showed that the pretreated barley straw had higher potential energy, so it had more affinity for norfloxacin (Yan & Niu 2018). Tamiris Chahm et al. used sulfuric acid to activate termite feces to study its adsorption performance to norfloxacin, and the results showed that sulfuric acid activated termite feces is an efficient and low cost adsorption material (Chahm et al. 2019). Most of these studies focus on the preparation and adsorption behavior of adsorbents. It is also assumed that the factors affecting adsorption capacity are independent of each other. However, it is quite different from the actuality that there is a complex nonlinear relationship between the two factors. Response surface methodology is an effective method to build a model that can predict the whole based on limited experimental data, which can reduce the required preliminary data collection and explore the interaction of multiple factors (Zhou et al. 2019; Zhou et al. 2020). Therefore, this study intends to use the response surface methodology to establish the nonlinear response relationship between the influencing factors and the output factors, analyze the nonlinear interaction between the factors, and obtain the optimal norfloxacin adsorption conditions.
Pomelo is a fruit widely cultivated all over the world. Pomelo peel accounts for about 45% of the total weight of pomelo (Liang et al. 2014), and pomelo peel contains a large amount of cellulose and hemicellulose. However, the majority of pomelo peel is treated as agricultural waste, resulting in a huge waste of resources and environmental pollution (Zhang et al. 2018). Therefore, waste grapefruit peel was used as the carbon source of biochar in this experiment. The batch adsorption test was designed by Design Expert software. Box-Behnken Design (BBD) was used to establish the response surface model of biochar dosage, solution pH and reaction temperature on the removal efficiency and adsorption capacity of norfloxacin by biochar. The interaction of factors was discussed and the adsorption conditions were optimized in order to provide reference for similar studies.
MATERIALS AND METHODS
Instruments and materials
Experimental instruments: UV-1810 UV-visible spectrophotometer (Youke, China). PHS-25 digital pH meter (Shanghai thunder magnetic, China). SHA-B digital display constant temperature water bath oscillator (Guohua, China). Norfloxacin (98% purity) reagent was purchased from Aladdin pharmaceutical company (Shanghai, China). Its physical and chemical properties are show in Table 1. The hydrochloric acid and sodium hydroxide used in the experiment were analytically pure. The experimental water was ultra-pure water (UP).
Basic properties of NOR
Molecular formula . | Molecular mass . | pKa . | lgP . | Chemical structure . |
---|---|---|---|---|
C16H18FN3O3 | 319.33 | pKa1 = 6.34 pKa2 = 8.75 | 0.46 | ![]() |
Molecular formula . | Molecular mass . | pKa . | lgP . | Chemical structure . |
---|---|---|---|---|
C16H18FN3O3 | 319.33 | pKa1 = 6.34 pKa2 = 8.75 | 0.46 | ![]() |
Preparation of biochar
The recovered grapefruit peel was washed with ultra-pure water and dried at 70 °C to a constant weight. The dried grapefruit peel was stored in a crucible and placed in a muffle furnace, and then heated to 400 °C at a heating rate of 20 °C/min and pyrolyzed at this pyrolysis temperature for 2 h. The biochar was taken out after pyrolysis, ground, passed through a 60-mesh sieve, and placed in a sealed plastic bottle for later use.
Adsorption conditions experiment
Biochar dosage: The prepared biochar (0.0025, 0.0050, 0.0075, 0.0100 and 0.0125 g) was placed in a 50 mL centrifuge tube and norfloxacin solution (25 mL, 2 mg/L) was added. It was then placed in a water bath oscillator and oscillated at room temperature for 2 h at 150 rpm.
Solution pH: keeping the above conditions unchanged, biochar dosage was controlled at 0.4 g/L and the pH of norfloxacin solution was 3, 4, 5, 6 and 7, respectively.
Reaction temperature: the biochar dosage was controlled at 0.4 g/L, and the norfloxacin solution concentration was 2 mg/L. Other conditions remained unchanged, and the biochar was oscillated for 2 h at 25, 30, 35, 40 and 45 °C, respectively.
Pseudo-first-order, pseudo-second-order kinetics and intra-particle diffusion models were used to analyze the process of biochar adsorption of NOR. The equation is as follows:
Response surface methodology (RSM): the Box-Behnken design
The effects of biochar dosage, solution pH and reaction temperature on the adsorption of norfloxacin by biochar and their interactions were studied. The experimental process was the same as 2.3. Box-Behnken experimental design was used to optimize the adsorption conditions of norfloxacin by biochar. The norfloxacin removal efficiency Y1 (%) and norfloxacin adsorption capacity Y2 (mg/g) were taken as response values to control the initial concentration of norfloxacin at 2 mg/L. Five groups of parallel center point experiments were designed. The coding value and experimental value of 3 factors at 3 level are shown in Table 2. According to the optimization results, the optimal conditions are verified.
Factors and levels of Box-Behnken design
. | . | . | Coded Levels . | ||
---|---|---|---|---|---|
Factor . | Symbols . | Unit . | − 1 . | 0 . | 1 . |
Biochar dosage | X1 | g/L | 0.3 | 0.4 | 0.5 |
pH | X2 | 3 | 4 | 5 | |
Temperature | X3 | °C | 25 | 35 | 45 |
. | . | . | Coded Levels . | ||
---|---|---|---|---|---|
Factor . | Symbols . | Unit . | − 1 . | 0 . | 1 . |
Biochar dosage | X1 | g/L | 0.3 | 0.4 | 0.5 |
pH | X2 | 3 | 4 | 5 | |
Temperature | X3 | °C | 25 | 35 | 45 |
RESULTS AND DISCUSSION
Effects of single factor on norfloxacin adsorption by biochar
Exploring the effect of time on the adsorption of pollutants by adsorbents is a necessary process to understand the mechanism. As shown in Figure 1, the adsorption of NOR by biochar reached equilibrium at about 120 min, and the Qe value was 3.75 mg/g. The commonly used pseudo first-order and second-order kinetics are used to fit the experimental data. The results are shown in Figure 1, and the fitting parameters are shown in Table 3. Compared with the pseudo-first-order kinetics, the pseudo-second-order kinetics model has a higher fitting correlation (R2 = 0.980) to the experimental data, and the fitted Qe value (3.81 mg/g) is closer to the experimental value (3.76 mg/g). Therefore, the adsorption of NOR by biochar is a chemical adsorption process rather than a simple and reversible physical adsorption. In addition, the intra-particle diffusion model is also taken into consideration. As shown in Figure 2, the adsorption of NOR by biochar can be divided into two stages. The first stage is that NOR is adsorbed to the surface of biochar from the liquid interface and has a large adsorption rate k = 0.124. The second stage is that NOR diffuses in the pores of the biochar and has a small adsorption rate k = 0.014. In addition, some reports on the adsorption of NOR by adsorbents were used to compare with the grapefruit peel biochar prepared in this study (Table 4). Therefore, 120 min was used as the reaction time for subsequent adsorption experiments.
Fitting parameters of kinetic models for adsorption of NOR by biochar
Kinetic model . | k . | Qe . | R2 . |
---|---|---|---|
Pseudo-first-order | 0.169 | 3.67 | 0.738 |
Pseudo-second-order | 0.104 | 3.81 | 0.980 |
Kinetic model . | k . | Qe . | R2 . |
---|---|---|---|
Pseudo-first-order | 0.169 | 3.67 | 0.738 |
Pseudo-second-order | 0.104 | 3.81 | 0.980 |
Comparison of the adsorption performance of adsorbents for NOR in other reports with this study
Adsorbents . | Dosage (g/L) . | Concentration (mg/L) . | Qe (mg/g) . | Reference . |
---|---|---|---|---|
Cauliflower roots biochar | 2 | 10 | 4.63 | Qin et al. (2017) |
KOH-modified biochar | 10 | 30 | 2.80 | Luo et al. (2018) |
Magnetic biochar | 4 | 10 | 2.30 | Wang et al. (2017) |
Hematite–biochar composites | 4 | 8 | 1.68 | Yang et al. (2019) |
Magnetic biochar-based manganese oxide | 2 | 10 | 4.64 | Li et al. (2018) |
Pomelo peel-based biochar | 0.4 | 4 | 3.75 | This study |
Adsorbents . | Dosage (g/L) . | Concentration (mg/L) . | Qe (mg/g) . | Reference . |
---|---|---|---|---|
Cauliflower roots biochar | 2 | 10 | 4.63 | Qin et al. (2017) |
KOH-modified biochar | 10 | 30 | 2.80 | Luo et al. (2018) |
Magnetic biochar | 4 | 10 | 2.30 | Wang et al. (2017) |
Hematite–biochar composites | 4 | 8 | 1.68 | Yang et al. (2019) |
Magnetic biochar-based manganese oxide | 2 | 10 | 4.64 | Li et al. (2018) |
Pomelo peel-based biochar | 0.4 | 4 | 3.75 | This study |
The influence of biochar dosage on norfloxacin removal efficiency and adsorption capacity is shown in Figure 3(a). The removal efficiency increases with the increase of biochar dosage. When the dosage of biochar increased from 0.1 g/L to 0.5 g/L, the removal efficiency increased from 37.15% to 41.62%, the change was not obvious. However, the adsorption capacity of norfloxacin by biochar decreased sharply from 7.43 mg/g to 1.6648 mg/g. The effect of biochar dosage on the adsorption capacity of norfloxacin was greater than that of removal efficiency.
The effect of biochar dosage (a), pH (b) and reaction temperature (c) on the removal efficiency and adsorption capacity of NOR by biochar.
The effect of biochar dosage (a), pH (b) and reaction temperature (c) on the removal efficiency and adsorption capacity of NOR by biochar.
The effect of solution pH on norfloxacin adsorption by biochar is shown in Figure 3(b). When the pH rose from 3 to 6, the removal efficiency of norfloxacin dropped rapidly from 58.78% at pH = 3 to 39.03% at pH = 6, with a decrease of nearly 20%. Meanwhile, the adsorption capacity of norfloxacin by biochar decreased from 2.9390 mg/g to 1.9519 mg/g. When the pH changed from 6 to 7, the removal efficiency and adsorption capacity were almost unchanged.
The influence of temperature on norfloxacin removal efficiency and adsorption capacity of biochar is shown in Figure 3(c). When the temperature rose from 25 °C to 45 °C, the removal efficiency increased by 2.74%, and the adsorption capacity increased from 1.9425 mg/g to 2.0795 mg/g, with a slight increase. This indicates that the adsorption of norfloxacin by biochar is endothermic (Wang & Zhang 2020), and the increase of temperature is conducive to the adsorption of norfloxacin by biochar.
Box-Behnken design (BBD) and model analysis
Box-Behnken design experiment design and results with independent factors
. | Coded levels . | . | . | ||
---|---|---|---|---|---|
Run . | X1 . | X2 . | X3 . | Removal efficiency Y1 (%) . | Adsorption capacity Y2 (mg/g) . |
1 | 0 | 0 | 0 | 41.62 | 2.0810 |
2 | 0 | 0 | 0 | 41.75 | 2.0875 |
3 | 0 | −1 | −1 | 56.91 | 2.8455 |
4 | 0 | 0 | 0 | 41.58 | 2.0790 |
5 | 1 | −1 | 0 | 67.00 | 2.6800 |
6 | 1 | 1 | 0 | 43.59 | 1.7436 |
7 | −1 | 0 | −1 | 39.80 | 2.6533 |
8 | 0 | 1 | −1 | 38.85 | 1.9425 |
9 | −1 | 0 | 1 | 41.80 | 2.7867 |
10 | 0 | 1 | 1 | 41.94 | 2.0970 |
11 | −1 | −1 | 0 | 44.43 | 2.9620 |
12 | 0 | 0 | 0 | 41.60 | 2.0800 |
13 | 1 | 0 | −1 | 49.65 | 1.9860 |
14 | 1 | 0 | 1 | 55.66 | 2.2264 |
15 | −1 | 1 | 0 | 40.78 | 2.7187 |
16 | 0 | −1 | 1 | 59.72 | 2.9860 |
17 | 0 | 0 | 0 | 41.65 | 2.0825 |
. | Coded levels . | . | . | ||
---|---|---|---|---|---|
Run . | X1 . | X2 . | X3 . | Removal efficiency Y1 (%) . | Adsorption capacity Y2 (mg/g) . |
1 | 0 | 0 | 0 | 41.62 | 2.0810 |
2 | 0 | 0 | 0 | 41.75 | 2.0875 |
3 | 0 | −1 | −1 | 56.91 | 2.8455 |
4 | 0 | 0 | 0 | 41.58 | 2.0790 |
5 | 1 | −1 | 0 | 67.00 | 2.6800 |
6 | 1 | 1 | 0 | 43.59 | 1.7436 |
7 | −1 | 0 | −1 | 39.80 | 2.6533 |
8 | 0 | 1 | −1 | 38.85 | 1.9425 |
9 | −1 | 0 | 1 | 41.80 | 2.7867 |
10 | 0 | 1 | 1 | 41.94 | 2.0970 |
11 | −1 | −1 | 0 | 44.43 | 2.9620 |
12 | 0 | 0 | 0 | 41.60 | 2.0800 |
13 | 1 | 0 | −1 | 49.65 | 1.9860 |
14 | 1 | 0 | 1 | 55.66 | 2.2264 |
15 | −1 | 1 | 0 | 40.78 | 2.7187 |
16 | 0 | −1 | 1 | 59.72 | 2.9860 |
17 | 0 | 0 | 0 | 41.65 | 2.0825 |
The quadratic regression model of ANOVA for removal efficiency of NOR by biochar
Sources . | Sum of squares . | df . | Mean square . | F-value . | ρ-value prob. > F . |
---|---|---|---|---|---|
Y1 | 1096.26 | 9 | 121.81 | 80.75 | <0.0001** |
X1 | 301.23 | 1 | 301.23 | 199.70 | <0.0001** |
X2 | 494.55 | 1 | 494.55 | 327.86 | <0.0001** |
X3 | 24.19 | 1 | 24.19 | 16.03 | 0.0052* |
X1X2 | 97.61 | 1 | 97.61 | 64.71 | <0.0001** |
X1X3 | 4.02 | 1 | 4.02 | 2.67 | 0.1466 |
X2X3 | 0.0196 | 1 | 0.0196 | 0.0130 | 0.9124 |
X12 | 23.08 | 1 | 23.08 | 15.30 | 0.0058* |
X22 | 103.95 | 1 | 103.95 | 68.91 | <0.0001** |
X32 | 31.76 | 1 | 31.76 | 21.05 | 0.0025* |
Residual | 10.56 | 7 | 1.51 | ||
Pure Error | 0.0178 | 4 | 0.0045 | ||
Cor Total | 1106.82 | 16 |
Sources . | Sum of squares . | df . | Mean square . | F-value . | ρ-value prob. > F . |
---|---|---|---|---|---|
Y1 | 1096.26 | 9 | 121.81 | 80.75 | <0.0001** |
X1 | 301.23 | 1 | 301.23 | 199.70 | <0.0001** |
X2 | 494.55 | 1 | 494.55 | 327.86 | <0.0001** |
X3 | 24.19 | 1 | 24.19 | 16.03 | 0.0052* |
X1X2 | 97.61 | 1 | 97.61 | 64.71 | <0.0001** |
X1X3 | 4.02 | 1 | 4.02 | 2.67 | 0.1466 |
X2X3 | 0.0196 | 1 | 0.0196 | 0.0130 | 0.9124 |
X12 | 23.08 | 1 | 23.08 | 15.30 | 0.0058* |
X22 | 103.95 | 1 | 103.95 | 68.91 | <0.0001** |
X32 | 31.76 | 1 | 31.76 | 21.05 | 0.0025* |
Residual | 10.56 | 7 | 1.51 | ||
Pure Error | 0.0178 | 4 | 0.0045 | ||
Cor Total | 1106.82 | 16 |
Note: **means the difference is very significant (ρ <0.05). *indicates significant difference (ρ < 0.05).
The quadratic regression model of ANOVA for adsorption capacity of NOR by biochar
Sources . | Sum of squares . | df . | Mean square . | F-value . | ρ-value prob. >F . |
---|---|---|---|---|---|
Y2 | 2.61 | 9 | 0.2898 | 42.45 | <0.0001** |
X1 | 0.7717 | 1 | 0.7717 | 113.03 | <0.0001** |
X2 | 1.10 | 1 | 1.10 | 161.68 | <0.0001** |
X3 | 0.0559 | 1 | 0.0559 | 8.19 | 0.0243* |
X1X2 | 0.1201 | 1 | 0.1201 | 17.59 | 0.0041* |
X1X3 | 0.0029 | 1 | 0.0029 | 0.4192 | 0.5380 |
X2X3 | 0.0000 | 1 | 0.0000 | 0.0072 | 0.9349 |
X12 | 0.1596 | 1 | 0.1596 | 23.38 | 0.0019* |
X22 | 0.2618 | 1 | 0.2618 | 38.35 | 0.0004* |
X32 | 0.0783 | 1 | 0.0783 | 11.47 | 0.0116* |
Residual | 0.0478 | 7 | 0.0068 | ||
Pure Error | 0.0000 | 4 | 0.0000 | ||
Cor Total | 2.66 | 16 |
Sources . | Sum of squares . | df . | Mean square . | F-value . | ρ-value prob. >F . |
---|---|---|---|---|---|
Y2 | 2.61 | 9 | 0.2898 | 42.45 | <0.0001** |
X1 | 0.7717 | 1 | 0.7717 | 113.03 | <0.0001** |
X2 | 1.10 | 1 | 1.10 | 161.68 | <0.0001** |
X3 | 0.0559 | 1 | 0.0559 | 8.19 | 0.0243* |
X1X2 | 0.1201 | 1 | 0.1201 | 17.59 | 0.0041* |
X1X3 | 0.0029 | 1 | 0.0029 | 0.4192 | 0.5380 |
X2X3 | 0.0000 | 1 | 0.0000 | 0.0072 | 0.9349 |
X12 | 0.1596 | 1 | 0.1596 | 23.38 | 0.0019* |
X22 | 0.2618 | 1 | 0.2618 | 38.35 | 0.0004* |
X32 | 0.0783 | 1 | 0.0783 | 11.47 | 0.0116* |
Residual | 0.0478 | 7 | 0.0068 | ||
Pure Error | 0.0000 | 4 | 0.0000 | ||
Cor Total | 2.66 | 16 |
Note: **means the difference is very significant (ρ < 0.05). *indicates significant difference (ρ < 0.05).
Relationship between predicated and experimental data for response Y1 (removal efficiency) (a) and Y2 adsorption capacity (b).
Relationship between predicated and experimental data for response Y1 (removal efficiency) (a) and Y2 adsorption capacity (b).
The model predicted the optimal conditions for norfloxacin adsorption by pomelo peel biochar as follows: biochar dosage (X1) = 0.5 g/L, solution pH (X2) = 3, reaction temperature (X3) = 45 °C (corresponding coding value is 1, −1, 1). The verification test is performed according to the optimal adsorption condition. The removal efficiency and adsorption capacity of norfloxacin were 75.68% and 3.0272 mg/g, respectively. The predicted values of the model were 73.30% and 3.003 mg/g, which were almost no different from the experimental values. Figure 4 shows the comparison between the experimental value and the predicted value of the model. It can be found that the relative deviation between the experimental value and the predicted value of the model is small and distributed on both sides of the line. At the same time, it can be drawn from Tables 6 and 7 that the ρ value of the model with two response values of Y1 and Y2 is <0.0001, indicating that the fitted regression equation is extremely significant. And the determination coefficient R2 was 0.9905 and 0.9820 respectively, indicating that the model fitted well with the actual situation and the experimental error was ignorable. Therefore, the regression Equations (6) and (7) can be used to predict and analyze the effects of biochar dosage, solution pH and reaction temperature on removal efficiency (Y1) and adsorption capacity (Y2) during norfloxacin adsorption by grapefruit peel biochar.
As for the removal efficiency Y1, it can be judged from the ρ-value in Table 6 that the primary terms X1 and X2 are extremely significant and X3 is significant. The interaction term X1X2 was very significant, while X1X3 and X2X3 were not significant. The quadratic term X22 is extremely significant, while X12 and X32 are significant. For the response value of norfloxacin adsorption capacity Y2 (Table 7), the primary terms X1 and X2 were extremely significant, while X3 was significant. The interaction term X1X2 was significant, while X1X3 and X2X3 were not significant. The quadratic terms X12, X22 and X32 were all significant. Comparing the F values in Tables 6 and 7, it can be found that the influence of each single factor on removal efficiency (Y1) and adsorption capacity (Y2) is X2 (solution pH) > X1 (biochar dosage) > X3 (reaction temperature) (Wang et al. 2021).
Interaction effects of factors and response surface
The three-dimensional space surface diagram composed of the response value to each experimental factor can directly reflect the influence of each factor and the interaction between each on the response value. The larger the slope of the response surface, the greater the influence of experimental factors on the response value (Wan et al. 2018). Using Design Expert software, quadratic multiple regression fitting was performed on the experimental data in Table 5, and the response surface diagram and contour diagram of the quadratic regression Equations (6) and (7) were obtained, as shown in Figures 5 and 6.
3D surface plots of effects of binary interactions among factors X1 (biochar dosage), X2 (pH), X3 (temperature) on response value Y1 (removal efficiency of NOR).
3D surface plots of effects of binary interactions among factors X1 (biochar dosage), X2 (pH), X3 (temperature) on response value Y1 (removal efficiency of NOR).
3D surface plots of effects of binary interactions among factors X1 (biochar dosage), X2 (pH), X3 (temperature) on response value Y2 (adsorption capacity of NOR).
3D surface plots of effects of binary interactions among factors X1 (biochar dosage), X2 (pH), X3 (temperature) on response value Y2 (adsorption capacity of NOR).
For the response values Y1 (NOR removal efficiency) and Y2 (NOR adsorption capacity), it can be seen from Figures 5 and 6 that the interaction between X1 (biochar dosage) and X2 (solution pH) is very significant, which is reflected in the steep surface. And the closer the contour shape is to the ellipse, the stronger the interaction (Pryseley et al. 2010). Both Y1 (NOR removal efficiency) and Y2 (NOR adsorption capacity) increased with the increase of X1 (biochar dosage) and the decrease of X2 (solution pH), showing an obvious inverse relationship. This is because when the concentration of norfloxacin remains unchanged, the increase of biochar dosage will provide more active sites that can adsorb norfloxacin, improving the removal efficiency and adsorption capacity of norfloxacin. Norfloxacin has two acid dissociation constants pKa values of 6.22 and 8.51 (Kong et al. 2014). Therefore, when the pH of the solution rose from 3 to 6, the cation of norfloxacin (NOR+) dissolved in the water gradually decreased. The zero point charge (pHpzc) of the grapefruit peel biochar is 9.82 > 6 (Wang et al. 2020). The surface of biochar was positively charged, and the cation exchange and H bond between norfloxacin cation (NOR+) and biochar were weakened, resulting in a great decrease in the adsorption capacity of norfloxacin on the grapefruit peel biochar. The slope of the 3D surface in Figures 5 and 6 is small, indicating that the interaction between X1 (biochar dosage) and X3 (reaction temperature), X2 (solution pH) and X3 (reaction temperature) has no significant effect on the removal efficiency and adsorption capacity of norfloxacin.
CONCLUSION
- (1)
Box-Behnken Design (BBD) was used to explore the effect of biochar dosage, solution pH and reaction temperature on adsorption of norfloxacin by pomelo peel biochar. Two quadratic multiple regression equations with norfloxacin removal efficiency Y1 and norfloxacin adsorption capacity Y2 as response values were constructed. The determination coefficients of the prediction model were 0.9905 and 0.9820, respectively, and the model was significant and reliable.
- (2)
Response surface analysis shows that the influence of each single factor on removal efficiency (Y1) and adsorption capacity (Y2) is X2 (solution pH) > X1 (biochar dosage) > X3 (reaction temperature). In the interaction, the influence of X1X2 (biochar dosage × solution pH) was extremely significant, while that of X1X3 (biochar dosage × reaction temperature) and X2X3 (solution pH× reaction temperature) was not significant.
- (3)
Using Design Expert software to optimize the adsorption of norfloxacin by the grapefruit peel biochar, the optimization scheme was X1 (biochar dosage) = 0.5 g/L, X2 (solution pH) = 3, X3 (reaction temperature) = 45 °C. The Y1 (NOR removal efficiency) and Y2 (NOR adsorption capacity) obtained in the verification test were 75.68% and 3.0272 mg/g, respectively, and the error was only 2.38% and 0.0242 mg/g. The model established by Box-Behnken Design is verified to be authentic, reliable and accurate.
ACKNOWLEDGEMENTS
This work was supported by the National Natural Science Foundation of China (Grant no. 51808001, 51409001) and Natural Science Foundation of Anhui Province (1808085QE146, 1708085QB45, 2008085ME159).
CONFLICT OF INTEREST
The authors have declared no conflict of interest.
DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.