Equilibrium and kinetics study on removal of arsenate ions from aqueous solution by CTAB/TiO₂ and starch/CTAB/TiO₂ nanoparticles: a comparative study

ATC

arsenic-treated SPC

CMT

CTAB-modified TiO₂

CTAB

N-cetyl-N,N,N-trimethyl ammonium bromide

SPC

starch-coated CMT

UTi

unmodified TiO₂

TiO₂ is currently considered as the most versatile nanomaterial. It has a wide range of applications in different fields (Xiaobo & Samuel 2007). These are attributed to its low cost, low toxicity, photostability in solution, redox selectivity and strong oxidizing power of holes. The application of TiO₂ at the nanoscale further improves efficiency due to the high surface area, large pore volume and presence of high affinity surface hydroxyl groups (Daimei et al. 2006; Gupta et al. 2012). TiO₂ and its different modified forms have also drawn the attention of researchers in the field of the removal of contaminants from water. It is capable of removing heavy metals as well as arsenite and arsenate from contaminated water (Gautham et al. 2010).

Gupta & Tripathi (2011) reviewed the use of TiO₂ for environmental clean-up of organic pollutants through photo-oxidation. Xu et al. (2007) studied the photocatalytic degradation of organic arsenic to inorganic form by TiO₂ and successful sorption on its surface.

Debnath & Ghosh (2009) applied nanostructured hydrous titanium oxide to adsorb Ni(II) from natural and industrial wastewater. The sorption of Se(IV) on TiO₂ (anatase) was investigated by Keliang et al. (2009). Ghosh et al. (2003) and Barakat (2005) studied the use hydrous titanium oxide for adsorption of Cr(VI) and Cu(II), respectively.

Xu et al. (2010) prepared hydrous titanium dioxide (TiO_2. xH₂O) by one-step hydrolysis. The adsorption capacity of sorption of As(III) was found to be very high and was pH dependent. The sorption of As(III) and As(V) on commercially available TiO₂ was also found to be dependent on pH as well as initial ion concentration (Dutta et al. 2004). Bang et al. (2005) experimented with a novel granular TiO₂ for removal of groundwater arsenic and found that more arsenate was adsorbed than arsenite on TiO₂ at pH 7.0. Nakajima et al. (2005) investigated the use of TiO₂ photocatalyst and adsorbent for removal of As(V) under photo-irradiation.

Barakat (2005) studied photocatalytic degradation using UV-irradiated TiO₂ suspension for destroying complex cyanide with a simultaneous removal of copper. Kajitvichyanukul et al. (2005) prepared TiO₂ thin films immobilized on glass plates for removal of Cr(VI).

Heterogeneous photocatalytic oxidation of arsenite to arsenate in aqueous TiO₂ suspensions has also been proved recently to be an effective and environmentally acceptable technique for the remediation of arsenite-contaminated water. Zhang & Itoh (2006) performed a process using an adsorbent developed by loading iron oxide and TiO₂ on municipal solid waste melted slag for oxidation of arsenite to arsenate.

However, the dispersibility, reusability and leaching of particles are a major concern for the use of TiO₂ in water purification. The nanoparticles have very high surface energy. Therefore, they tend to agglomerate so that they can reduce their surface energy by decreasing the surface area and thus stabilize (Gupta et al. 2012). Certain modifications are necessary to overcome these limitations. Adequate surface functionalization and coating with polymeric materials, for example, are some of the promising aspects regarding selectivity and aqueous stability of this material.

Use of surfactant-like CTAB increases the stability of TiO₂ nanoparticles in water by forming micelles around them and thus can enhance the dispersibility of the particles in aqueous medium (Deka & Maji 2011). CTAB also contains a hydrophilic end (Miao et al. 2014), which might enhance the solubility of modified TiO₂ particles in aqueous medium. Thus some particles may get leached in to water during wastewater treatment. This difficulty may be avoided by the use of supporting materials or by coating the nanoparticles with polymeric materials.

Starch, one of the cheapest and nontoxic polymers, is abundantly available in nature. Starch and starched-based materials used as bio-compatible materials have been cited in the literature (Marques et al. 2002; Saikia et al. 2010). He & Zhao (2005) have reported the advantage of coating starch onto bimetallic nanoparticles (Fe-Pd) for degradation of chlorinated hydrocarbons in water. A significant improvement in arsenic removal rate from brine by using starch-coated magnetite nanoparticles has been reported (An et al. 2011). Stabilization with starch causes less agglomeration and improves removal rate compared to the bare nanoparticles. The processing of starch for stabilizing nanomaterials is quite easy. Starch can be easily solubilized in hot water and re-precipitated by simply adding acetone or ethanol. The coating of the nanomaterial by starch can be done in aqueous medium without using any other toxic chemicals.

The aim of the present work is to prepare CTAB-modified TiO₂ nanoparticles and stabilize them by coating with starch, and finally use the particles in the treatment of As(V)-contaminated water. A comparative study was made using unmodified TiO₂ (UTi), CTAB-modified TiO₂ (CMT) and starch-coated CMT (SPC). Low concentrations of arsenic solutions were used in the experiments considering the practical applicability of the material.

Analytical grade nanosized TiO₂ (mixture of rutile and anatase), potassium sulfate (K₂SO₄), and sodium arsenate dibasic heptahydrate (HAsNa₂O₄.7H₂O) were purchased from Sigma-Aldrich. Starch polymer was purchased from SRL, India. Gluteraldehyde solution (25%), used as cross-linking agent, and di-sodium hydrogen phophate (Na₂HPO₄) were purchased from E-Merck, India. CTAB (N-cetyl-N,N,N-trimethyl ammonium bromide) was purchased from LOBA Chemie, India. Ethanol was supplied by Jiangsu Huaxi International Trade Co. Ltd, China. Deionized water was obtained from Milli-Q water purification system (Millipores.A.S.67 120 Molsheim, France). All other reagents and chemicals used were of analytical grade.

Coating of starch over TiO₂ particles was carried out in two steps following the method reported by Deka & Maji (2011). Initially TiO₂ was modified with CTAB and then was coated with starch.

One gram of TiO₂ was taken in a round-bottomed flask containing 1:1 ethanol–water mixture. The flask was fitted with a spiral condenser and kept at 80 °C for 12 h under magnetic stirring. In a similar set up, 1.2 g of CTAB was refluxed with 1:1 ethanol–water mixture at 80 °C for 3 h. This mixture was poured into the TiO₂ solution mixture and was stirred again for another 6 h at 80 °C. It was then filtered and washed with lukewarm deionized water several times and oven-dried at 45 °C. Finally, the material was ground and stored in a desiccator to avoid moisture absorption.

One gram of CMT was dispersed in a 250 mL beaker under continuous magnetic stirring for 1 h and then sonicated for 30 min to completely disperse the CMT nanoparticles in the solvent.

In the meantime, a solution of starch was prepared by dissolving 0.5 g starch in 50 mL of hot water (80 °C). The CMT solution was then mixed with the starch solution and stirred for 1 h. Starch molecules typically have many –OH groups on them, which are polar. These –OH groups interact with the polar ends of CTAB and thus form a coating around them. The mixture was then poured slowly into acetone under stirring to precipitate the SPC particles.

The temperature of the beaker was brought down below room temperature, then glutaraldehyde solution (5% w/w) was added to the mixture containing SPC particles under magnetic stirring. The temperature was again raised gradually to 50 °C and the reaction was continued for another 2 h under continuous stirring condition. The temperature was then decreased to room temperature. The nanoparticles thus formed were filtered, washed with deionized water, dried and kept ready for subsequent applications.

Fourier transform infrared spectrophotometry (FT-IR) (NICOLET Impact I-410) was used to study the probable interaction between different component materials of the prepared nanoparticles as well as As(V) binding to the adsorbent. The study was carried out in the range 500–4,000 cm⁻¹.

The distribution of the clay minerals within the polymer matrix in the nanoparticle was characterized by powder X-ray diffractometry (XRD) using a Miniflex instrument (Rigaku, UK). The measurement conditions were: anode material = Cu; K-alpha, λ = 1.54 Å and scanning rate 2°/min with an angle ranging from 2° to 80°.

A scanning electron microscope (SEM) (JEOL 6390LV) was used to study the presence of elements as well as morphology of the samples before and after As(V) treatment.

The pH of the arsenic solutions was adjusted using NaOH (0.1 M) and/or HCl (0.1 M) solutions as required and measured by Cyberscan pH 510 (Eutech) instrument.

The As(V) concentration was measured by using an Analyst 200 Atomic Absorption Spectrophotometer (Perkin Elmer). All the measurements were based on integrated absorbance and performed at 193.7 nm by using a quartz tube analyser (Perkin Elmer) at an atomization temperature of 2,000 K.

A stock solution of 1,000 mg/L As(V) was prepared by dissolving the requisite amount of sodium arsenate (Na₃AsO₄) in 1,000 mL deionized water. The stock solution was diluted in steps to obtain standard solutions containing 0.1–0.7 mg/L of arsenic solution. The material dose and time for removal of As(V) was optimized by using the solution containing 400 μg As(V)/L. Batch adsorption studies were carried out with 2 g of adsorbent per litre of solution (as determined previously) of desired concentration at neutral pH in 100 mL conical flasks, which were agitated at 160 rpm for predetermined time intervals at room temperature on a mechanical shaker. After agitation, the suspension was centrifuged and filtered, and the arsenic concentration left in the aliquot was determined by atomic absorption spectrophotometry. A comparative study on the efficacy of UTi, CMT and SPC for As(V) removal was done.

The kinetic study for the adsorption of As(V) was conducted at neutral pH. In order to investigate the sorption mechanism for the removal process, pseudo first and second order kinetic models were used in different experimental conditions. The adsorption of As(V) was carried out for 90 min with initial As(V) concentration 0.1–0.7 mg/L for a material dose of 2 g/L. The kinetic parameters for sorption of As(V) for both CMT and SPC were calculated by using Lagergren's equations (Bhatt et al. 2012).

Freundlich, Langmuir and Temkin isotherm models were applied for the study of equilibrium adsorption of As(V) on CMT and SPC.

The reusability test was performed for CMT and SPC four times. An As(V) solution with an initial concentration of 400 μg/L was treated for 90 min with 2 g/L of CMT and SPC, respectively. It was then filtered and the concentration of As(V) in the filtrate was measured to determine the removal efficiency of the materials (CMT or SPC). The material left as residue was washed with dilute HCl acid for desorption of the As(V) ions adsorbed on its surface. It was then dried, weighed and used for further study to access the removal efficiency against the As(V) solution with same initial concentration. The process was repeated under the same conditions.

In the spectrum of CTAB, shown in Figure 1(A(b)), the absorption bands at 2,916 and 2,848 cm⁻¹, respectively, are attributable to asymmetric and symmetric modes of –CH₂ stretching in the methylene chains (Kungt & Hayes 1993). Band near 1,474 cm⁻¹ may be due to symmetric stretching of CH₃⁺-N bond (Thakur 2015).

Bands near 2,916 cm⁻¹, 2,850 cm⁻¹ and 1,463 cm⁻¹ (corresponding to CTAB) along with bands at 1,636 cm⁻¹ and near 530 cm⁻¹ (corresponding to TiO₂) in the CMT spectrum indicate the probable interactions between the individual components.

Figure 1(B) shows the FT-IR spectra of starch, SPC and arsenic-treated SPC (ATC). The spectrum for starch (Figure 1(B(a))) shows a broad band near 3,440 cm⁻¹ for –OH stretching, 2,914 cm⁻¹ for –CH stretching and 1,656 cm⁻¹ for –OH bending. The two peaks in the region 1,183–966 cm⁻¹ are attributable to C–O bond stretching of C–O–H group (Kousalya et al. 2010). Incorporation of CMT into starch can be well predicted from the curve (b). Increase in intensity of bands corresponding to –OH stretching and –OH bending as well as appearance of a new band near 1,410 cm⁻¹ (CH₃–N⁺ of CMT) is an indication for interaction between the two components.

In the FT-IR spectrum of ATC (Figure 1(B(c))), it was observed that the intensity of bands assigned to the hydroxyl groups and C–N stretching in SPC decreased prominently. This might be due to the participation of those groups in arsenic binding. The band appearing near 632 cm⁻¹ is attributable to the adsorbed arsenate ions on to the material (Sham et al. 1998).

CMT nanoparticles (Figure 2(b)) became less agglomerated than those of unmodified TiO₂ (Figure 2(a)) due to the presence of long hydrocarbon chains of CTAB which prevented them from agglomerating. Jiang et al. (2013) studied the morphology of CTAB-modified Ni nanoparticles and observed that modification with CTAB resulted in irregular sphere-like structures which were less aggregated than Ni-only nanoparticles. ATC (Figure 2(d)) composite appeared slightly rough compared to SPC (Figure 2(c)) composites. This might be due to sorbed arsenic on the composite structure. The surface of arsenic desorbed (i.e. reused) composite (Figure 2(e)) was rougher than SPC (Figure 2(c)). This might be attributed to the disintegration of the surface of the composite caused by the repeated washing with acid during sorption-desorption treatment. Similarly the shape of CMT nanoparticles (Figure 2(f)) became irregular and rough due to the acid treatment.

The relative content of carbon and oxygen was greater in SPC (Figure 3(d)) than CMT (Figure 3(c)) suggesting the participation of starch polymer into the system. Presence of arsenic in the spectra of both arsenic adsorbed CMT and SPC confirmed the successful sorption of As(V) onto the systems. This further supports the findings of FT-IR data as discussed earlier (Figure 1(d)).

The TGA curve for CMT (Figure 4(A(c))) shows some weight loss in the temperature range ∼220–320 °C; UTi (Figure 4(A(b))) shows hardly any weight loss in that region. The weight loss in CMT is due to the incorporated CTAB on the surface of TiO₂ nanoparticles. Coating of CMT with starch in SPC showed much more weight loss as compared to CMT. But at the same time the weight loss was less than the pristine polymer. This indicated that interaction had taken place between CMT and starch, and starch had successfully coated the surface of CMT.

No separate degradation patterns were observed for the TGA curve of SPC indicating that the mixing of starch and CMT nanoparticles did not form a blend. Rather, starch properly interacted with CMT nanoparticles.

The XRD spectra of UTi, CMT and SPC are shown in Figure 4(B). The XRD patterns of UTi (Figure 4(B(b))) exhibit strong diffraction peaks at approximately 2θ = 25 ° (101), 37 ° (111), 47 ° (210), 53 ° (211) and 55 ° (220). The peaks occur due to the crystalline portion of the TiO₂. Similar findings were reported by Juang & Shiau (2000) and Wang et al. (2008). The diffraction peak at 21° is for crystallinity of starch.

In the diffractogram of CMT (Figure 4(B(c))) and SPC (Figure 4(B(d))), no remarkable change in either peak positions/or peak intensity for TiO₂ crystals can be observed. This suggests that the starch coating process did not result in any phase change of TiO₂ particles.

CMT particles exhibited better removal efficiency than UTi particles and brought down the arsenic concentration below 10 μg/L. CTAB prevented the aggregation of UTi particles and thus exposing more surfaces for interaction with arsenate ions in solution. However, part of the CMT entered into the aqueous phase during the experiment causing a significant loss of the material. This could be reduced by coating with crosslinked starch (SPC). Hence both CMT and SPC were used in the subsequent experiments.

It was found that for a solution of concentration 400 μg/L, a dose of 2 g/L of CMT was sufficient to bring down the As(V) concentration below 10 μg/L. No prominent change was observed with the increase of material dosage thereafter. Therefore, the material dose was fixed at 2 g/L for further study. Figure 6(a) shows that CMT has better efficiency compared to SPC for As(V) removal. However, reusability of CMT is a major problem as discussed later whereas starch coating is a good alternative.

Figure 6(b) shows the effect of agitation time on percentage removal of the As(V) ion. With increase in treatment time, the percentage removal of As(V) reached maximum (99% for CMT and 83% for SPC) after 90 min treatment and after that remained almost constant throughout the time period studied. This might be due to the rapid achievement of equilibrium between the arsenic adsorbed on the surface of sorbent and arsenic present in solution. Therefore, the agitation time was kept fixed at 90 min for the subsequent batch experiments.

The effect of initial arsenic(V) ion concentration was studied in the concentration range 0.1 to 0.7 mg/L. Material dose and agitation time were kept fixed as optimized before. Figure 6(c) shows the effect of As(V) concentration on percentage removal of the ion. With the increase in metal ion concentration, the amount of metal ion uptake per unit weight of adsorbent (mg/g) had also increased. It was observed that the percentage removal of As(V) decreased with the increase in initial metal ion concentration. This might be attributed to a delayed establishment of equilibrium between the adsorbent sites and the As(V) ions. The saturation of sorbent sites with the increase in initial ion concentration might also be the reason for diminished removal rate.

Similarly the different pseudo second order kinetics parameters were determined from the linear plot of t/q_t vs t. The correlation coefficients (R²) were computed and all the values are shown in Table 2. The q_e values were found to increase with the increase in the initial As(V) concentrations. The curve for pseudo second order kinetics exhibited higher correlation co-efficient compared to pseudo first order model.

The Langmuir parameters viz. q_m, b and R² values were found to be 1.024 mg/g, 1.31 × 10² and 0.989, respectively, for SPC and 1.423 mg/g, 10.82 and 0.998, respectively, for CMT.

In Table 3, it can be seen that the sorption was favourable for both materials as R_L values (0.0187 for SPC and 0.187 for CMT) occurred between 0 and 1 for each case. Also q_e vs C_e showed a linear relationship.

Freundlich parameters R² and n values, as shown in Table 3, also indicated good fitting of sorption data with the model. The Temkin parameters were also derived and the correlation coefficient (R² ∼ 0.987 and 0.984) values showed that Temkin model could well explain the sorption pattern.

It was observed that the R² values for all the models showed good correlations of the data. Thus the sorption could be reasonably explained by all the models under consideration. The preferential order for sorption, among these models is as follows: Langmuir > Temkin > Freundlich.

A comparison of the maximum adsorption capacity (q_o) calculated from the Langmuir model with some other established materials is shown in Table 4. Ferruginous manganese ore has a high adsorption capacity as shown in Table 4, but arsenic does not get desorbed easily even on varying the pH (Chakravarty et al. 2002). The material under present investigation showed comparable results.

The weight loss after each time of reuse was also calculated for both CMT and SPC. It was found that the weight loss (%) of material was low (Figure 9(b)) for SPC compared to CMT. This implied the successful stabilization of CMT particles inside the starch matrix resulting in a decreased leaching rate. Use of crosslinking agent also might have some contribution for observing lower weight loss in the case of SPC.

With the increase in their respective concentrations, both phosphate and sulfate ions were found to notably reduce the removal rate of arsenate. This might be due to the similarity in chemistry of As(V) with phosphate and sulfate in aqueous solution at neutral pH. The effect of phosphate is more pronounced (Figure 10) compared to that of sulfate. The resemblance of ionization constants and ionic structure of phosphate with arsenate is responsible for this anomaly (Guan et al. 2012).

FT-IR analysis indicated successful modification of TiO₂ particles with CTAB surfactant. Good interaction of the modified form with starch polymer was also observed. These findings were further supported by TGA, SEM and EDX analysis. XRD data showed that the phase of TiO₂ remained unaltered during the modification or coating process. Successful sorption of As(V) was confirmed by FT-IR, SEM and EDX analysis. Both CMT and SPC were capable of reducing the As(V) concentration from 400 μg/L to less than 10 μg/L.

The equilibrium data were fitted to Langmuir, Freundlich and Temkin adsorption isotherm models. The Langmuir isotherm model was found to be better fitted than the other two models as judged by their equilibrium data. The pseudo second order kinetic model showed better correlation than pseudo first order model for sorption of As(V) on CMT. Competitive ions like phosphate and sulfate reduce the arsenate removal rate by CMT.

Sorption of As(V) occurred in all the materials, i.e., UTi, CMT and SPC. SPC showed the best result when efficiency, reusability and integrity all are taken under consideration.

The authors thank DRL, Tezpur for providing instrumental facilities.