47 By plotting ln KC versus 1/T, ∆H0 and ∆S0 values can be obtained from the slope and intercept, respectively.
The negative value of ∆G0 indicates the spontaneity of the adsorption process, and greater values (in module) reflect a more energetically favourable adsorption reaction [113]. A positive value for ∆G0 indicates that energy must be input, and that the reaction is nonspontaneous. Enthalpy ∆H0 is used to identify the nature of adsorption, in which a positive value of ∆H0 indicates an endothermic process and negative ∆H0 indicates the adsorption is exothermic. A positive value of ∆S0 indicates increased randomness of adsorbate molecules on the solid surface than in solution [108]. If ∆S0 is positive, along with negative ∆G0, the adsorption is spontaneous and is reaching equilibrium [114].
Determining the best model: Error analysis
Nonlinear optimization techniques require an error function to be defined to evaluate the fit of various isotherm or kinetic equations to the experimental data. Eight different error functions, including the sum of the squares of the errors (SSE), the sum of the absolute errors (SAE), the mean absolute deviation (MAD), the average relative error (ARE), the mean square error (MSE), Root mean square error (RMSE), Chi-square (χ2) and the coefficient of determination (R2), were employed in this study to find the most suitable isotherm and kinetic models to represent the experimental data [109, 115, 116]. In each cases the parameters were determined by minimizing the respective error functions using the Solver add-in Excel. The calculated expressions of the error functions used are included from equations from 9 to 16 in Table 4-4.
48 adsorbed on biochar surface showed a higher dependence on pH solution than anionic dyes MO and BG. An increase in the pH value of the dye solution resulted in an increase in the removal amount of the cationic dyes MB and SO, whereas an opposite trend was observed for anionic dyes MO and BG.
For example, the amount of MB adsorbed of VRS700, JRS700, VRH700 and JRH700 generally increased from 20.57 to 26.54 mg/g, from 19.32 to 25.80 mg/g, from 10.74 to 20 mg/g, from 10.16 to 19.53 mg/g, respectively. In contrast, the amount of MO adsorbed VRS700, JRS700, VRH700 and JRH700 slightly decreased from 21.41 to 17.69 mg/g, from 19.79 to 16.13 mg/g, from 12.83 to 8.78 mg/g and from 11.43 to 8.66 mg/g, respectively. Adsorption for SO and BG dyes were the same trends but slightly smaller values than MB and MO, respectively. This behavior was explained by the fact that at high pH values pH > pHpzc, the negatively charged sites of all biochars (−B−O¯) increases because of deprotonation of surface. The electrostatic attraction force between negatively-charged biochar surface and positively-charged dye MB+ and SO+ molecules is therefore enhanced:
−B−OH + OH¯ −B−O¯ + H2O
−B−O¯ + MB+ −B−O¯ ··· MB+
−B−O¯ + SO+ −B−O¯ ··· SO+
Electrostatic attraction, particularly, between the oxygenated surface functional groups (i.e., –OH and –COOH) of biochars and N+ of MB and SO may govern the adsorption process. At very low pH values, MB and SO dye removal were also obtained, although in lower percentages, confirms the existence of other physical interactions, such as hydrogen bonding, porous diffusion, π-π or π+-π interaction. In addition, biochar surface is always negative at very low pH due to a
considerable excess in the number of negatively charged groups compared to positively charged groups [118 – 120]. Since the surface of the biochars was strongly negatively-charged at pH > pHpzc, the electrostatic repulsion occurs between the negatively-charged dye MO and BG molecules and the surface of the biochar, resulting the lower amount of anionic dyes to be adsorbed by biochars.
In this study, the maximum value of MB and SO dye removal appeared when the pH > 7, whereas MO and BG appeared when the pH < 3. However, an increasing pH solution from 7 to 10, the amount of MB and SO adsorbed kept basically unchangeable. As a result, for all subsequent studies, pH ~ 7 and pH ~ 2 were selected for cationic and anionic dye experiments, respectively.
4.3.2 Effect of biochar dosage
The biochar dosage is particularly important because it determines the extent of dye removal and gives an idea of the effectiveness of a biochar and the ability of a dye to be adsorbed with a minimum dosage [121]. In this study, the effect of biochar dosage was performed in 10 mL of dye solution of approximately 50 mg/L, where different dosages of biochar were added to achieve dye-biochar ratios between 1 to 5 g/L. Fig. 4-3 shows the percentage of dye removal increased with an increase in
49 biochar dosage (see the solid lines with vertical axis of the left side figure), but, at the same time, the amount of dye adsorbed decreased generally (the dashed lines with vertical axis of the right side figure).
With increasing the biochar dosage, the number of available adsorption sites increases, leading to the percent removal increases. However, the high biochar dose can attribute to unsaturation or
aggregation of adsorption sites. Such aggregation might result in a decrease in total biochar surface area available to the dyes and an increase in diffusion path length [122]. As a result, the amount of dye (mg) absorbed per gram of biochar decreases with increasing biochar dosage. Fig. 4-3 shows that VRS700 maintained a higher level of adsorption than other biochars for different amount. For examples, the percent removal of dyes by VRS700 at various dosage of biochar (0.01 – 0.05 g) showed that the percentage of MB, SO, MO, and BG removal increased from 95 to 98%, from 86 to 96%, from 78 to 97%, and from 71 to 90%, respectively. As expected, the amount of MB, SO, MO and BG adsorbed decreased from 46 to 10 mg/g, from 41 to 9.9 mg/g, from 38 to 9.8 mg/g, from 33 to 9.7 mg/g, respectively. Similar trends were observed in cases of JRS700, but at the slightly lower adsorbed values. For rice husk biochars, the amount absorbed per unit mass also decreased, but there is only a little difference between VRH700 and JRH700, and their adsorbed values were much lower than VRS700. It is noted that when the biochar dosage was increased from 0.01 to 0.02 g, the percentage of MB, SO, MO, and BG removal increased rapidly, however, above 0.02 g of biochars, the percentage of dye removal kept basically unchangeable (slightly varied within 2%). Therefore, the remaining experiments were conducted for a biochar dosage of 2 g/L.
4.3.3 Effect of pyrolysis temperature
Biochars produced at 300, 500 and 700 °C were chosen for evaluation of the effect of pyrolysis temperature on the adsorption capacity of MB, SO, MO and BG. Specific surface area and pore parameters of these biochars were described in our previous paper [101] and were summarized in Table SM4-1. Briefly, the values of surface area increased from 300 to 700 °C, where Vietnamese rice variety generally showed higher values than Japanese rice variety (VRS>JRS, VRH>JRH).
VRS700 and JRS700 obtained the highest BET surface area of 378 and 293 m2/g, respectively, followed by VRH700 and JRH700 with 245 and 236 m2/g, respectively.
The adsorption results in Fig. 4-4 shows that the removal of dyes increased significantly with the higher pyrolysis temperature biochars. In other words, biochars produced at 700 °C obtained the higher percent removal compared to those produced at 300 and 500 °C. In addition, all biochars exhibited the higher cationic dye removal efficiency, i.e. MB and SO, compared with the anionic dyes, i.e. MO and BG. This behavior may be attributed to the surface characteristics of biochars and the ionic charges of the dye molecules. Biochar surfaces are usually negatively charged, which mean that positively charged cationic dyes would be absorbed via electrostatic attraction, as confirmed on
50 the effect of pH. The binding of anionic dyes to the biochar surface was primarily by physical forces [100]. It is observed that the enrichment of specific surface area by high pyrolysis temperature at 700 °C led to the rising of biochar removal efficiency. In other words, the higher surface area of biochars produced at 700 °C was believed to provide more adsorptive sites for dye molecules, therefore, they were chosen to investigate the adsorption isotherms and kinetics of the studied dyes.
4.3.4 Effect of contact time and adsorption kinetics
The effect of contact time (1-720 min) on the qt values at 25 °C is shown in Fig. 4-5. It may be seen that the MB, SO, MO and BG dyes were rapidly adsorbed in the first 1–20 min, and then the
adsorption rate decreased gradually and reached equilibrium in about 240 min. At the beginning, the uptake rate for all the dyes is very high as many adsorption sites are available for adsorption at the onset of the process. As the sites are gradually filled up, adsorption become slow due to dye aggregation at the surface. The aggregation of dye molecules with increasing contact time makes it almost impossible for the dye molecules to diffuse deeper into the biochar structure at the highest energy sites. Another explanation, the dyes are quickly adsorbed on the exterior surface, but, when the exterior surface is saturated, the dye enters into the pores of the biochar particles and is adsorbed on the interior surface [123]. These may be the reasons why rate of adsorption became slower at higher contact time. The percentage removal at 720 min contact time were found to be higher by a maximum of ~3% than those obtained after 240 min contact time. For this reason, the optimum contact time was chosen as 240 min for adsorption in this study.
To understand mechanisms of dye uptake by biochars, the kinetic experimental data was analysed using three kinetic models, namely, the Pseudo-second-order, Elovich and Intra-particle diffusion (Table 4-2). The plots of nonlinear forms of the three models obtained at 25 °C are shown in Fig. 4-5 and their nonlinear kinetic parameters are tabulated in Table SM4-2. The linear kinetic parameters of the Pseudo-second-order and Elovich are shown in Table SM4-3. The values of eight non-linear error functions in Table SM4-4 show that SSE, SAE, MAD, ARE, MSE, RMSE and χ2 values in pseudo-second-order kinetic model were smallest, and its linear regression coefficient R2 value closed to unity for all dye cases. Moreover, the calculated equilibrium adsorption capacities (qe,cal) of Pseudo-second-order were in close agreement with the experimental values (qe,exp). These results suggest that the adsorption data was well presented by pseudo-second-order kinetic, implying that the adsorption process was controlled by chemical surface interaction via electrostatic attraction, as confirmed previously in the effect of solution pH.
Although the experimental data was satisfactorily described by the pseudo-second-order, this kinetic model cannot predict or confirm any particular adsorption mechanism for adsorption. It is well known that the adsorption process of the dye molecules onto a porous biochar is mainly controlled by a multi-step process, therefore, the intra-particle diffusion model was studied. Fig. 4-6 presents the plots
51 of the linearized form of the intra-particle diffusion model for all the biochars and its parametric values are given in Table 4-5.
In Fig. 4-6, the plots of qt versus t1/2 for the intra-particle diffusion models of MB, SO, MO and BG adsorption onto biochars showed three similar interdependent linear lines, indicating that there were three stages take place during adsorption process [124]. External diffusion or liquid film diffusion occurred in the first sharper linear stage (stage I: t = 1 – 10 min) where the dye uptake was quite rapid.
This stage involved the transport of dye molecules from the bulk liquid phase to the external surface of biochars through a liquid boundary layer. This was followed by an intermediate stage (stage II: t = 20 – 120 min) that showed a slower uptake, which was due to intra-particle pore diffusion of dye molecules from the exterior of the biochars into macropores, mesopores and micropores of biochars.
Until the final plateau was reached (stage III: t >120 min) due to adsorption equilibrium. The final stage was the intra-particle diffusion starting to slow down due to low dye concentration in solution phase as well as less available biochar adsorptive sites. The parameters and correlation
coefficients obtained for each stage is provided in Table 4-5. The results showed that the regression curves were straight lines with relative high correlation coefficient values (R2 = 0.74 – 0.99), implying that porous diffusion mechanism would have a significant effect on the dye-biochar adsorption process. It could also be seen that, for all dyes, the rate constant for the first stage (kp1) was higher by 4 to 12 times than the rate constant of second phase (kp2) and higher by 17 to 232 times than the final stage (kp3), denoting that the diffusion resistance of the boundary layer was much smaller than the diffusion resistance of the pore diffusion steps [125]. The calculated values of C1, C2, C3 were all non-zero which confirmed that diffusion into the biochar pores was not the only rate-controlling step.
Therefore, the dye-biochar adsorption proceeded via a complex mechanism consisting of both surface adsorption and intra-particle transport within the pores of biochars.
4.3.5 Effects of initial dye concentration and Adsorption isotherms
The effect of initial dye concentration in the adsorption of MB, SO, MO and BG onto biochar was evaluated at various initial dye concentration, varied from 10 to 200 mg/L and is presented in Fig. 4-7.
When the dosage of biochar was kept unchanged (∼ 0.02 g), the amount of the dye adsorbed per unit mass biochars (qe) increased with an increase in the initial dye concentration. At low concentrations there were unoccupied adsorptive sites on the biochar surface, offering a greater chance for dye adsorption. When the dye concentration in solution was increased, the adsorption sites available in the biochars became more quickly saturated, thereby reducing the efficiency of dye adsorption capacity.
The nonlinear fitting of Langmuir, Freundlich, and Temkin isotherms is also presented in Fig. 4-7, and their parameters are summarized in Table SM4-5 (for nonlinear forms) and SM4-6 (for linear forms). An inspection of error functions inTable SM4-7 showed that the SSE, SAE, MAD, ARE, MSE, RMSE and χ2 values for Langmuir model were smallest, together with the highest R2 values
52 obtained from both nonlinearized and linearized models. Additionally, the calculated data from this model was quite similar to the experimental data. These results confirmed that Langmuir model appropriately described the isotherms of the dye adsorption process, which indicated monolayer adsorption on a finite number of identical sites of equal energies of adsorption.
The theoretical monolayer maximum adsorption capacities (qm) for MB calculated from the nonlinear Langmuir equation at 25 °C were ordered as follow: VRS700 (67.69 mg/g) > JRS700 (56.88 mg/g) >
VRH700 (33.28 mg/g) > JRH700 (32.81 mg/g); and for SO dye were the same order but slightly smaller than MB, with VRS700 (60.15 mg/g) > JRS700 (52.99 mg/g) > VRH700 (32.36 mg/g) >
JRH700 (30.16 mg/g) (Table SM4-5). The adsorption of anionic dyes followed the same order of cationic dyes, however, the Langmuir adsorption capacity values were much lower than for cationic dyes. For example, MO adsorption capacities were ordered as follow: VRS700 (48.31 mg/g) >
JRS700 (40.44 mg/g) > VRH700 (18.58 mg/g) > JRH700 (15.20 mg/g), and BG adsorption capacities also followed the same order but slightly smaller than MO, with VRS700 (45.47 mg/g) > JRS700 (37.79 mg/g) > VRH700 (18.08 mg/g) > JRH700 (14.78 mg/g). These results confirm that biochars from Vietnamese IR50404 showed higher dye adsorption capacity than Japanese Koshihikari variety and was significantly more effective for cationic dyes, as already found in the effect of pyrolysis temperature. Notably, the adsorption capacity of rice straw biochars (VRS700 & JRS700)
to remove dyes was almost double that shown by rice husk biochars (VRH700 & JRH700). Reduced adsorption in rice husk biochars (VRH700 and JRH700) was probably due to the inability of the dye molecules to penetrate all the internal pore structures and less available adsorption sites.
In addition, adsorption of MB was higher than SO and MO was higher than BG. It was found that the difference in the absorbed amount was attributed to the difference in the molecular size of dyes [126], and this may apply in this study. As listed in Table 4-1, molecular size of MB (319.85) is smaller than SO (350.84), MO (327.3) molecular size is only about one-half BG size (698). In other words, small molecules MB and MO could gain easier access to the internal pore network of the biochar surface.
The dimensionless constant separation factor (RL) values obtained from Langmuir equation were between 0.04 and 0.92, together with the nonlinearity index (1/n) from Freundlich equation was between 0.36 and 0.53 at 25 °C (data not shown), indicating the favourability of the dyes adsorption onto biochars under studied conditions. Therefore, the selected biochars were suitable adsorbents for MB, SO, MO and BG from aqueous solution.
4.3.6 Effect of temperature and adsorption thermodynamics
The study on the effect of temperature helps to reveal whether adsorption is exothermic (decrease of adsorption capacity with increasing temperature) or endothermic (increase of adsorption capacity with increasing temperature). A temperature range of 25 – 45 °C (298 – 318K) was used to evaluate the
53 effect of temperature on dye adsorption. As illustrated in Figs. from SM4-1 to SM4-4, the amount of MB, SO, MO and BG adsorbed on biochars was found to increase slightly with increasing
temperature, thereby indicating the endothermic nature of the ongoing adsorption. It is explained that increasing temperature produces a swelling effect within the internal structure of the biochar, which enables dye molecules penetrate further into smaller pores of biochars. Also, as the temperature increases, the rate of diffusion of the dye molecules across the external boundary layer and into the internal pores of the biochar particles is increased as a result of the reduced viscosity of the solution, resulting in higher adsorption capacities of biochars [127].
The nonlinear Langmuir plots for the adsorption of MB, SO, MO and BG onto biochars at
temperatures 25, 35 and 45 °C are presented in Figs. SM4-1, SM4-2, SM4-3 and SM4-4, respectively.
It was observed that in all the temperature cases of the Langmuir isotherm model, the obtained correlation coefficients were close to unity and the calculated equilibrium adsorption capacity values were very close to the values obtained experimentally (data not shown), indicating verification of the Langmuir adsorption model and involvement of monolayer adsorption in each case. Since the
adsorption isotherm data fitted well to the Langmuir model, KL constant obtained from both linear and nonlinear optimization technique was used to calculate thermodynamic parameters, including Gibbs free energy change ∆G0, change in enthalpy ∆H0 and change in entropy ∆S0. As shown in Table SM8, the negative ∆G0 values were obtained in all studied dyes, indicating that the adsorption process of cationic and anionic dyes was both thermodynamically favorable and spontaneous under experimental conditions [113]. It could be observed that when the temperature increased from 298 to 318 K,
the ΔG0 values decreased by 7.3 to 10.9 % for MB, 7.2 to 10.2 % for SO, 7.9 to 11.1 % for MO and 7.6 to 9.7 % for BG adsorption, respectively. This denotes the increase in spontaneity of all adsorption systems, leading to the higher removal efficiency of these dyes at higher temperatures. Furthermore, the ΔG0 values calculated from MB adsorption at 298K were correspondingly lower by as much as 2.66, 3.69 and 2.72 kJ/mol with respect to the ΔG0 values from SO, MO and BG adsorption,
respectively. This confirms that the adsorption of MB on biochars was more spontaneous and favored than that of SO, MO and BG dyes, which could explain the higher adsorption capacities obtained from the isotherm study of MB adsorption.
The positive values for ∆H0 indicates the endothermic nature of the adsorption process of dyes onto the studied biochars. The endothermicity of the adsorption process requires the absorption of heat from its surroundings during the adsorptive process. In general, physisorption is considered when the ∆H0 value is less than 40 kJ/mol, while chemisorption is assumed to take place between 50 and 200 kJ/mol [108]. In this study, the low ΔH0 values obtained for all dye adsorption (ΔH0 = 2.51 - 23.70 kJ/mol), implying that physical interactions have a significant effect on the adsorption rate. The ΔS0 was also positive, corresponding to an increased randomness on the biochar surface during the
54 uptake of dye molecules. In conclusion, values of ΔG0 and ΔH0 confirm the spontaneous and
endothermic nature of the adsorption process, which was consistent with the results of several studies [128, 129].
4.3.7 Possible adsorption mechanism
The low enthalpy values obtained from thermodynamic study is an indication that physisorption was the main mechanismof adsorption for both anionic dyes and cationic dyes. Physical adsorption mechanism may occur via porous diffusion, hydrogen bonding, π-π interaction and π+-π interaction between the dye molecules and biochars. These physical mechanisms are proposed in Fig. 4-8.
Porous diffusion
The porous diffusion mechanism of MB, SO, MO and BG dye molecules was suggested to consist of two distinct adsorption steps. The first step may be the external diffusion of the dye molecules from the boundary layer to the surface of the biochar. This may follow by the internal diffusion in which the penetration of the dye molecules into the pores of the biochar occurs.
Hydrogen bonding
Generally, hydrogen bonding is favoured on material with oxygen-containing groups and in particular hydroxyl functionalities as in the case of the biochar material. Hydrogen bonding interaction may occur between the H-bonds (i.e., −COOH or −OH which act as the H-donors) on the biochar surface and the nitrogen or oxygen atoms in dyes (i.e., nitrogen atoms in the MB and SO molecules, or the oxygen atoms in the MO and BG molecules, which act as the H-acceptors).
π-π interaction and π+-π interaction
π-π interaction is a noncovalent interaction between π-acceptor and π-donor molecules, while π+-π interaction (or cation-π interaction) is a noncovalent interaction between a surface of an aromatic π-donor and cation. In π-π interaction, the aromatic rings in biochar may act as π-electron π-donors and the aromatic rings in MB, SO, MO and BG dyes may act as π-electron acceptors. Also, molecular structures of MB, SO and MO possess a cation N+, which may lead to the binding of cation N+ to the π-face of aromatic rings in biochar surface, forming π+-π interaction.
The experimental results of pH effects show that adsorption of cationic dyes MB and MO was favoured as pH increases. However, the adsorption of anionic dyes was not favoured at high pH values because the electrostatic repulsion and the presence of OH- ions in excess competing with the dye anions for the adsorption sites. Therefore, besides physical interaction as found, another major mode of adsorption of cationic dyes could be chemisorption. Chemical adsorption mechanism of MB and SO on biochars may likely be due to an electrostatic attraction between the oxygenated surface
55 functional groups (i.e., –OH and –COOH) of biochars and N+ of MB and SO molecules. This
electrostatic attraction is illustrated in Fig. 4-9.