1. Introduction
The contamination of water sources by various arsenic species are a critical worldwide issue due to the toxicity and carcinogenicity of these arsenic compounds [1]. Arsenic is one of the most toxic chemicals in the environment. It enters drinking water through a variety of sources including the mobilization under natural conditions, the mining activities, the combustion of fossil fuels, the use of crop desiccants and arsenic herbicides [2]. Excess arsenic in drinking water can cause several health problems, including neurological complications, respiratory problems, and skin lesions [3]. The European Commission and the United States [4] have set the maximum acceptable contaminant level of 10 g/L of arsenic in drinking water. The most common arsenic species in water are arsenate (AsO43−) and arsenite (AsO33−). Arsenite is considered to be more toxic and mobile than arsenate and is difficult to be removed from waters [5]. Various techniques have been employed to remove arsenic from water, such as adsorption [4], [6], bioreactors [7], chemical precipitation [8], electrokinetic processes [9], [10], electrocoagulation [11], ion exchanges [12], phytoremediation [13], nanofiltration and reverse osmosis [14] and vertical flow wetlands [15] and thin film membrane [16] [17]. Among these methods, adsorption is of interest of many researchers because of low cost, easy operation and capacity to process a large amount of polluted water [18]. To promote arsenite removal efficiency, pre-treatment processes generally involve a pre-oxidation of arsenite to arsenate. Many oxidation technologies, including the addition of conventional oxidants [19] or oxygen and/or ozone [20], electrochemical or biological oxidation [21]–[23], Fenton process [24], and photocatalytic oxidation [25], have been widely explored for the oxidation of arsenite. However, high cost and complicated requirements operational requirements have limited the application of these approaches. An alternative technique which involves the use of an adsorbent has been studied to remove arsenic species. Recent progress in porous materials has led to an increasing attention in the application of these adsorbents for arsenic removal. The effectiveness of nanostructured Mn-Ce binary oxide [26], zirconia nanoparticles [4], hydroxyl iron modified montmorillonite nanoclay [27] in removing arsenate and arsenite have been reported.
It is well-known that zirconia nanoparticles possesses ion exchange property and the affinity of Zr(IV) species for several oxo-anions, including selenite, selenite, arsenate and arsenite has been well reported [28][29][30]. One of the challenges when applying nano ZrO2 in adsorption process is the difficulty in recycling the material as their particle sizes are too small. Fortunately, this can be solved by dispersing them onto a porous matrix. It is reported that nano ZrO2 can be highly dispersed in porous matrix such as multiwall carbon nanotube [31] for arsenic removal and alumina [32] for methylene blue and congo red removal, and reduced graphene oxide [33] for fluoride removals. Activated carbon (AC) is a low-cost material with unique properties such as desired surface functionalization, high porosity, and high surface area. Because of that, AC can be diversely employed to effectively adsorb and remove pollutants [34]. The materials based on the combination of nano zirconia and activated carbon are expected to have distinguishable properties over individual components.
In the present work, the synthesis of zirconia nanoparticle/activated carbon composite and its use to adsorb arsenite are described. The adsorption of arsenite in terms of kinetics and isotherms is addressed.
2. Experimental
2.1. Materials
The activated carbon (AC) was obtained from Tra Bac Company (Vietnam). Its physico-chemistry properties are listed as follows: density: 530 – 550 Kg m−3, carbon content >90%; particle size: 0.075 – 4.75 mm, specific surface area based on BET model: 930 ± 30m2 g−1, humidity <5 %. Zirconia oxide chlorideoCtahydrate (ZrOCl2.8H2O, India, 99%), arsenic trioxides (Merck, 99%), absolute ethanol (C2H5OH, 99.7%, Merck), ammonia (NH3, Merck, 28 %, d = 0.91 mL−1), hydrogen peroxide (30 %, Merck, 1.44 g.mL−1) were used as received from the manufacturer. 1000 ppm arsenite standard solution was prepared by dissolving 1.304 gram of As2O3 in 10 mL of distilled water containing 10 grams of NaOH. The pH of the solution was adjusted to 5–6 using 6M HCl and then the solution was made up to the mark with distilled water.
2.2. Synthesis of ZrO2 and ZrO2-activated carbon (ZrO2/AC)
1.0 M Zr(IV) solution was prepared by mixing 178.85 gram of ZrOCl2.8H2O in absolute ethanol in a 500 mL volumetric flask. This mixture was shaken until it was uniform and made up to the mark with ethanol. This 1.0 M Zr(IV) solution was diluted using ethanol to the desired concentration.
The ZrO2/AC was synthesized by a hydrothermal process. 3 mL of NH3 (28%) was added dropwise into the mixture of 3.6 g of AC and 30 mL of 0.4 M Zr(IV) under vigorous stirring for 2 hours, then the suspension was transfered into Teflon autoclave for hydrothermal treatment for 24 hours at 180°C. The solid was separated by centrifugation and dried at 100°C for 24 hours and calcined at 300°C for 3 hours under nitrogen flow to obtain the ZrO2/AC composite. The ZrO2 nanoparticles were also synthesized in a similar manner without the addition of AC.
2.3. Instruments
Thermal gravimetric analysis (TGA) and differential scanning calorimetry (DSC) was conducted on LABSYS evo DTA/DSC instrument. X-ray diffraction (XRD) was recorded using D8-Advance 5005 with CuK = 0.154 nm. Transmission electron microscopy (TEM) was performed using JEOL JEM-1010 Electron Microscope and scanning electron microscopy (SEM) was studied using Hitachi S-4800. X-ray photoelectron spectroscopy (XPS) was performed using ESCALAB 220iXL (Thermo Fisher Scientific) with Al Kα monochromatic. FTIR spectra were recorded with an Infrared Affinity-1S spectrophotometer (Shimadzu). Nitrogen adsorption/desorption isotherms were measured using TriStar 3000 V6.07A. The specific surface area was calculated by BET model in the relative pressure of 0.01 – 0.25. The concentration of arsenic species was determined by using AGILENT 7700x LC-ICP-MS.
2.4. Adsorption Study
A flask containing a mixture of ZrO2/AC (m = 0.5 g) and 50 mL arsenite solution (Co = 10 – 600 ppm) is shaken for 24 hours. The arsenite-ZrO2/AC suspensions were centrifuged. The supernatants were collected by filtering through 0.2 m-membrane and the concentrations of arsenite (Ce) in the supernatants were determined by ICP-MS. The equilibrium adsorption capacity (qe) is calculated as follows:
2.5. Kinetics Studies
Kinetic studies were conducted by adding 0.5 mg of ZrO2/AC (m) to 50 mL of 10 ppm arsenite (Co) under vigorous stirring. After a certain time interval, 1 mL of suspension was drawn and was centrifuged to separate the solid. The concentration of arsenite in the supernatant (Ct) was determined by ICP-MS. The adsorption capacity at time t (qt) is expressed as follows:
2.6. Recyclable Study
1.5 g of ZrO2/AC (m1) was added in a 250 mL flask containing 150 mL (
) and this mixture was stirred magnetically for 360 min to ensure equilibrium adsorption and separate the solids from solution. The remaining arsenite concentration in the supernatant (
) was determined by ICP-MS method. The used adsorbent was then added into 100 ml of 0.5 M NaOH solution with vigorous stirring for 180 minutes to completely desorb arsenic species. Finally, the adsorbent was washed with distilled water to neutral medium, soaked for 30 minutes in 0.01 M HCl solution before being dried at 100°C for 24 hours to obtain the used ZrO2/AC.
In a similar manner, 1.2 g (m2), 0.9 g (m3), 0.6 g (m4) of recovered ZrO2/AC material after each consecutive run was subsequentially employed to adsorb arsenite in 120 mL, 90 mL and 60 mL solution (
), respectively, in the next run. The concentration of arsenite in the supernatant (
, i: the number of time) was determined. The adsorption capacity of each adsorbent is obtained by the expression:
where i =1,2,3 and 4 is the number of times used.
3. Result and Discusion
3.1. Characterization of Materials
The thermal analysis for ZrO2/AC is presented in Fig. 1a where two weight losses at around 100 and 485°C can be observed. The first loss of ~17 % at around 25 – 160°C accompanied with an endothermic peak can be assigned to the loss of volatile compounds and the physicowater on the surface of the sample. The second one of ~73% with exothermic peak at around 300°C – 750°C is due to the combustion of AC. The remaining 10% mass of solid can be atributed to the ZrO2 content in the composite. The XRD patterns of AC, ZrO2 and ZrO2/AC are illustrated in Fig. 1b. XRD pattern of AC presents two broad bands at 25 and 43° indicating the graphitic hexagonal structure of the carbon [35].
The XRD pattern of ZrO2 exhibits the characteristic peaks of t-ZrO2 at 2θ = 50 and 59° corresponding to diffraction planes (112), and (121) (JCPDS No. 81-1544) and m-ZrO2 at 2θ = 26, 29, 32, 36, 41, 45 và 55° corresponding to (110), (−111), (111), (002), (211), (202) and (221) planes (JCPDS No. 81-1314). XRD pattern of ZrO2/AC retains the characteristic pattern of AC but with lower diffraction intensity. However, the characteristic peaks of ZrO2 cannot be observed in ZrO2/AC XRD pattern which can be due to its low mass percentage in the composite as well as its structural disorder during composite formation. The FTIR spectrum of ZrO2/AC is presented in Fig. 1c. Clear absorptions at 3420, 1626 and 1557 cm−1 correspond to the stretching (▪ (−OH)) and bending (δ (−OH)) vibrations of absorbed water or coordinated molecules on the ZrO2 surface [36]. The vibrations peaks at 3420; 2158; 2020 and 1971 cm−1 can be assigned to CO2 molecules in the pores of the composite [37]. The interaction between the oxygen-containing groups of AC and Zr(IV) can lead to the formation of either a monodentate or bidentate complex [38] which is confirmed by the appearance of a vibrational at 1460 cm−1. The Zr-O vibration of zirconia shows the absorption peaks at 505 and 561 cm−1 [39]. These results reconfirm the interaction between AC and ZrO2 in the composite.
Macropores of around several micrometres embroided with micropore in AC are visible in SEM observation (Fig. 2a). This is favorable for accessible absorbent. The zirconia nanoparticles of around 50–100 nm are observed in Fig. 2b. Meanwhile, the zirconia nanoparticles are found to be highly dispered on the surface of AC in the composite (Fig. 2c).
The specific surface areas of the obtained materials were determined via nitrogen adsorption/desorption isotherms in the relative pressure of 0.01 – 0.25 (Fig. S1 and Table S1). The specific surface area of AC, ZrO2 and ZrO2/AC were found to be 809; 170 and 549 m2.g−1, respectively. The surface area of AC decreased with the introduction of zirconia, due to the accumulation of zirconia particles in the AC pores. However, the ZrO2/AC composite exhibits the compatible surface area compared with several other porous materials.
3.2. Adsorption of Arsenite on ZrO2/AC
3.2.1. Arsenite adsorption on different adsorbents
The adsorption kinetics of arsenite at an initial concentration of 10 ppm on AC, ZrO2 and ZrO2/AC are presented in Fig. 3. AC material shows negligible adsorption kinetics while nano ZrO2 exhibits significantly fast kinetics and high adsorption capacity. However nano ZrO2 in the suspension form is hard to filter and recycle. Whereas, although not exhibiting fast adsorption kinetics as pristine ZrO2, ZrO2/AC still shows high adsorption capacity and is easy to reuse. Therefore, ZrO2/AC was used for further experiments.
The adsorption of arsenite onto ZrO2/AC surface was further studied by XPS spectra. The spectra of ZrO2/AC before arsenic adsorption are shown in Fig. S2a including XPS spectrum of O(1s), C (1s) and Zr(3d) of the material (Fig. S2a1, S2a2, S2a3 and S2a4, respectively). Fig. S2b is the XPS spectrum of ZrO2/AC after arsenic adsorption including the spectrum of O(1s), C(1s), Zr(3d) and As(3d) (Fig. 2b1, 2b2, 2b3, 2b4 and 2b5, respectively). The XPS spectrum of O(1s) of the initial material was separated into 2 peaks with binding energies of 528.32 and 530.96 eV with peak area percentages of 48% and 52%, which are assigned to the binding energies of the oxygen in the metal oxide Zr-O and Zr-OH, respectively. After arsenic adsorption, the XPS spectrum of O(1s) (Fig. S2b2) exhibits a single peak at the intermediate binding energy of 530.83 eV due to the formation of new Zr-O-As bonds upon arsenic adsorption [40][41]. The XPS spectrum of C(1s) of the adsorbent (Fig. S2a3) includes 3 peaks with corresponding binding energies of 284.31; 285.24 and 288.39 eV which are attributed to C(1s) in C=C bond, C-O (C=O; C-OH) and C-O-Zr in the carbonyl group COOH or Zr-O-C= O, respectively. The presence of these functional groups, which play an important role in forming bonds with metal ions, is predictable with the activation of AC [37], [42]. After arsenic adsorption, the XPS spectrum of C(1s) (Fig. S2b3) still shows 3 main peaks with a slight shift of the binding energy to 284.85, 285.71, and 289.01 eV, respectively. This may be due to the adsorption of arsenic on the surface leading to new bonds such as C-O-As or As-O-C=O. XPS spectrum of Zr(3d) of the non-arsenic material consists of two double peaks at binding energy levels of 179.55, 181.95; and 182.48, 184.88 eV which are typical for the binding energy of Zr in the Zr-O and Zr-OH bonds [40], [41]. After arsenic adsorption, the XPS spectrum of Zr(3d) (Fig. S2b4) still exhibits two double peaks, yet at higher binding energy levels of 180.84; 182.21; 182.83 and 185.29 eV. This is due to the conversion of the Zr-OH bond to the new Zr-O-As(III) and Zr-O-As(V) bonds corresponding to the peaks at 182.83 eV and 185.19 eV with a comparable area percentage (32.7% and 20.3%, respectively). The XPS spectrum of As(3d) (Fig. S2b5) of the material after arsenic adsorption includes two peaks at 44.36 and 45.06 eV with the peak’s area percentages of 60 and 40%, respectively. According to the previous studies [43], [44], these energy levels correspond to the energies of As(III) and As(V) in the Zr-O-As(III) and Zr-O-As bonds (V). This result is consistent with the XPS spectrum of Zr(3d) in terms of area percentages of As(III) and As(V) in the respective bonds. These results also confirm the simultaneous adsorption of both As(V) and As(III) on the surface of ZrO2/AC materials. This can be explained when considering the oxidation of a part of As(III) to As(V) during the experiment, leading to the simultaneous adsorption of As(III) and As(V).
The effect of pH on arsenite adsorption is presented in Fig. 3b. With increasing solution pH, the adsorption efficiency of arsenite onto ZrO2/AC increases to a maximum at pH ~5, and then decreases at higher pH. The pH of point of zero charged (pHPZC) based the pH drift method [45] was found to be 5.4 (Fig. 3c). The three acid–base dissociation constant (pKa) values of arsenite are 9.22, 12.13 and 13.46 [46]. The As(III) species can be in the forms of H3AsO3, H2AsO3−, HAsO32−, and AsO33−, which is dependent on the solution pH. The neutral form H3AsO3 mainly exists at pH < 9.22 while arsenite oxoanions are predominant at pH > 9.22. One would expect that the pH-dependent adsorption of arsenite onto ZrO2/AC would be in agreement with that of metal hydroxides if the process is predominantly controlled by electrostatic interaction. Thus, the abovementioned discrepancy suggests that the initial increase of arsenite adsorption at pH 2–5 is likely resulted from other non-electrostatic interactions. At pH > 5, the negative charges of ZrO2/AC and the fraction of arsenite oxoanions increase, resulting in an enhanced electrostatic repulsion between arsenite and the ZrO2/AC. Thus, the flux of arsenite adsorption toward the surface of ZrO2/AC would decrease. The pHPZC of the adsorbent in As(III) solution is also shown in Fig. 3c. The pHPZC of ZrO2/AC after arsenite adsorption shifts to low pH indicating that high adsorption efficiency of ZrO2/AC under neutral and weak acidic/alkaline conditions could be attributed to the complexation among H2AsO3− and H3AsO3 and surface hydroxyl groups of the ZrO2/AC material. The formation of outer-sphere surface complexes could not shift the point of zero charge of absorbent because there is no specific chemical reaction between the arsenite and ZrO2/AC that could change the surface charge. The shift of point of zero charge to a lower pH range is an evidence of the formation of anionic negatively charged surface complexes [44]. Hence the decrease to 4.8 in pHPZC implies that the adsorption of arsenic could be a result of the formation of both outer-sphere complexes and the negatively charged inner-sphere complexes between arsenite and adsorbent [47].
3.2.2. Adsorption kinetics
It could be noted that the adsorption efficiency increases rapidly at the initial stage then it slows down before reaching equilibrium state. The fast adsorption at initial stage might be owing to the availability of higher numbers of active adsorption sites at early stage. Further adsorption leads to a drop in the number of available active adsorption sites, slowing down the adsorption process. Moreover, it is hard to occupy the remaining active adsorption sites because of repulsive interactions among surface molecules on ZrO2/AC and the bulk solution.
The three well-known models, i.e. pseudo-first-order [48], second order kinetic models [49] and intraparticle diffusion model [50] were employed to evaluate the adsorption kinetics on ZrO2/AC as described below.
where qe and qt (ppm) are the amount of arsenite adsorbed at equilibrium and time t, respectively, k1 and k2 (min−1 and ppm−1 min−1) are the equilibrium rate constants of these models, respectively. Kd is the intraparticle diffusion constant (mg.g−1 min−1/2) and xi (mg. g−1) a constant related to diffusion resistance.
The kinetics parameters were calculated by fitting the experimental data in non-linear form. The results show that the pseudo-first order kinetic model fits better with higher determination coefficients (R2).
The intraparticle diffusion model (the Weber-Morris’s model) was applied for diffusion kinetics analysis. The values of model parameters were calculated from the slopes of the obtained linear plots as given in Table 1. The plot of qt versus square root of time t0.5 illustrates a significantly linear relationship for the adsorption of arsenite (R2 = 0.813 and p =0.0008 <0.05) (Fig. 4b). The intra-particle diffusion rate constant (kid) has a value of 0.1573 mg g−1 min−0.5 for arsenite adsorption on ZrO2/AC. Also, the linear relationship depicts macropore diffusion process. Furthermore, such zero intercept indicates that the intra-particle diffusion is the sole rate-controlling step.
3.2.3. Adsorption equilibrium
The adsorption equilibrium is one of the critical physico-chemical aspects in the description of the adsorption behavior for solid-liquid system. Two well-known model with two parameters is used in this work, e.g. Langmuir and Freudlich isotherm models
where qmom is a maximum monolayer adsorption capacity (mg.g−1).
Adsorption data fits well to either Langmuir or Freudlich isotherm (R2 > 0.999) (Fig. 4c, Table 1) indicating that the adsorption processes involves both monolayer adsorption and heterogeneous interaction. The fitting data shows that the maximum monolayer adsorption capacity (qmax) for arsenite on ZrO2/AC is 64 mg.g−1. In consideration of the adsorption capacities for arsenite species, ZrO2/AC appears to be compatible or better than other adsorbents as shown in Table 2 making it a promising adsorbent for arsenic treatment in aqueous solution.
3.2.4. Effect of coexisting ions
The anions such as fluoride, nitrate, phosphate and sulfate were used as interferents to study their effect on arsenite adsorption (Fig. 4d). The result shows that the arsenite adsorption was not affected by the coexisting ions in this study except for PO43−. At a phosphate concentration of 300-fold higher, the arsenite adsorption efficiency decreases significantly which might due to the competitive adsorption of phosphate and arsenite onto the surface through inner-sphere complexation [55].
3.2.5. Recyclability
The regeneration of adsorbent was carried out by treating with 0.5 M NaOH followed by 0.1 M HCl. Adsorption efficiency decreases slightly from 3 % for the first reuse, 4 % for second reuse and 6 % for the third reuse (Fig. 5a). Fig. 5b presents XRD patterns of ZrO2/AC after several usages. It is found that its patterns are unchangeable indicating a stable structure of ZrO2/AC after three recycles. ZrO2/AC with the high adsorption capacity and stability is a promising adsorbent for arsenite removal.
4. Conclusions
ZrO2/activated carbon composite has been synthesized by hydrothermal process. Zirconia nanoparticles are highly dispersed on the carbon matrix to form a composite with high surface area. The obtained composite exhibits an excellent adsorption capacity towards arsenite in aqueous solution. Both Langmuir and Freundlich models could be successfully applied to describe the adsorption process where both monolayer adsorption and heterogeneous interaction are possible. The anions such NO3−, Cl−, and SO42− do not significantly affect arsenite adsorption capacity. Phosphate was a powerful competitor over arsenic for adsorptive sites on the adsorbent. The shift of PZC of adsorbent to a lower value after adsorption of arsenic indicates that the formation of outer- or inner-sphere complexes between arsenic and adsorbent may be involved. The used adsorbent could be recycled easily with NaOH solution. The adsorption efficiency decreased by about 5 % after recycling. The results from this study suggest that ZrO2/activated carbon composite can act as excellent adsorbent towards arsenite.
PDF Links
PubReader
Full text via DOI
Download Citation 
Supplement
Print 
