Application of magnetically modified sewage sludge ash (SSA) in ionic dye adsorption

Abstract

Incineration is a traditional method of treating sewage sludge and the disposal of derived ash is a problem of secondary waste treatment. In this study, sewage sludge ash (SSA) was coated with ferrite through a ferrite process and then used as an adsorbent for ionic dyes (methylene blue [MB] and Procion Red MX-5B [PR]). The modified SSA possessed surface potential that provided electrostatic attraction toward MB and PR. Adsorbent FA10 (named on the basis of being produced from 10 g of SSA in the ferrite process) was used for the adsorption of MB. Ideal pH for adsorption was 9.0 and maximum adsorption capacity based on Langmuir isotherm equation was 22.03 mg/g. Adsorbent FA2.5 (named on the basis of being produced from 2.5 g of SSA in the ferrite process) was used for PR adsorption. Ideal pH for adsorption was 3.0 and the maximum adsorption capacity (calculated as above) was 28.82 mg/g. Kinetic results reveal that both MB and PR adsorption fit the pseudo-second-order kinetic model better than the pseudo-first-order model. The values of activation energy calculated from rate constants were 61.71 and 9.07 kJ/mol for MB and PR, respectively.

Implications:

Magnetic modified adsorbent could be synthesized from sewage sludge ash (SSA). In this study, the adsorption ability of SSA toward ionic dye (methylene blue [MB] and Procion Red MX-5B [PR]) was enhanced by ferrite process. The synthesized Fe₃O₄ can act as an active site and provide electrostatic attraction toward cationic dye and anionic dye at different pH. The application of magnetic modified adsorbent in wastewater treatment can not only recycle the SSA, but also make SSA become an environmentally friendly material.

Introduction

Incineration is a traditional method of sewage sludge disposal. It possesses several advantages, including the reduction of sludge volume, thermal destruction of toxic organic constituents, and energy recovery during sludge combustion (Werther and Ogada, 1999; Wang et al., 2012). The major solid product of sewage sludge incineration is sewage sludge ash (SSA). Although SSA is an inorganic and stable material, it still needs final disposal. Therefore, SSA recycling techniques have been developed. One use of incinerated ash has been as a base material for the production of lightweight aggregates used in construction (Cheeseman and Virdi, 2005; Hu et al., 2012). Other uses include its use as the raw material for cement (Lin and Lin, 2005; Garcés et al., 2008), bricks (Anderson, 2002; Anderson et al., 2002), and tiles (Chen and Lin, 2009; Kikuchi, 1998). On the other hand, SSA can be recycled as fertilizer to increase the nutrient level of agricultural soils (Weigand et al., 2013; Franz, 2008).

Besides the above applications, SSA can be utilized as an adsorbent in wastewater treatment. It is a win-win strategy in that it uses SSA as an inexpensive, environmentally friendly material for turning waste into useful material (Moghaddam et al., 2009). The adsorbates of SSA adsorbent could be heavy metal ions or organic dyes. Liu et al. (2011a, 2011b) and Elouear et al. (2009) utilized sewage sludge incineration slag as an adsorbent for removing Cu, Ni, or Cd from wastewater; the adsorption pH was either weak acid or neutral. Weng (2002) studied the adsorption characteristics of an anionic dye (New Coccine) by SSA. The experimental data correlated well with a nonlinear multilayer adsorption isotherm. The adsorption capacities ranged between 1.96 and 3.45 mg/g, with pH being the key factor affecting adsorption. Dye adsorption density increased abruptly below pH 5.0; this phenomenon was attributed to SSA having a positively charged surface when the solution pH was less than the pH zero point of charge (pH_zpc). Below this pH, the surface acidity of the hydrolyzed ash and its positively charged surface provided favorable electrostatic conditions for the single anionic dye species to be adsorbed. Moghaddam et al. (2012) studied the adsorption of Acid Red 119 by mixtures of dried sewage sludge (DSS) and SSA. Results revealed that a DSS mixture of 40 wt%, 5 g/L of adsorbent dose, 200 mg/L of initial dye concentration, initial pH 6, and 60 min of contact time were the optimal experimental conditions for giving a removal efficiency of higher than 95%.

The above studies support the potential of SSA as a dye adsorbent; however, there has not been abundant research into its adsorption application. According to other reports (Weng and Pan, 2006; Weng, 2002), the adsorption capacity of SSA is 1.6 mg/g for MB and 1.96–3.45 mg/g for New Coccine dye. This study examines a process by which these results can be improved. A ferrite synthesis process is proposed and used to produce crystalline Fe₃O₄ on the surface of SSA particles. Ferrite is a spinel structure that crystallizes in a cubic close-packed arrangement (Mathew and Juang, 2007). The chemical mechanism for the formation of Fe₃O₄ is listed in eq 1 (Zhu et al., 2007):

(1)

The surface potential of Fe₃O₄ can change with pH adjustment (Barale et al., 2008); it acts as an adsorption site for cationic or anionic dye in aqueous solution. Furthermore, the magnetic adsorbent can be easily separated by magnetic field, which provides a more convenient recovery method than conventional separation techniques. The adsorbates of this experiment were methylene blue (MB) and Procion Red MX-5B (PR). Langmuir isotherm and Freundlich isotherm are adopted to evaluate the adsorption isotherm, and adsorption kinetics are studied using pseudo-first-order and pseudo-second-order kinetic models. Finally, the activation energy is obtained from the kinetic rate constant and Arrhenius equation.

Materials and Methods

Characteristics analysis of sewage sludge

The sewage sludge from a wastewater treatment plant in northern Taiwan was incinerated at 900 °C for 1 hr and the residue was named as sewage sludge ash (SSA). The SSA was ground and sieved with a 0.59-mm sieve before adsorption experiments. Chemical analysis, X-ray diffraction (XRD) analysis, and ignition-loss testing were utilized to analyze the characteristics of SSA. For chemical analysis, 0.1 g SSA powder was mixed with 16 mL of aqua regia and 8 mL of hydrofluoric acid at 180 °C for 3 hr. After cooling and filtration, the filtrate was diluted and the concentrations of Si, Al, Fe, Na, K, Ca, and Mg were analyzed by atomic absorption spectroscopy (Varian AA240; CA, USA). X-ray diffraction (XRD) analysis was conducted with an X-ray diffractometer (Rigaku D/max-2500, Tokyo, Japan). It was scanned from 10° to 80° (2θ) at a scan rate of 4° per minute with voltage of 30 KV and current of 50 mA. For ignition-loss testing, 5 g (W ₁) of SSA were heated in the furnace at 800 °C 3 hr. The heating rate was fixed at 15 °C per minute. The weight (W ₂) of the residue was recorded while the sample was cooled. The results of ignition loss were calculated according to eq 2:

(2)

Synthesis and characteristics of modified magnetic adsorbent

Different dosages (0, 2.5, 5, 10 g) of SSA were mixed with 250 mL of solution containing 0.2 M FeSO₄ (FeSO₄·7H₂O, 99%; Panreac, xx, xx) and 0.2 M FeCl₃ (FeCl₃·6H₂O, 99%; Panreac). The mixture was agitated and its pH adjusted to 11 with 10 N NaOH. Then, the mixture was heated to 90 °C for 2 hr. After the ferrite process, the mixture was cooled, filtrated, washed with deionized (DI) water twice, dried, and ground to powder. The magnetic adsorbents produced from 0, 2.5, 5, and 10 g of SSA were named FA0, FA2.5, FA5, and FA10.

XRD, magnetic properties, specific surface area, and zeta potential analyses were conducted to understand the characteristics of adsorbents. The XRD analysis procedure was the same as that for SSA analysis. Magnetization testing was conducted at room temperature in a vibrating sample magnetometer (MPMS SQUID VSM; Quantum Design Inc, CA, USA). The magnetic field was set from 20,000 to −20,000 Oe. The specific surface area was measured from N₂ adsorption isotherms of 77 K with adsorption apparatus (Autosorb-1; Quantachrome, FL, USA) and calculated with the Brunauer-Emmett-Teller (BET) equation. Zeta potential of the magnetic adsorbents was determined by a zeta potential analyzer (Nano Z; Malvern, Worcs, UK), and 0.1 g of adsorbent was mixed with 1000 mL of 0.001 M NaCl solution. The mixture was dispersed in an ultrasonic cleaner for 30 min and 1 N HCl or NaOH solution was used to adjust the pH within a range of 2–10 for measurement.

The effect of pH on dye adsorption

MB and PR were chosen as adsorbates, and the adsorption of dyes was studied in batch operation mode and the concentration of residual dye in solution was determined by ultraviolet/visible (UV/Vis) spectrophotometry (Lambda 25; PerkinElmer, Taiwan, China) at wavelengths of 664 nm for MB and 538 nm for PR. SSA and FA5 were chosen as adsorbents and the ratio of adsorbent versus dye solution was fixed at 0.5 g/50 mL. The initial concentration of dye solution was 100 mg/L. Adsorbents and dye solution were added to the Erlenmeyer flask, which was placed into the water bath equipment (BT-150D; Yihder, Taiwan, China) and the shake speed was fixed at 100 rpm. NaOH (1 N) and HCl (1 N) was used to adjust the pH value of the dye solution, and adsorption experiments were conducted from pH 2 to 10 at 30 °C for 6 hr. After the adsorption experiments, the solutions were filtered with Advantec 5C filter paper (Tokyo, Japan) by gravitational filtration and the concentrations of MB and PR determined with UV-Vis spectrophotometry. The adsorption percentage was obtained from eq 3:

(3)

where C ₀ is the initial concentration of dye solution, and C is the dye concentration of the filtrate.

Isotherm adsorption

The equilibrium time of adsorption was determined before isotherm adsorption testing. The experimental conditions were 200 mg/L of initial dye concentration, 0.5 g/50 mL of adsorbent versus dye solution at 30 °C for 0.5–16 hr. The pH value for MB adsorption was 9.0 and that for PR was 3.0. Isotherm adsorption experiments were conducted for different initial concentrations (100, 200, 400, 600, and 800 mg/L) of MB and PR. The ratio of adsorbent versus dye solution and adsorption temperature were as above. Adsorption duration was 4 hr. After adsorption equilibrium of 4 hr, the experimental results were examined using Langmuir ( eq 4) and Freundlich ( eq 5) isotherm models, as follows (Gobi et al., 2011):

(4)

(5)

where C _e is the dye concentration in solution at equilibrium; Q _e is the adsorption capacity of the adsorbent at equilibrium; Q _m is the maximum capacity of the adsorbent (mg/g); K _L is the Langmuir adsorption constant (L/mg); K _F is the Freundlich constant (L/mg); and n is the heterogeneity factor.

Dimensionless separation factor, R _L, was calculated from the Langmuir isotherm equation to determine the type of adsorption process (Tseng, 2007):

(6)

where C _i is the initial concentration of dye and R _L is classified as R _L > 1, unfavorable; 0 < R _L < 1, favorable; R _L= 0, irreversible adsorption.

Adsorption kinetics

The results of the kinetic study were obtained for different temperatures (30, 45, and 60 °C) and durations (20, 40, 60, and 80 min). Other experimental conditions were the same as adsorption isotherm testing and the initial concentration of dye was 400 mg/L. The results were analyzed with two models, namely, the pseudo-first-order ( eq 7) and pseudo-second-order kinetic models ( eq 8) (Lin et al., 2011):

(7)

(8)

where Q _e (mg/g) is the adsorption capacity at equilibrium; Q_t (mg/g) is the adsorption uptake at time t (min); and k ₁ and k ₂ are the rate constants of the pseudo-first-order and pseudo-second-order models, respectively.

The temperature dependence of rate constants can be described by the Arrhenius equation:

(9)

where A (g/mg/min) is the frequency factor; E _a (kJ/mol) is activation energy; R (J/mol/K) is the gas constant; and T (K) is absolute temperature.

Desorption experiment

Desorption experiments were conducted to illustrate the adsorption process. 0.1 N of HCl and NaOH were used as desorption reagent for FA10 and FA2.5, respectively. The adsorption experiments were conducted using 200 mg/L of initial dye concentration, 0.5 g/50 mL of adsorbent dosage, pH 3 for PR, pH 9 for MB, and a shake speed of 100 rpm at 30 °C for 4 hr. After adsorption, the dye-loaded adsorbent was dried and mixed with 0.1 N HCl or NaOH solution and the desorption conditions were 0.5 g/50 mL of adsorbent and 100 rpm shake speed at 30 °C for 10 min to 4 hr. After desorption, the concentrations of MB and PR in the filtrate were determined with UV-Vis spectrophotometry. The desorption efficiency (DE) was calculated from eq 10:

(10)

where C (mg/L) is the concentration of MB and PR of the desorption solution; V (mL) is the volume of the desorption solution; Q (mg/g) is the adsorption capacity of adsorbent before desorption; and m (g) is the amount of the adsorbent used in the desorption experiments.

Results and Discussion

Characteristics analysis of sewage sludge ash

SSA is a residue of sewage sludge incineration. Its major components were inorganic compounds. The results of chemical analysis and ignition-loss testing are shown in Table 1. The value of ignition loss was 0.1%, which indicates organic compounds having been burnt off during initial incineration. SiO₂ constituted 44.6% of SSA. XRD analysis is shown in Figure 1. It shows that the major crystal phase of SSA was SiO₂, which corresponds to the results of chemical analysis.

Table 1. Results of chemical analysis and ignition loss of SSA

	Chemical Composition (%)						Ignition Loss
Sample	SiO2	Al2O3	Fe2O3	Na2O	K2O	CaO	(%)
SSA	44.6	16.0	7.3	6.2	6.11	4.5	0.1

Figure 1. X-ray diffraction analysis of sewage sludge and magnetically modified adsorbent. FA0 = SSA not subject to the ferrite process.

Formation and characteristics of modified magnetic adsorbent

Modified SSA was synthesized in the ferrite process. The ferric and ferrous ions were reacted at pH 11 for 2 hr at 90 °C. Fe₃O₄ coated the surface of SSA particulates. XRD analysis of magnetically modified SSA is shown in Figure 1. It is obvious from the diffraction patterns that Fe₃O₄ and SiO₂ are the major phases of these samples. Magnetization of magnetically modified adsorbent is shown in Figure 2. The results show that the value of saturated magnetizations decreased with the increased SSA dosage. The magnetization values of FA0, FA2.5, FA5, and FA10 were 73.7, 53.6, 45.8, and 33.0 electromagnetic units (EMU)/g, respectively. It is thought that the higher SSA dosages led to lower relative coverage by Fe₃O₄ on the surface of the adsorbent, leading to a decrease in saturated magnetization.

Figure 2. Magnetization versus applied magnetic field.

Effect of pH on dye adsorption

pH is an important factor in ionic dye adsorption because the surface charge of adsorbents varies with pH value. The adsorption results of SSA, FA5, and ferrite from pH 2 to pH 12 are shown in Figure 3a and b.The experimental conditions were 100 mg/L of initial dye concentration and 0.5 g/50 mL of adsorbent at 30 °C for 6 hr. For MB, adsorption percentages by SSA and FA5 increased with increased pH. Pan et al. (2003) studied copper ion adsorption by SSA and mentioned that silicon or aluminum oxide may be hydrolyzed. The chemical reaction listed below may occur under different pH:

Figure 3. Adsorption of ionic dyes at different pH values for different adsorbents: (a) MB and (b) PR. (Initial dye concentration: 100 mg/L.)

(11)

Methylene blue is a heterocyclic aromatic compound with a dimethylazanium group. It represents a single positive charge when it is dissolved in water. Under alkaline conditions, the hydrolyzed silicon tends to become ≡Si–O⁻ (a negatively charged group). At the same time, Fe₃O₄ also represents a negative charge under alkali conditions (Wang et al., 2010) so that both hydrolyzed silicon and Fe₃O₄ can adsorb MB by electrostatic attraction. Therefore, the MB adsorption ability of FA and ferrite are higher than that of SSA (the adsorption percentage of FA5 was higher than 98% when pH was higher than 9.0 but that of SSA was lower than 65%). For PR adsorption, it was found that the adsorption percentage of PR in the three samples decreased with increased pH, and that of SSA was lower than 10% in the tested pH range. Although (eq 11) showed that ε Si-OH₂ ⁺ can form under acid conditions, the adsorption percentages of SSA show that the electrostatic attraction between SSA and PR is very weak when pH is less than 6. On the other hand, ferrite possesses a higher adsorption ability than FA5 because of electrostatic repulsion provided by the negative charge of SSA in FA5. Based on this result, it is thought that the positive surface charge of Fe₃O₄ under acid conditions provides the major source of electrostatic attraction toward PR, enhancing the adsorption ability of modified SSA toward the anionic dye. Figure 3 shows the adsorption percentage of MB to be 98.4% at pH 9.0 and that of PR to be 89.1% at pH 3.0. Therefore, suitable pH values for MB and PR adsorption are set at pH 9.0 and pH 3.0, respectively.

Effect of SSA dosage on dye adsorption

Fe₃O₄ acts as an active site for dye adsorption; therefore, SSA dosing may affect the adsorption ability of the adsorbent because it affects Fe₃O₄ content. The adsorption abilities of FA2.5, FA5, and FA10 toward MB and PR are shown in Figure 4. The experimental conditions were 100 mg/L, initial dye concentration, and 0.5 g/50 mL adsorbent at 30 °C for 6 hr. The pH value of MB adsorption was 9.0 and that of PR was 3.0. For MB, adsorption percentages were close among the adsorbents (98.2% for FA2.5, 98.6% for FA5, and 99.0% for FA10). For PR, adsorption percentage increased with decreased SSA dosing. The respective adsorption percentages were 93.7%, 89.8%, and 80.6% for FA2.5, FA5, and FA10, respectively. Due to electrostatic attraction being an important driving force for ionic dye adsorption, the zeta potentials for FA2.5 and FA10 from pH 2 to pH 10 are shown in Figure 5.The zeta potentials of FA2.5 and FA10 are negative at pH 9 and the adsorption percentages of MB for the two adsorbents are higher than 98%. Under acidic conditions, the zeta potential of FA2.5 and FA10 are positive. The value for FA2.5 is about +40 mV and that for FA10 is about +30 mV at pH 3. It is thought that FA2.5 can provide higher electrostatic attraction toward PR and obtain a higher adsorption percentage than FA10. According to Figure 5, the value of the zeta potential decreased with increased SSA dosing under acidic conditions. Therefore, the adsorbent's PR adsorption ability is enhanced by increasing Fe₃O₄ content (or decreasing SSA content). As a result, FA2.5 and FA10 were chosen as adsorbents for PR and MB, respectively, as they provided the highest adsorption abilities (FA2.5 adsorbed 93.7% of PR and FA10 adsorbed 99.0% of MB). These two adsorbents were studied for their adsorption isotherm and kinetics.

Figure 4. Adsorption of ionic dyes on different adsorbents.

Figure 5. Zeta potentials of SSA, FA2.5, and FA10 from pH 2 to pH 10.

Isotherm adsorption

FA2.5 is chosen as the PR adsorbent and FA10 for MB. The specific surface areas calculated from BET are 61.55 and 44.94 m²/g, respectively. The adsorption percentages at different time are shown in Figure 6.The experiments were conducted at 30 °C, 0.5 g/50 mL of adsorbent, and 200 mg/L of initial dye concentration. The pH of MB adsorption was controlled at 9.0 and that of PR at 3.0. According to Figure 6, it took more than 2 hr to reach adsorption equilibrium. Therefore, the equilibrium time for both MB and PR was chosen to be 4 hr. This time was adopted for isotherm adsorption.

Figure 6. Adsorption percentage versus time for MB and PR.

The results of isotherm adsorption are shown in Figure 7 and Table 2. Based on eqs 5 and 6, the values of Q _m and K _L in Table 2 are derived from the slope and intercept of the linear plots of C _e/Q _e versus C _e, whereas 1/n and K _F are from the slope and intercept of the linear plots of log Q _e versus log C _e. The r ² value of the Langmuir equation is higher than that of Freundlich, which implies the adsorption of MB and PR suitably fit the Langmuir isothermal equation. The dimensionless separation factor R _L of MB and PR in this study is between 0 and 1 when C _i was between 100 and 800 ppm. The R _L value indicates that the adsorption types for MB and PR for FA10 and FA2.5 are favorable.

Figure 7. Results of isotherm adsorption. (a) Fitting of Langmuir equation; (b) fitting of Freundlich equation.

Table 2. Parameters of adsorption isotherm model fitting for MB and PR

Dye	Langmuir			Freundlich
	Q m (mg/g)	K L (L/mg)	r ^2	n	K F	r ^2
MB	22.03	0.13	0.9991	8.81	11.04	0.9392
PR	28.82	0.03	0.9984	3.27	4.42	0.9308

Adsorption kinetics

Adsorption kinetics can be analyzed using pseudo-first-order and pseudo-second-order kinetic models. The equations for pseudo-first-order and pseudo-second-order kinetics are given by eqs 7 and 8, respectively. The experimental conditions were 30, 45, and 60 °C for adsorption temperature, 400 mg/L for initial dye concentration, and 0.5g/50 mL for adsorbent concentration. The pH value for MB adsorption was 9.0 and that of PR was 3.0. The results are shown in Table 3 and the r ² values of pseudo-first-order and pseudo-second-order models are both higher than 95%. Besides the r ² value, the applicability of both kinetic models can be verified through the sum of error squares, SSE(%) (Hameed et al., 2007; Anirudhan and Tharun, 2012). SSE(%) is given by

Table 3. Results of kinetic model fitting for MB and PR

	T	Q e,exp	First Order		Second Order		E a
Dye	(°C)	(mg/g)	k 1 (1/min)	r ^2	k 2 (g/mg/min)	r ^2	(kJ/mol)
MB	30	20.20	0.0355	0.9785	0.00072	0.9981	61.71
	45	19.18	0.0532	0.9803	0.00198	0.9959
	60	21.64	0.0629	0.984	0.00654	0.9998
PR	30	23.20	0.0350	0.9758	0.00239	0.9886	9.07
	45	19.96	0.0520	0.9878	0.00276	0.9991
	60	20.42	0.020	0.9936	0.00331	0.9992
Note: Q e,exp = adsorption capacity from experiments.

(12)

where Q _e,exp is the experimental adsorption capacity at equilibrium; Q _e,cal is the calculated adsorption capacity from kinetic models; and N is the number of data points. For MB adsorption, the SSE(%) of pseudo-first-order and pseudo-second-order kinetics are 5.01 and 0.89. On the other hand, those of PR adsorption are 7.73 and 5.03. The higher value of r ² and lower value of SSE represented the better model fitting. Therefore, the pseudo-second-order kinetic model is a better-fitted model for MB and PR adsorption.

Activation energy E _a can be obtained by the slope of plotting ln k ₂ against 1/T in Arrhenius equation and the values of E _a for MB and PR adsorption are 61.71 and 9.07 kJ/mol. The magnitude of E _a can be considered a parameter for the differentiation between chemical or physical adsorption (Han et al., 2009). The E _a value of physical adsorption is usually less than 4.2 kJ/mol, because the forces involved in physical adsorption are weak. Chemical adsorption involves forces much stronger than those in physical adsorption. The E _a of activated chemical adsorption is between 8.4 and 83.7 kJ/mol but that of nonactivated chemical adsorption is near zero (Aksu, 2002). In this study, the E _a values of PR and MB adsorption range between 8.4 and 83.7 kJ/mol, indicating that the adsorption of PR and MB on magnetically modified SSA adsorbent is possibly a process of chemical adsorption.

Other adsorption studies on MB and PR were gathered to compare the data of isotherm adsorption and kinetic data (see Table 4). The adsorption capacity of FA10 is higher than some inorganic adsorbents but lower than sewage sludge and bamboo-based activated carbon due to diverse force or other interaction between organic functional group. On the other hand, small differences in adsorption capacity are observed between FA2.5 and other adsorbents. In terms of kinetic modeling, most dye adsorption best fits the pseudo-second-order model, with only PR adsorbed on Bokbunja wine industry waste (BWIW) fitting pseudo-first-order kinetics well.

Table 4. Comparison of MB and PR adsorption on different adsorbents

Adsorbent	Dye	Q m (mg/g)	Well-Fitted Kinetic Model	E a (kJ/mol)	Reference
FA10	MB	22.03	Pseudo-second order	61.71	This study
Fly ash	MB	0.7	Pseudo-second order	5.42	Lin et al., 2008
Fly ash	MB	7.9	Pseudo-second order	NA	Wang et al., 2005
Zeolite	MB	15.9	Pseudo-second order	NA
Unburned carbon	MB	79.9	Pseudo-second order	12.4, 39.3
Zeolite	MB	19.94	Pseudo-second order	16.2–25.1	Han et al., 2009
Sewage sludge	MB	24.9–114.9	Pseudo-second order	NA	Otero et al., 2003
BBAC	MB	454.2	Pseudo-second order	NA	Hameed et al., 2007
FA2.5	PR	28.82	Pseudo-second order	9.07	This study
BWIW	PR	29.37	Pseudo-first order	NA	Binupriya et al., 2008
CNT	PR	39.84	Pseudo-second order	33.35	Wu et al., 2007
Notes: NA = not available; BBAC = bamboo-based activated carbon; BWIW = Bokbunja wine industry waste; CNT = carbon nanotube.

Desorption experiment

Desorption results are shown in Figure 8. It is found that the DE value of PR is higher than MB (67.7% for PR and 6.0% for MB at 4 hr). The results correspond to the E _a value calculated from kinetic modeling whereby the higher E _a of MB adsorption (61.71 kJ/mol) reveals strong interaction between MB and FA10. Besides this, it was observed that the color of the solution changed to purple during MB desorption. After UV-Vis spectral scanning from 300 to 900 nm, it was found that the maximum adsorption wavelength changed from 663 to 598 nm. The color change implies a new compound being produced during desorption. It is thought that desorption occurred at a different position to the adsorption site, resulting in the desorbed compound having a different chemical structure, and, therefore, an altered maximum adsorption wavelength. This variation between the desorption site and adsorption site reveals that the interaction between MB and FA10 is strong and adsorption can be classified as chemical adsorption. On the other hand, the high value of DE (%) for PR desorption implies that the interaction between PR and FA2.5 is not strong and this weak interaction may be seen by the lower E _a value.

Figure 8. Desorption results of MB and PR.

Conclusion

SSA was modified by ferrite process. The synthetic Fe₃O₄ crystal on SSA not only provides electrostatic attraction toward ionic dyes, but also acts as active adsorption sites, leading to increased adsorption ability. Furthermore, magnetized modified SSA can be recovered using a magnetic field, which provides a more convenient recovery method than conventional separation techniques. In this study, Fe₃O₄ was synthesized on SSA particle by mixing SSA with a solution containing ferric and ferrous ions, adjusting the pH to 11 and reacting the mixture at 90 °C for 2 hr. Through the Fe₃O₄ process, magnetically modified SSA was produced. The products were termed FA2.5, FA5, and FA10 according to their respective dosing with SSA (i.e., 2.5, 5.0, and 10.0 g, respectively). The adsorption ability of modified SSA toward MB and PR was based on electrostatic attraction. Adsorbents had negative surface charge under alkaline condition, allowing them to adsorb a cationic dye (MB). On the other hand, a positive surface charge existed on the surface of adsorbents under acid conditions, allowing for anionic dye (PR) adsorption. SSA not subject to the ferrite process had weak adsorption ability, with the adsorption percentage for MB and PR being lower than 65% and 10%, respectively (100 mg/L of initial dye concentration). By comparison, the adsorption ability among FA2.5, FA5, and FA10 showed that FA10 had a very high adsorption ability toward MB (99% of adsorption percentage) at pH 9.0. For PR adsorption at pH 3.0, FA2.5 had the highest adsorption percentage (93.7%). As a result, FA2.5 and FA10 were chosen as PR and MB adsorbents and were utilized in the study of adsorption isotherms and kinetic models. For adsorption isotherms, the experimental data indicate that MB and PR adsorption best fit the Langmuir isotherm, with the adsorption capacities for MB and PR being 22.03 and 28.82 mg/g, respectively. The kinetic study revealed that MB and PR adsorption fit well the pseudo-second-order kinetic model, and the respective values of activation energy calculated from rate constants were 61.71 and 9.07 kJ/mol for MB and PR. According to the activation energy, it is thought that MB and PR adsorption can be described as chemical adsorption.

Funding

The authors would like to thank the National Science Council of the Republic of China for financially supporting this research under contract no. NSC 99-2221-E-122 -001-MY2 and NSC 102-2221-E-122-001.