The optimization of As(V) removal over mesoporous alumina by using response surface methodology and adsorption mechanism

Journal of Hazardous Materials 254–255 (2013) 301–309

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

The optimization of As(V) removal over mesoporous alumina by using response surface methodology and adsorption mechanism Caiyun Han a , Hongping Pu a , Hongying Li a , Lian Deng a , Si Huang a , Sufang He b , Yongming Luo a,∗ a b

Faculty of Environmental Science and Engineering, Kunming University of Science and Technology, Kunming 650500, PR China Research Center for Analysis and Measurement, Kunming University of Science and Technology, Kunming 650093, PR China

h i g h l i g h t s • • • •

Mesoporous alumina was synthesized at room temperature by using P123 as a template. Box–Behnken Design was employed to optimize arsenic adsorption process. Interactive effects of adsorption parameters on arsenic adsorption capacity were investigated. As(V) adsorption mechanisms over MA under different pH conditions were illustrated in detail.

a r t i c l e

i n f o

Article history: Received 20 December 2012 Received in revised form 25 March 2013 Accepted 8 April 2013 Available online xxx Keywords: Mesoporous alumina As(V) adsorption Optimization FT-IR characterization Adsorption mechanism

a b s t r a c t The Box–Behnken Design of the response surface methodology was employed to optimize four most important adsorption parameters (initial arsenic concentration, pH, adsorption temperature and time) and to investigate the interactive effects of these variables on arsenic(V) adsorption capacity of mesoporous alumina (MA). According to analysis of variance (ANOVA) and response surface analyses, the experiment data were excellent fitted to the quadratic model, and the interactive influence of initial concentration and pH on As(V) adsorption capacity was highly significant. The predicted maximum adsorption capacity was about 39.06 mg/g, and the corresponding optimal parameters of adsorption process were listed as below: time 720 min, temperature 52.8 ◦ C, initial pH 3.9 and initial concentration 130 mg/L. Based on the results of arsenate species definition, FT-IR and pH change, As(V) adsorption mechanisms were proposed as follows: (1) at pH 2.0, H3 AsO4 and H2 AsO4 − were adsorbed via hydrogen bond and electrostatic interaction, respectively; (2) at pH 6.6, arsenic species (H2 AsO4 − and HAsO4 2− ) were removed via adsorption and ion exchange, (3) at pH 10.0, HAsO4 2− was adsorbed by MA via ion exchange together with adsorption, while AsO4 3− was removed by ion exchange. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Removing arsenic from contaminated water has attracted considerable attention due to high toxicity for human and other organisms. Moreover, arsenic concentrations (levels) even in natural water of Argentina, Australia, Bangladesh, Vietnam, West Bengal, India, Taiwan and China are far higher than the corresponding rules of these country and area [1–7]. Therefore, several methods including electrocoagulation, precipitation, filtration, reverse osmosis, ion exchange, membrance, biological treatment together with adsorption have been developed for disposing arsenic-contaminated water [8–12]. Among them, adsorption has been recognized as an effective and most extensive technique

∗ Corresponding author. Tel.: +86 871 5103845; fax: +86 871 5103845. E-mail address: [email protected] (Y. Luo). 0304-3894/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jhazmat.2013.04.008

owing to its high removal efficiency, low cost as well as simple operation. In the past few decades, lots of researches involved arsenic adsorption have been devoted to test the uptake and isotherms of various adsorbents, and the corresponding adsorbents included activated carbon, zeolite, metal oxide (Xm On , X = Al, Fe, Ti, Mn, Cu, Zr and their composites), biosorbent, synthetic resin, industrial/agriculture byproducts or wastes etc. [13]. According to the classification made by United Nations Environmental Program agency (UNEPA), activated alumina is one of the most available adsorbents for removing arsenic from contaminated water. Nevertheless, traditional commercial activated alumina (TCAA) generally suffers from the drawbacks of low adsorption capacity, slow adsorption rate and narrow working pH region, which should be closely associated with its ill-defined pore structure together with small surface area. A perfect example can be found in Zhang and co-worker’s report that the optimum pH for removing As(V) over TCAA is in a narrow range of 5.5–6.0 and

302

C. Han et al. / Journal of Hazardous Materials 254–255 (2013) 301–309

the removal of As(V) sharply decreased beyond the pH region [14]. Recently, many researchers have tended to improve the performance of TCAA by doping alum, manganese and copper [15–17], while the corresponding adsorption capacity of these modified alumina-based adsorbents is still low. Since the discovery of the M41S family (silica-based mesoporous materials) has sparked considerable interest in the synthesis of mesoporous materials for the use in the many fields including catalysis [18,19], adsorption [20,21], separation [22,23], sensors [24], optics [25] as well as fabrication of novel nano-object materials [26]. In the past decades, mesoporous alumina (MA) has been synthesized via various methods, and MA with respect to TCAA exhibits more excellent performance in adsorption for many pollutants such as simazine [27], and CO2 [28], 4-chloro-2-methylphenoxyacetic acid [29]. Of late, limited efforts have been devoted to synthesize MA as high-performance adsorbent for arsenic(V) removal [30–32]. However, these synthesis routes included high-temperature crystallization (100–120 ◦ C) together with the use of expensive aluminum alkoxide (such as aluminum sec-butoxide) and organic solvent (such as sec-BuOH). Moreover, it was documented that nonionic polyethylene oxide (PEO) surfactant with respect to cation and anion templates has attracted much attention due to low-cost, nontoxic and biodegradable [33]. In addition, compared with high-temperature synthesis, low-temperature synthesis is an environmentally friendly method, which will be in favor of energy-saving. Therefore, despite significant progress in the synthesis of MA, further studies in this area, especially in relation to low-temperature synthesis route by using low-cost, nontoxic and biodegradable templates are desirable. In general, arsenic removal is affected by many factors including As(V) concentration, pH, adsorption temperature and time etc. [34]. In order to assess the effect of experiment parameters on adsorption capacity, the proper use of an adequate experimental design is of particular importance. Response Surface Methodology (RSM), a collection of mathematical and statistical techniques, has been found to be a useful method for studying the mutual interaction between the variables and optimizing the variables in the adsorption process. This contribution was aimed at the investigation of As(V) adsorption performances over MA, which was prepared with nonionic surfactant P123 and aluminum isopropoxide (AIP) at room temperature. The Box–Behnken Design (BBD) of the RSM [35–37] was employed to investigate the effects of significant operating parameters including initial arsenic concentrations, pH, adsorption time and temperature on arsenic adsorption capability and to find the most suitable combination of variables resulting in maximum As(V) adsorption capability. The second-order polynomial equation (regression model) provided an excellent explanation of the relationship between the response (arsenic adsorption capacity) and these independent parameters. More important, according to the results of arsenic species definition, FT-IR characterization together with pH change during the whole adsorption process, the arsenic adsorption mechanisms over MA under various pH conditions were investigated and illustrated in detail.

respectively, and the third step was the synthesis of MA. In the first step, 20.4 g of AIP and a small amount of nitric acid (65 wt.%) was added into 160 mL of hot denionized water with vigorously stirring for 2–4 h to yield aluminum hydroxide sol. In the second step, 7.5 g of P123 and 0.24 mol of HCl were added into 150 mL of denionized water with vigorously stirring under room temperature (RT) for 2–4 h. After P123 was completely dissolved, the template solution was formed. In the third step, the solution contained template P123 was added into the aluminum hydroxide sol. After that the mixture was stirred at 40 ◦ C for 12–24 h, and the corresponding pH value was adjusted to 7.0 by using sodium hydroxide solution. Subsequently, the resulting mixture was kept at RT under static conditions for 36–72 h, and the reaction products were filtered, washed with the mixture of water and ethanol, and dried at 105 ◦ C for 24–48 h. Finally, the dried sample was calcined at 400 ◦ C in air for 5 h. The batch experiments of arsenic adsorption were carried out by mixing adsorbent MA with arsenic solution in a series of 100 mL conical flask under magnetic stirring conditions, and the resulting mixtures were centrifuged after adsorption. Arsenic(V) concentration before and after adsorption was measured by atomic fluorescence spectrometer (AFS-230E), and the As(V) adsorption capacity of MA was calculated using following equation:

2. Experimental

The validity of the equation was analyzed by using ANOVA (analysis-of-variance), and fit quality of the equation was judged from the coefficients of correlation and P-value.

Mesoporous alumina (MA) was synthesized by employing nonionic triblock copolymer P123 (EO20 PO70 EO20 ) and aluminum tri-isopropoxide (AIP) as a structure-directing agent and an aluminum source, respectively. The following procedures were used to prepare MA alumina, which were mainly composed of three steps: the first step and the second step consist of the synthesis of aluminum hydroxide gel and the preparation of template solution,

qt =

(C0 − Ct ) × V m

(1)

where qt denotes arsenic(V) uptake capacity of MA at contact time t, C0 and Ct are the arsenic(V) concentrations before and after adsorption, V is the volume of adsorption solution, and m is the weight of adsorbent MA. 2.1. Box–Behnken Design and As(V) adsorption optimization RSM was employed to investigate the effects of different operating factors on arsenic adsorption capacity, and reveal the optimum conditions for As(V) removal as well as build models. BBD was applied to evaluate the interactive effects of adsorption variables and optimize the adsorption process. The effects of variables (adsorption parameters including pH value, initial concentration, adsorption time and temperature) on adsorption capacity were selected for RSM. The corresponding experiments were carried out in a system with 0.05 g MA and 50 mL As(V) solution, and pH was adjusted by 1.0 M HCl (or 1.0 M NaOH). The range and levels for these variables are coded according to Eq. (2) and summarized in Table 1. xi =

Xi − X0 x

(2)

where xi and Xi are the coded and the real values of variables. X0 and x are the center point of Xi and the step change in Xi , respectively. Generally, the mathematical relationship between the response Y (arsenic (V) adsorption capacity) and these variables can be described by the following second-order polynomial equation: Y = ˇ0 + ˇ1 A + ˇ2 B + ˇ3 D + ˇ11 A2 + ˇ22 B2 + ˇ33 C 2 + ˇ44 D2 + ˇ12 AB + ˇ13 AC + ˇ14 AD + ˇ23 BC + ˇ24 BD + ˇ34 CD

(3)

2.2. Characterization Powder XRD patterns were performed on a Rigaku D/max 2550PC diffractometer using Cu Ka radiation, operating at 40 kV and 300 mA. N2 adsorption–desorption isotherm and BET surface

C. Han et al. / Journal of Hazardous Materials 254–255 (2013) 301–309

303

Table 1 Process variables and their level for the adsorption of As(V) by BBD. Factors

Name

Units

Low actual

X1 X2 X3 X4

Time Temperature pH Initial concentration

Min ◦ C – mg/L

30 15.0 2.0 11.18

High actual 1440 60.0 10.0 130

Low coded

Middle coded

High coded

−1 −1 −1 −1

0 0 0 0

1 1 1 1

60.6◦ and 66.8◦ ) are expressed in the wide XRD pattern of the sample (Fig. 2, inset), which are well matched to the diffraction peaks of ˚ ␥-Al2 O3 crystalline phases (JCPDS Card No. 10-0425, a = 7.9103 A). Fig. 3 shows TEM images of MA. The sample MA showed a wormlike motif of the micellar structure, which is associated with the result of small angle XRD that only one diffraction peak is observed (Fig. 2). In light of N2 adsorption–desorption, XRD and TEM characterization resutls, the sample (MA) was identified as a disordered mesoporous alumina. 3.2. Regression model The main objective of RSM is to determine regression model of adsorption process, and the quadratic model was used to find out the relationship between the response and variables in this article. According to experiment data, the final empirical model of As(V) adsorption over MA was described by using Eq. (4). Y = 28.94 + 1.30X1 + 0.90X2 − 6.03X3 + 12.70X4 − 0.37X1 X2 − 0.11X1 X3 + 1.07X1 X4 − 0.65X2 X3 + 0.12X2 X4 − 4.84X3 X4 − 1.11X12 − 0.91X22 − 6.37X32 − 8.98X42

Fig. 1. N2 adsorption–desorption isotherm of MA.

−196 ◦ C.

area were carried out on an ASAP 2020 apparatus at All samples were degassed at 250 ◦ C for 2 h prior to analysis. BET specific surface area was calculated from adsorption data in the relative pressure range from 0.05 to 0.25. Fourier transforms infrared (FTIR) spectra of the samples in the form of KBr pellets were recorded by using a Nicolet 560 IR spectrometer with the wavenumber range of 4000–400 cm−1 . 3. Results and discussion

where Y is the predicted response (arsenic adsorption capacity), and X1 , X2 , X3 and X4 are the coded values of time, temperature, pH and initial concentration, respectively. In order to ensure the adequacy of employed model, an adequate fit of the model should be given to avoid poor or ambiguous results. The significance of quadratic regression model was tested by the value of F, P and correlation coefficient, and the corresponding results of ANOVA were tabulated in Table 3. The model F-value (65.03) and a very low p-value (less than 0.0001) implied that the model was highly significant for As(V) adsorption on MA. “Adequate Precision” measures the signal to noise ratio, and the ratio greater than 4.0 is desirable [41]. The “Adequate Precision” ratio

3.1. Textural properties N2 adsorption–desorption isotherm of the sample MA are presented in Fig. 1, and the corresponding physicochemical properties including BET surface, pore size and pore volume are tabulated in Table 2. It is obvious that the N2 adsorption–desorption isotherm is found to be of type IV and a clear type-H2 hysteresis loop is observed for the sample. The isotherm of sample MA is characterized by a well-defined step at P/P0 = 0.60–0.85, which arises from capillary condensation in the mesoporous. BET surface area and pore volume of the sample are 312 m2 /g and 0.5 cm3 g−1 , respectively. Small and wide angle XRD patterns of the sample were shown in Fig. 2. The presence of a single diffraction peak in 2 region blow 2◦ for the sample was an indicative of the formation of mesostructure [38–40]. It is clear that no diffraction peaks assigned to long range ordered structure were detected in 2 range between 0.5◦ and 8.0◦ . Moreover, six diffraction peaks (at 2 = 32.2◦ , 36.9◦ , 39.3◦ , 45.8◦ , Table 2 Textural properties of the adsorbent MA. BET surface area (m2 /g)

BJH pore size (nm)

Pore volume (cm3 g−1 )

312

5.8

0.5

(4)

Fig. 2. Small and wide angle (inset) XRD patterns of MA.

304

C. Han et al. / Journal of Hazardous Materials 254–255 (2013) 301–309

Table 3 the results of ANOVA for the response surface quadratic model. Source

Sum of squares

D.f.

Mean square

F-value

p-value > F

Mode X1 -time X2 -temperature X3 -pH X4 -initial concentration X1 X2 X1 X3 X1 X4 X2 X3 X2 X4 X3 X4 X1 2 X2 2 X3 2 X4 2 Residual Lack of fit Pure error

3196.39 20.16 9.76 436.69 1934.46 0.56 0.046 4.57 1.69 0.056 93.75 7.96 5.33 262.97 522.88 49.15 49.15 0.00

14 1 1 1 1 1 1 1 1 1 1 1 1 1 1 6 10 4

228.31 20.16 9.76 436.69 1934.46 0.56 0.0046 4.57 1.69 0.056 93.75 7.96 5.33 262.97 522.88 3.51 4.92 0.00

65.03 5.74 2.78 124.39 551.00 0.16 0.013 1.30 0.48 0.016 26.70 2.27 1.52 74.90 148.94

<0.0001 0.0311 0.1177 <0.0001 <0.0001 0.6950 0.9103 0.2729 0.4992 0.9014 0.0001 0.1543 0.2384 <0.0001 <0.0001

R2 = 0.9849, adjusted R2 = 0.9697, predicted R2 = 0.9528, adequate precision = 27.797.

of this model (27.797) is far greater than 4.0, which indicates the presence of adequate signal for the model. It is noticeable that the value lack of fit is insignificant, which demonstrate the quadratic model for As(V) adsorption over MA is very valid. The value of determination coefficient (R2 = 0.9849) of Eq. (4) indicates that the regression model is best suited for predicting the performance of As(V) adsorption with MA and only 1.5% of the total variations were not satisfactorily explained by the model. The value of predicted multiple correlation coefficient (R2 = 0.9528) is reasonable in agreement with the value of adjusted multiple correlation coefficient (R2 = 0.9697). As can be seen from Table 3, all the p-values of X1 , X3 , X4 , X3 X4 , X3 2 and X4 2 are less than 0.05, which indicates that these variables are significant and have great influence on arsenic adsorption capacity. The fitted quality of Eq. (4) was also expressed by comparing arsenic adsorption capacity between experiment and model predicted, as shown in Fig. 4. It is clear that the predicted values are quite close to the actual experiment, thus confirming that the regression model exhibits excellent stability for As(V) adsorption on MA. Therefore, it can be concluded the response surface model

Fig. 3. TEM images of MA obtained on a JEM-2010 HR transmission electron microscope. (A) low magnification image and (B) high magnification image.

Fig. 4. The predicted values versus experimental values of As(V) adsorption capacity over MA.

C. Han et al. / Journal of Hazardous Materials 254–255 (2013) 301–309

305

Fig. 5. 3-D surface plots for interactive effect of (A) pH and initial concentration at adsorption time 720 min, adsorption temperature 54 ◦ C, (B) time and pH at initial concentration 60 mg/L, adsorption temperature 54 ◦ C, (C) temperature and pH at adsorption time 720 min, initial concentration 60 mg/L, (D) time and initial concentration at initial pH 3.90, adsorption temperature 54 ◦ C, (E) temperature and initial concentration at initial pH 3.90, adsorption time 720 min, (F) time and temperature at initial pH 3.90, initial concentration 130 mg/L.

306

C. Han et al. / Journal of Hazardous Materials 254–255 (2013) 301–309

Fig. 6. Speciation digram of As (V) defined by computer program Mineql+. (a) H3 AsO4 , (b) H2 AsO4 − , (c) HAsO4 2− , (d) AsO4 3− .

developed in this study (Eq. (4)) was considered to be satisfactory for the prediction of As(V) adsorption system. 3.3. Interactive effects of adsorption variables Fig. 5 illustrates the mutual interactive effects of the combination of independent variables on arsenic (V) adsorption capacity in the manner of 3-D surface plots, and these plots were represented as a function of two factors by holding other factors at a fixed level. As can be seen from Fig. 5, the interactive effect of pH and initial concentration on As(V) adsorption capacity is of the highest significant (Fig. 5(A)), which is excellent in agreement with the results observed from Table 3 (p = 0.0001). As shown in Fig. 5(B) and (C), the adsorption capacity firstly increased and then kept constant with pH in the range of 2.0–3.9 and 3.9–5.0, respectively. However, the adsorption capacity was decreased when the pH increased from 5.0 to 10.0. The phenomena might be closely associated with the different interactions between adsorption sites and As (V) species under various pH. As can be seen from Fig. 5(D) and (E), As(V) adsorption capacity increased with increasing As(V) initial concentration from 11.18 mg/L to 130 mg/L due to larger amount of As(V) ions competing for the available adsorption sites of MA. Moreover, it also noted that the contribution of adsorption time to adsorption capacity for high initial concentration is larger than that of low initial concentration (Fig. 5(D)), which might be ascribed to the quick saturation of adsorption sites for MA under low As(V) concentration. In addition, the interactive effect of adsorption temperature and adsorption time on the response of As(V) adsorption capacity is far less and can be ignored, as shown in Fig. 5(F). The optimum predicted point of maximum adsorption capacity obtained by Design-Expert 8.0.6 is about 39.06 mg/g, and the corresponding optimal parameters of adsorption process are listed as below: adsorption time 720 min, temperature 52.8 ◦ C, initial pH 3.9 and initial concentration 130 mg/L. 3.4. Adsorption mechanism 3.4.1. Arsenate species definition In order to investigate As(V) adsorption mechanism over MA, the arsenic species under different pH conditions were defined by using a well-known computer program (Mineql+, 4.6), as shown in Fig. 6. According to the diagram, with pH value increasing from 1.0 to 13.0, the As(V) species were changed from H3 AsO4 to

Fig. 7. FT-IR spectra of (a) Fresh MA, (b) MA after adsorption at pH 2.0, (c) MA after adsorption at pH 6.6, (d) MA after adsorption at pH 10.0.

AsO4 3− via H2 AsO4 − and HAsO4 2− in turn. It is noted that the arsenic specie is dominatingly presented as H3 AsO4 or AsO4 3− when pH value is below 1.0 or is above 13.0. Furthermore, H2 AsO4 − and HAsO4 2− are the major Arsenic species (>95%) in the pH range of 3.5–5.5 and 8.0–10.0, respectively. In light of the optimization results obtained from RSM (Fig. 5), it can be concluded that MA is prefer to adsorb H2 AsO4 − rather than other As(V) species and the adsorption ability of MA for these species is in the order of H2 AsO4 − > H3 AsO4 > HAsO4 2− > AsO4 3− . 3.4.2. FT-IR characterization FT-IR spectra of MA samples before and after adsorption As(V) species are presented in Fig. 7. The band centered around 570 cm−1 , assigned to the stretch vibration of Al–O(AlO6 ) [17,42,43], is detected for all the four samples. The band centered around 845 cm−1 is well–expressed in the spectra of the samples after adsorption, but not for fresh MA, which has been attributed to characteristic vibration of As–O in H2 AsO4 − [17,44], thus indicating that non-surface complex As–O bond of As(V) adsorption was formed under various pH conditions. Two distinct IR vibration bands centered at 948 and 1259 cm−1 are only detected for the sample after adsorption at pH 2.0, which is the characteristic of asymmetrical stretching vibration of As–O in H3 AsO4 [11]. The result is excellent consistent with the existence of arsenic specie (Fig. 6) and implies the presence of interaction between MA and H3 AsO4 via Van der Waals force. Two bands centered near 1029 and 1090 cm−1 , ascribed to the symmetrical and asymmetrical stretching vibrations of Al-O-H [43,45,46], are expressed in the spectrum of fresh MA. However, these two bands shift to higher wavenumber (1070 and 1122 cm−1 ) for the samples after adsorption, which might be attributed to the interaction between Al-O-H group and arsenic species. In addition, analogous shift phenomenon (from 1440 cm−1 to 1456 cm−1 ) was detected for the band at 1440 cm−1 ascribed to the stretching vibration of O–H [47,48], which should be closely associated with the interrelationship between MA and arsenic species.

C. Han et al. / Journal of Hazardous Materials 254–255 (2013) 301–309

307

Fig. 8. FT-IR spectra of (a) MA adsorbed at pH 6.6 and (b) regenerated sample (a).

As can be seen by comparing Fig. 7, the band at 1389 cm−1 is observed in the spectra of the MA samples after adsorption at pH 6.6 and pH 10.0, but not for fresh MA as well as the MA sample after adsorption at pH 2.0. Combing with the result of arsenate species definition (Fig. 6) that only HAsO4 2− was simultaneous presented at pH 6.6 and pH 10.0 media. Therefore, the band can be conceivably assigned to the As–O vibration of HAsO4 2− , which was further demonstrated by comparing MA after adsorption at pH 6.6 with the corresponding sample regenerated by using 0.05 M NaOH (Fig. 8). 3.4.3. pH change In light of RSM results, pH is one of the most important factors affecting the adsorption performances of MA. Therefore, in order to obtain As(V) adsorption mechanism over MA, the pH values of arsenic solution before and after adsorption with MA were measured, and the corresponding results were summarized in Table 4. As shown in Table 4, the pH values of arsenic solution after adsorption increase under acidic and near-neutral media, which should be assigned to the fact that adsorbent MA was protonated. This means that some hydrogen protons (H+ ) in solution have been adsorbed by MA. However, under alkaline conditions, a distinct reduce in pH was detected for the solution after adsorption with respect to initial arsenic solution, which might be ascribed to the presence of competition adsorption between HAsO4 2− , AsO4 3− and OH− for the active sites of MA. In order to further demonstrate our postulation, the pH changes of arsenic solution during the whole adsorption process under various initial pH values are presented in Fig. 9. It is noticeable that the increase in pH value is detected for As(V) adsorption under acidic condition (pH 2.0), which have been attributed to the fact that some hydrogen protons (H+ ) of the solution were adsorbed on the surface of MA with adsorption processing, thus not only causing MA protonation but reducing the amounts of H+ in the adsorption solution. Table 4 The changes of solution pH before and after adsorption. Before adsorption After adsorption

2.0 2.4

3.0 4.6

4.2 5.4

5.1 6.1

6.6 6.9

7.6 7.9

8.6 8.0

10.0 8.4

Fig. 9. Plots of pH change for arsenic solution under various initial pH values with time in the whole adsorption process. (A) pH 2.0, (B) pH 6.6, (C) pH 10.0.

In addition, it was demonstrated that the isoelectric point of Al2 O3 is estimated to be about 8.3 (pHZPC ≈ 8.3) [15], which is an intrinsic property of solid-water interface and is only determined by the chemical nature of materials. Therefore, it can be concluded that MA is positively charged when solution pH is lower than 8.3, thus providing an indirect and authentic proof for our postulation reported in this article. As shown in Fig. 9 (C), a distinct decrease in pH is observed for As(V) adsorption in the first hour at alkaline condition (pH 10.0), and then the solution pH value slightly increase with adsorption time. The phenomena can be conceivably explained as

308

C. Han et al. / Journal of Hazardous Materials 254–255 (2013) 301–309

Fig. 10. Schematic diagrams of arsenic adsorption mechanisms for MA under different pH value conditions.

follows: large amounts of hydroxide ion (OH− ) of the solution were preferentially adsorbed by MA due to it’s weak acidity [49], thus leading to the quick reduce in pH value. After that, some hydrogen protons (H+ ) were also adsorbed by MA, which is likely because that MA is positively charged when solution pH is lower than alumina isoelectric point (pHZPC ≈ 8.3). On the other hand, the adsorbed hydroxide ions should be exchanged with arsenic species of the solution to release some OH− ions into solution. The ion exchange between arsenic species and the adsorbed hydroxide ions together with MA protonation caused the increase in pH for As(V) adsorption. Analogous phenomena were observed for As(V) adsorption under near-neutral condition (pH 6.6), the decrease in pH value during the first hour has been ascribed to the synergistic effects of the preferential adsorption of hydroxide ions by the acidic centers of MA as well as MA protonation, whereas the increase of pH in the following adsorption process has assigned to the ion exchange between the adsorbed hydroxide ions and arsenic species. Based on the above results of arsenate species definition, FT-IR characterization and pH change, the possible mechanisms of As(V) adsorption over MA were put forward and illustrated in Fig. 10, which can be explained as follows: 1) under acidic solution (pH 2.0), H2 AsO4 − is preferentially adsorbed to the acidic centers and the protonated hydroxyl groups of MA by electrostatic interaction. Furthermore, some H3 AsO4 were simultaneously adsorbed by MA via hydrogen bond between H3 AsO4 and the unprotonated hydroxyl group, as confirmed by FT-IR characterization (Fig. 7); 2) at near-neutral media (pH 6.6), the acidic centers as well as the protonated hydroxyl groups of MA are in favor of H2 AsO4 − adsorption via electrostatic interaction with respect to HAsO4 2− . Moreover, the alkalized acidic centers of MA were preferred to exchange with H2 AsO4 − rather than HAsO4 2− to release OH− into the adsorption solution; 3) under alkaline condition (pH 10.0), arsenic species are

almost presented in forms of HAsO4 2− . The hydroxide ion (OH− ) of the alkalized acidic centers was preferentially exchanged with HAsO4 2− to release it into the adsorption solution. At the same time, few HAsO4 2− were adsorbed on the free weakly acidic centers and the protonated hydroxyl groups of MA, respectively. 4. Conclusion Mesoporous alumina (MA), prepared with nonionic surfactant P123 in water media under room temperature, was found to be an effective adsorbent for As(V) removal. The BBD was employed to optimize the important adsorption variables including As(V) concentration, pH, adsorption time and temperature and to investigate the interactive effects of these variables on the response (arsenic adsorption capacity) for MA. The experiment data were excellent fitted to the quadratic model, and the relationship between the response and the variables was well described by a second-order polynomial equation (regression model). Based on response surface analysis derived from the empirical model, the interactive influence of initial concentration and pH on As(V) adsorption capacity is highly significant, while the interactive effect of adsorption temperature and adsorption time on As(V) adsorption capacity is insignificant. The predicted maximum adsorption capacity is about 39.06 mg/g, and the corresponding optimal parameters of adsorption process are listed as below: time 720 min, temperature 52.8 ◦ C, initial pH 3.9 and initial concentration 130 mg/L. In light of the analysis results of arsenate species definition, FT-IR characterization and pH change, As(V) adsorption mechanisms over MA under different media were proposed as follows: 1) at acidic solution (pH 2.0), H3 AsO4 and H2 AsO4 − were adsorbed by MA via hydrogen bond and electrostatic interaction, respectively; 2) under nearneutral media (pH 6.6), arsenic species (H2 AsO4 − and HAsO4 2− )

C. Han et al. / Journal of Hazardous Materials 254–255 (2013) 301–309

were removed by MA via adsorption together with ion exchange, 3) under alkaline condition (pH 10.0), the hydroxide ion of the alkalized acidic centers was preferentially exchanged with HAsO4 2− rather than AsO4 3− , and few HAsO4 2− were adsorbed by the free weakly acidic centers and the protonated hydroxyl groups of MA via electrostatic interaction. Acknowledgements We gratefully acknowledge the support of this work by Natural Science Foundation of China (grant nos.: 51068010, 21003066 and 21267011) and Young Academic and Technical Leader Raising Foundation of Yunnan Province (grant no. 2008py010). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.jhazmat.2013.04.008. References [1] J. Matschullat, Arsenic in the geosphere—a review, Sci. Total Environ. 249 (2000) 297–312. [2] P. Mondal, C.B. Majumder, B. Mohanty, Laboratory based approaches for arsenic remediation from contaminated water: recent developments, J. Hazard. Mater. B 137 (2006) 464–479. [3] R.N. Ratnaike, Acute and chronic arsenic toxicity, Postgrad. Med. J. 79 (2003) 391–396. [4] C.M. Liao, H.H. Shen, T.L. Lin, S.C. Chen, C.L. Chen, L.I. Hsu, C.J. Chen, Arsenic cancer risk posed to human health from tilapia consumption in Taiwan, Ecotoxicol. Environ. Saf. 70 (2008) 27–37. [5] D.B. Menzel, H.K. Hamadeh, E. Lee, D.M. Meacher, V. Said, R.E. Rasmussen, H. Greene, R.N. Roth, Arsenic binding proteins from human lymphoblastoid cells, Toxicol. Lett. 105 (1999) 89–101. [6] V.K. Sharma, M. Sohn, Aquatic arsenic: toxicity, speciation, transformations, and remediation, Environ. Int. 35 (2009) 743–759. [7] A. Basu, J. Mahata, S. Gupta, A.K. Giri, Genetic toxicology of a paradoxical human carcinogen, arsenic: a review, Mutat. Res. 488 (2001) 171–194. [8] A. Violante, M. Ricciardella, S. Del Gaudio, M. Pigna, Coprecipitation of arsenate with metal oxides: nature, mineralogy, and reactivity of aluminum precipitates, Environ. Sci. Technol. 40 (2006) 4961–4967. [9] J. Floch, M. Hideg, Application of ZW 1000 membranes for arsenic removal from water sources, Desalination 162 (2004) 75–83. [10] S.Y. Thomas, Choong, T.G. Chuah, Y. Robiah, F.L. Gregory Koay, I. Azni, Arsenic toxicity, health hazards and removal techniques from water: an overview, Desalination 217 (2007) 139–166. [11] X.H. Guan, J.M. Wang, C.C. Chusuei, Removal of arsenic from water using granular ferric hydroxide: macroscopic and microscopic studies, J. Hazard. Mater. 156 (2008) 178–185. [12] M. Vaclavikova, G.P. Gallios, S. Hredzak, S. Jakabsky, Removal of arsenic from water streams: an overview of available techniques, Clean Technol. Environ. Policy 10 (2008) 89–95. [13] D. Mohana, C.U. Pittman Jr., Arsenic removal from water/wastewater using adsorbents—a critical review, J. Hazard. Mater. 142 (2007) 1–53. [14] Y. Zhang, M. Yang, X. Huang, Arsenic(V) removal with a Ce(IV)-doped iron oxide adsorbent, Chemosphere 51 (2003) 945–952. [15] S.S. Tripathy, A.M. Raichur, Enhanced adsorption capacity of activated alumina by impregnation with alum for removal of As(V) from water, Chem. Eng. J. 138 (2008) 179–186. [16] S.M. Maliyekkala, L. Philip, T. Pradeep, As(III) removal from drinking water using manganese oxide-coated-alumina: performance evaluation and mechanistic details of surface binding, Chem. Eng. J. 153 (2009) 101–107. [17] P. Pillewan, S. Mukherjee, T. Roychowdhury, S. Das, A. Bansiwal, S. Rayalu, Removal of As(III) and As(V) from water by copper oxide incorporated mesoporous alumina, J. Hazard. Mater. 186 (2011) 367–375. [18] S. Endud, K.L. Wong, Mesoporous silica MCM-48 molecular sieve modified with SnCl2 in alkaline medium for selective oxidation of alcohol, Micropor. Mesopor. Mater. 101 (2007) 256–263. [19] A. Martínez, C. López, F. Márquez, I. Díaz, Fischer-Tropsch synthesis of hydrocarbons over mesoporous Co/SBA-15 catalysts: the influence of metal loading, cobalt precursor, and promoters, J. Catal. 220 (2003) 486–499. [20] C.Y. Han, H. Wang, L.Y. Zhang, R.T. Li, Y.Y. Zhang, Y.M. Luo, X.M. Zheng, Characterization and investigation on the difference of hydrothermal stability for ordered mesoporous aluminosilicate sieves, Adv. Powder Technol. 22 (2011) 20–25. [21] Y.M. Luo, Z.Y. Hou, R.T. Li, X.M. Zheng, Rapid synthesis of ordered mesoporous silica with the aid of heteropoly acids, Micropor. Mesopor. Mater. 109 (2008) 585–590.

309

[22] H. Wang, C.Y. Han, H.P. Pu, L.Y. Zhang, Z.H. Zou, Y.M. Luo, Rapid synthesis of mesoporous SBA-16 silica in weakly acidic media, Bull. Chem. Soc. Jpn. 83 (2010) 852–854. [23] Y. Ueno, A. Tate, O. Niwa, H.S. Zhou, T. Yamada, I. Honma, High benzene selectivity of uniform sub-nanometer sized pores of self-ordered mesoporous silicate, Chem. Commun. 6 (2004) 746–747. [24] C.X. Song, X.L. Zhang, C.Y. Jia, P. Zhou, X. Quan, C.Y. Duan, Highly sensitive and selective fluorescence sensor based on functional SBA-15 for detection of Hg2+ in aqueous media, Talanta 81 (2010) 643–649. [25] L. Laskowski, A. Kassiba, M. Makowska-Janusik, N. Errien, A. Mehdi, J. Swiatek, Synthesis and optical behaviour of mesoporous silica functionalized by organometallic molecules, J. Phys.: Conf. Ser. 289 (2011) 012024. [26] H.P. Pu, C.Y. Han, H. Wang, S.W. Xu, L.Y. Zhang, Y.Y. Zhang, Y.M. Luo, Characterization and investigation mechanism of hydrothermal stability for micropore-free ordered mesoporous molecular sieves, Appl. Surf. Sci. 258 (2012) 8895–8901. [27] V. Addorisio, D. Pirozzi, S. Esposito, F. Sannino, Decontamination of waters polluted with simazine by sorption on mesoporous metal oxides, J. Hazard. Mater. 196 (2011) 242–247. [28] C. Chen, W.S. Ahn, CO2 capture using mesoporous alumina prepared by a sol–gel process, Chem. Eng. J. 166 (2011) 646–651. [29] V. Addorisio, S. Esposito, F. Sannino, Sorption capacity of mesoporous metal oxides for the removal of MCPA from polluted waters, J. Agric. Food Chem. 58 (2010) 5011–5016. [30] A.K. Patra, A. Dutta, A. Bhaumik, Self-assembled mesoporous ␥-Al2 O3 spherical nanoparticles and their efficiency for the removal of arsenic from water, J. Hazard. Mater. 201–202 (2012) 170–177. [31] Y. Kim, C. Kim, I. Choi, S. Rengaraj, J. Yi, Arsenic removal using mesoporous alumina prepared via a templating method, Environ. Sci. Technol. 38 (2004) 924–931. [32] M.J. Yu, X. Li, W.S. Ahn, Adsorptive removal of arsenate and orthophosphate anions by mesoporous alumina, Micropor. Mesopor. Mater. 113 (2008) 197–203. [33] S.A. Bagshaw, E. Prouzet, T.J. Pinnavaia, Templating of mesoporous molecular sieves by nonionic polyethylene oxide surfactants, Science 269 (1995) 1242–1244. [34] P. Lakshmipathiraj, B.R.V. Narasimhan, S. Prabhakar, G. Bhaskar Raju, Adsorption of arsenate on synthetic goethite from aqueous solutions, J. Hazard. Mater. B 136 (2006) 281–287. [35] R. Singh, R. Chadetrik, R. Kumar, K. Bishnoi, D. Bhatia, A. Kumar, N.R. Bishnoi, N. Singh, Biosorption optimization of lead(II), cadmium(II) and copper(II) using response surface methodology and applicability in isotherms and thermodynamics modeling, J. Hazard. Mater. 174 (2010) 623–634. [36] Z. Wu, H. Li, J. Ming, G.H. Zhao, Optimization of adsorption of tea polyphenols into oat ␤-glucan using response surface methodology, J. Agric. Food Chem. 59 (2011) 378–385. [37] P. Rana-Madaria, M. Nagarajan, C. Rajagopal, B.S. Garg, Removal of chromium from aqueous solutions by treatment with carbon aerogel electrodes using response surface methodology, Ind. Eng. Chem. Res. 44 (2005) 6549–6559. [38] Y.M. Luo, Z.Y. Hou, D.F. Jin, J. Gao, X.M. Zheng, Template assisted synthesis of Ga2 O3 –Al2 O3 nanorods, Mater. Lett. 60 (2006) 393–395. [39] Q. Yuan, A.X. Yin, C. Luo, L.D. Sun, Y.W. Zhang, W.T. Duan, H.C. Liu, C.H. Yan, Facile synthesis for ordered mesoporous ␥-aluminas with high thermal stability, J. Am. Chem. Soc. 130 (2008) 3465–3472. [40] B.J. Xu, T.C. Xiao, Z.F. Yan, X. Sun, J. Sloan, S.L. Gonzalez-Cortes, F. Alshahrani, M.L.H. Green, Synthesis of mesoporous alumina with highly thermal stability using glucose template in aqueous system, Micropor. Mesopor. Mater. 91 (2006) 293–295. [41] J.H. Zhang, S.X. Chen, R. Yang, Y.Y. Yan, Biodiesel production from vegetable oil using heterogenous acid and alkali catalyst, Fuel 89 (2010) 2939–2944. [42] C. Liu, Y.C. Liu, Q.X. Ma, H. He, Mesoporous transition alumina with uniform pore structure synthesized by alumisol spray pyrolysis, Chem. Eng. J. 163 (2010) 133–142. [43] S. Music, D. Dragcevic, S. Popovic, Hydrothermal crystallization of boehmite from freshly precipitated aluminium hydroxide, Mater. Lett. 40 (1999) 269–274. [44] Y. Ma, Y.M. Zheng, J. Paul Chen, A zirconium based nanoparticle for significantly enhanced adsorption of arsenate: synthesis, characterization and performance, J. Colloid Interface Sci. 354 (2011) 785–792. [45] Y.M. Chen, M.K. Wang, P.M. Huang, T.M. Tsao, K.C. Lin, Influence of catechin on precipitation of aluminum hydroxide, Geoderma 152 (2009) 296–300. [46] H.W. Hou, Y. Xie, Q. Yang, Q.X. Guo, C.R. Tan, Preparation and characterization of ␥-AlOOH nanotubes and nanorods, Nanotechnology 16 (2005) 741–745. [47] G. Blázquez, M.A. Martin-Lara, G. Tenorio, M. Calero, Batch biosorption of lead(II) from aqueous solutions by olive tree pruning waste: equilibrium, kinetics and thermodynamic study, Chem. Eng. J. 168 (2011) 170–177. [48] B. Saha, S. Chakraborty, G. Das, A mechanistic insight into enhanced and selective phosphate adsorption on a coated carboxylated surface, J. Colloid Interface Sci. 331 (2009) 21–26. [49] M. El Doukkali, A. Iriondo, P.L. Arias, J.F. Cambra, I. Gandarias, V.L. Barrio, Bioethanol/glycerol mixture steam reforming over Pt and PtNi supported on lanthana or ceria doped alumina catalysts, Int. J. Hydrogen Energy 37 (2012) 8298–8309.