Synthesis of iron oxyhydroxide-coated rice straw (IOC-RS) and its application in arsenic(V) removal from water

Arsenic (As) is a toxic metalloid which can pollute water, land, crops, and the environment at large, ultimately affecting human health (Zhao et al. 2010a, b). A high level of As (>10 mg/L) may cause skin lesions, rhagades, damage mucous membranes, digestive, respiratory, circulatory, nervous system, and moreover it is associated with skin, liver, and lung cancers (Choong et al. 2007; Nguyen et al. 2009; Lizama et al. 2011; Bulut et al. 2014). During the last 20 years, naturally occurring As has been found to be widespread in natural water in many countries around the world, especially in Bangladesh, India (West Bengal), China (including Inner Mongolia), Vietnam, Argentina, and some parts of West Africa due to their exposure to high As drinking water sources (Zhang et al. 2003; Smedley et al. 2007; Pehlivan et al. 2013). Works on ground waters from different parts of Burkina Faso have shown that the Yatenga province population has been exposed to As contaminated water with over 0.5–1,600 mg/L (Smedley et al. 2007; Barro-Traoré et al. 2008).

Arsenic cannot be destroyed in the environment; however, it can be changed to different forms and accumulated in different biota and environmental media. Various technologies available for removal of As from contaminated water are based mainly on six principles: (i) oxidation and filtration; (ii) biological oxidation: oxidation of As(III) to As(V) by micro-organisms and then removal of As(V) by iron and manganese oxides; (iii) co-precipitation: oxidation of As(III) to As(V) by adding suitable oxidizing agent followed by coagulation, sedimentation, and filtration; (iv) adsorption: activated alumina, activated carbon, iron-based sorbents, zero-valent iron, and hydrated iron oxide, etc.; (v) ion-exchange through suitable cation and anion exchange resins; and (vi) membrane technology: reverse osmosis, nanofiltration, and electrodialysis (Mondal et al. 2006; Lizama et al. 2011; Jain & Singh 2012). Among these, a few treatment technologies are efficient but expensive whereas some are cheaper but not efficient. Because of the simplicity of the adsorption process, it has become one of the most promising and applied methods in As removal (Kanel et al. 2006; Bulut et al. 2014). The most commonly employed adsorbents used for removal of As include magnetic iron oxide nanoparticles (Song et al. 2013), activated alumina (Singh & Pant 2004), resin-MnO₂ (Lenoble et al. 2004), TiO₂ (Dutta et al. 2004), nanostructured ZrO₂ spheres (Cui et al. 2013), ores (Chakravarty et al. 2002; Zhang et al. 2004), zeolites (Elizalde-Gonzalez et al. 2001), activated carbon (Kalderis et al. 2008), functionalized graphene (Mishra & Ramaprabhu 2011), p(4-vinylpyridine)-based hydrogels (Sahiner et al. 2011), iron-modified: resin (Rau et al. 2003), sponge (Munoz et al. 2002), and sand (Thirunavukkarasu et al. 2003). Unfortunately, most of these materials are considered as expensive adsorbent in many countries.

During recent years, interest has been primarily focused on the production of low-cost sorbents from agricultural wastes or byproducts. Lignocellulosic materials are evaluated to be very economic precursors for the production of adsorbents that have been used extensively for many purposes of separation and purification. However, the adsorption capacities of lignocellulose for removing As are low and slow (Amin et al. 2006; Elizalde-Gonzalez et al. 2008; Urík et al. 2009; Ranjan et al. 2009; Anirudhan et al. 2012). Therefore, chemical modifications have been employed to synthesize appropriated and efficient adsorbents for removing As ions; these include iron-impregnated: biomass (Aryal et al. 2010), chitosan (Gupta et al. 2009; Gang et al. 2010), bead cellulose (BCF) (Guo & Chen 2005), lignocellulose substrate (iron(III)-LS) (Dupont et al. 2007), sawdust (Urík et al. 2009), and rice husk (RH-FeOOH) (Pehlivan et al. 2013).

Rice straw (RS) is one of the most abundant natural sources in the world. Its annual production is about 731 million tonnes, which is distributed in Asia (667.6 million tonnes), America (37.2 million tonnes), Africa (20.9 million tonnes), Europe (3.9 million tonnes), and Oceania (1.7 million tonnes) (Kim & Dale 2004; Binod et al. 2010; Hsu et al. 2010). It is a kind of lignocellulosic biomass which contains about 32–47% cellulose, 19–27% hemicellulose, and 5–24% lignin (Karimi et al. 2006; Wattanasiriwech et al. 2010). This material is one of the most agriculture wastes listed in Burkina Faso. In this context, we decided to probe the adsorption capabilities of RS-based matrixes toward As(V).

This paper presents the synthesis of an inorganic–organic hybrid adsorbent of iron oxyhydroxide-coated rice straw (IOC-RS) by RS: (i) pretreatment in sulfuric acid and sodium hydroxide solutions; and (ii) impregnation in ferric nitrate and sodium hydroxide solutions. This work focused on investigating how various experimental parameters influence As(V) adsorption. These parameters included: pH, As(V) initial concentration and the interference of dissolved NaNO₃, MgSO₄, and Na₃PO₄.

H₂SO₄ and NaOH used for RS pretreatment were purchased from Merck. Fe(NO₃)₃ procured from Merck was used for biosorbent modification. A Titrisol ampule with As₂O₅ in H₂O used for batch sorption experiments was purchased from Merck, Germany. As(V) stock solution of 1,000 mg/L As(V) was prepared by transferring the Titrisol ampule content with As₂O₅ in H₂O (Merck, Darmstadt, Germany) into a 1 L volumetric flask, which was filled up to the mark at 20 °C and stored at 16 °C. Diluted solutions of As(V) were prepared daily before starting the batch studies. NH₄OH and HCl purchased from Merck, Germany, were used to adjust the pH solutions. MgSO₄·7H₂O (Fluka, Seelze, Germany), Na₃PO₄·12H₂O (Sigma-Aldrich, Seelze, Germany) and NaNO₃ (Merck) were used for studying the effects of ionic strength. KI used for reduction of As(V) to As(III), was obtained from Roth, Germany. NaBH₄ (Fluka, Seelze, Germany), NaOH (Merck, Darmstadt, Germany), and HCl (Merck, Darmstadt, Germany) were used for hydride generation (HG). All reagents used throughout this work were of analytical grade. All glassware was cleaned with diluted nitric acid solution and rinsed with deionized water. Pure water was obtained with a water system filtration (Seralpur, Pro90C).

RS was collected from local farm near the Ouagadougou area, Burkina Faso. The straws were cut into small pieces with lengths of ≤6 cm and then washed three times with tap water. The air-dried material (at 60 °C) was made into powder in an electrical ball mill (BLB Braunschweig) and sieved in an electrical sieving machine (Retsh, West of Germany). The particles of sizes between 0.125 and 0.200 mm were rewashed thoroughly with deionized water to remove the fine particles, and dried in a hot-air oven (at 70 °C) for 24 h. The air-dried and powdered RS (60 g) was successively treated with: (i) 2 mol/L H₂SO₄ (1/1 w/w of dry matter, at 80 °C for 30 min) to remove starch, proteins, and sugars; and (ii) 0.5 mol/L NaOH (ratio straw/sodium hydroxide = 5, stirring for 24 h at 22 °C) to remove the low molecular weight lignin compounds after filtration. The substrate was thoroughly washed, and then the material was air-dried in an oven at 60 °C for 24 h. The dried material was stored in a vacuum desiccator.

Twenty-five grams of the dry acid and alkali-treated RS was soaking with 100 mL deionized water for 2 h. The soaked substrate was crushed in a porcelain mortar to get a solid paste. Coating was performed by reacting the solid pasty material with 800 mL of ferric nitrate Fe(NO₃)₃ solution (0.05 mol/L) in a 2 L flask. Sodium hydroxide (1 mol/L) was slowly added into the mixture, under continuous stirring, until the pH rose to a value between 2.8 and 3.2. The loading process was left for 24 h, and the pH was checked and readjusted with acid (0.1 mol/L HCl) and base (0.1 mol/L NaOH) solutions during the process. The sorbent was then filtered, washed with deionized water several times, dried in an oven at 50 °C for 24 h and stored in desiccators.

The thermogravimetric analysis (TGA) of RS and IOC-RS were analyzed using a SETSYS Evolution TGA 16/18 Setaram Instrumentation. The ramping rate was 10 °C/min up to 900 °C in a nitrogen environment. The scanning electron microscope (SEM) was performed on JEOL JSM-6480 microscope to collect the SEM images of the adsorbents. The Fourier transform infrared spectroscopy (FTIR) was carried out on a Bruker Tensor 27 spectrometer, Diamant attenuated total reflectance. The spectra were recorded in the region of 4,000–520 cm⁻¹ for 32 scans. The specific surface area was determined by N₂ adsorption at −196 °C using an ASAP 2020 Micromeritics instrument and the Brunauer–Emmett–Teller method (Sing et al. 1985). The zeta potential measurement of the sorbent was performed by mixing 0.2 g of sample with 50 mL of deionized water at 22 ± 2 °C. The pH_i (initial pH) values of the solution were adjusted roughly from pH 4 to 10 by a pH meter, via addition of 0.1 M HCl or NH₄OH. The mixture was shacked to the equilibrium time of 4 h. After completion of the equilibration time, the admixture was filtered and the pH_f (final pH) values were measured. The difference between pH_i and pH_f values (ΔpH = pH_f − pH_i) was plotted versus the pH_i. The pH value at the point of zero charge (pH_pzc) of the sorbent was determined from the point of intersection of the resulting curve, at which ΔpH = 0. For As(V) concentrations access, HG attached with atomic absorption spectrometry with a Zeeman correction (AAS-Hitachi Z-2000) was used, with a calibration range from 1 to 20 μg/L. Arsine obtained from As(III) gives the best signal; therefore, all As(V) in the sample was previously reduced to As(III). The reduction was carried out with the optimized protocol: 30% HCl (2.5 mL) and 20% KI (2.5 mL) solution as a reduction agent were added to all sample solutions (10 mL), and a waiting time of 1 h allowed for complete reduction of As(V); then the whole were made up to 25 mL before analysis. For HG, the following solutions were used: (i) 1.2 M HCl; and (ii) NaBH₄–NaOH solution (solute 10 g NaBH₄ in 1 L of H₂O Seralpure by adding 4 g of NaOH); the solution was prepared immediately before use. The principle is to volatilize As from As(III) to arsine AsH₃(g), due to a reaction of nascent hydrogen on As. This nascent hydrogen comes from the decomposition of BH₄⁻ ions.

The effect of solution pH was carried out by adding 0.2 g of sorbent in 50 mL of As(V) solution of 50 mg/L initial concentration at different pH values (2.0–10.0). The mixture was gently agitated with a rotary shaker (Retsch, Berlin, Germany) for 8 h at a temperature of 22 ± 2 °C.

The influence of initial solution concentration on As(V) sorption with sorbent was carried out by adding 0.2 g of sorbent into 50 mL of As(V) solution of 10–300 mg/L initial concentration. The pH values were adjusted at 4 and 6, and the mixtures were shaken for 24 h at 22 ± 2 °C. The As sorption experiments were repeated with solution containing the mixture ions of dissolved NaNO₃, MgSO₄, and Na₃PO₄, usually present in water with As(V) solution of 50 mg/L.

The SEM pictures (Figure 1) reveal the surface textures and porosities of RS and IOC-RS. It was found that after the surface of RS had been coated by the iron oxyhydroxide in the iron reclaim system, the apparent color of RS changed from dark yellow to dark orange (red), and the surface morphology became much finer and smoother. On the other hand, the surface area of the IOC-RS (7.86 m²/g) was also different from that of RS (2.85 m²/g).

RS is a lignocellulosic (cellulose, hemicellulose, and lignin) and silica matrix, which likely consists of alkene, esters, aromatic, ketones, and alcohols with different oxygen-containing functional groups. The FTIR spectrum of RS (Figure 2(a)) showed bands at 897 cm^–1, 1,060 cm^–1, and 1,380 cm^–1, which are ascribed to β-glycosidic linkages, C–O–C stretching and O–H bending, respectively (Sun et al. 2004). The bands observed at 3,332 cm^–1 and 1,735 cm⁻¹ can be assigned to O–H stretching and C=O stretching, respectively. The presence of C=C groups is reflected by vibration bands at 1,606 and 1,515 cm⁻¹ due to aromatic rings of lignins (Fernandez-Bolanos et al. 1999). On the other hand, the adsorption peaks at 1,645 cm⁻¹ and 796 cm⁻¹ are, respectively, assigned to the H–O–H bending mode (Liu et al. 2006) and the symmetric Si–O–Si stretching vibration (Farook et al. 2006). The loaded RS spectrum (Figure 2(b)) shows some changes mainly in the decrease of the peak intensities at 1,735 and 1,645 cm⁻¹, which can be explained by the complexation of Fe³⁺ ions with carboxylate and hydroxyl groups in the matrix (Pehlivan et al. 2013). The decrease of peaks observed at 1,606, 1,515, and 796 cm⁻¹ is due to a partial removal of lignins and silica during the pretreatment with NaOH (Sun et al. 2004; Liu et al. 2006). The strong band between 600 and 520 cm⁻¹ belongs to the stretching mode Fe–O (Correa et al. 2010; Bordoloi et al. 2013; Bulut et al. 2014).

The thermal degradation (Figure 3) of RS and IOC-RS showed that there was a similar weight loss between 150 and 550 °C due to degradation of lignocellulose polymers. The minor weight loss of ∼1.5% was first observed around 160 °C in the RS, which may be due to the decomposition of volatile matters such as low molecular weight sugars. Following this, a substantial weight loss by ∼50% took place from 200 to 400 °C. This was expected to be the decomposition of hemicellulose and cellulose (Wattanasiriwech et al. 2010); the overlapping peaks in the differential thermogram (DTG) are, respectively, 322 and 366 °C. Another important loss of ∼14% due to the decomposition of lignin occurred between 350 and 540 °C. For IOC-RS, the result clearly shows that the thermal stability changed by modification of RS with iron. The DTG shows one peak at 354 °C, suggesting that it was likely that hemicellulose and cellulose simultaneously decompose within this temperature. The percentage weight loss of the residue after heating up to 700 °C was ∼4% higher for IOC-RS when compared to unmodified RS. This implies that there was presence of iron in IOC-RS which was not degraded when the samples were heated up to 700 °C. In addition, it was found that ash content in the pretreated RS was only 3.53% whereas that in the IOC-RS was increased to 22.84%; this higher percentage of residual ash in IOC-RS is assumed to be from iron(III)–lignocellulose complex. The form of iron in the IOC-RS was iron oxyhydroxide (–FeOOH), as previously reported (Pehlivan et al. 2013).

The pH_zpc (zero proton charge (zpc)) of an adsorbent is a very important characteristic that determines the pH at which the sorbent surface has net electrical neutrality, and at which value the acidic or basic functional groups no longer contribute to the pH of the solution (Wan Ngah & Hanafiah 2008). Figure 4 shows a plot of the zeta potential (ΔpH) of sorbent versus initial pH (pHi). The values of the zeta potential in all suspensions decreased as the pH was increased. According to the zeta potential curve of sorbent, it was found that the pH_zpc of IOC-RS was about 5.9. This result is in agreement with the pH_zpc of Fe(III)-coated rice husk (Pehlivan et al. 2013). Based on the pH_zpc value, it can be deduced that the IOC-RS surface charge is globally positive, as the solution pH is less than 6. On the contrary, with a solution pH higher than the pH_zpc, the surface is negatively charged and there will be an electrostatic repulsion between an anionic species and the surface of IOC-RS.

To determine the optimum pH for adsorption of As over RS and IOC-RS, the uptake of As(V) was studied at pH range of 2–10 and the removal data are shown in Figure 5. Results from the present study clearly show that As(V) adsorption by RS is not effective whatever the solution pH; while the percent removal of As(V) by IOC-RS was reduced from 99.6 to 64.3% as pH shifted from 2 to 10. Optimal As(V) adsorption by IOC-RS was found in the range of pH 2–4 in which the removal rate was above 95%. The pH affects significantly the speciation of As(V) in solution and the surface charge of the solid particles. As(V) species and their corresponding stability pH values are H₃AsO₄ (pH <2), H₂AsO₄⁻ (pH 2–7), HAsO₄²⁻ (pH 7–11), and AsO₄³⁻ (pH >12). As pH increased from 4 to 10, the amount of multivalent species (HAsO₄²⁻ (pH 5–9) and AsO₄³⁻ (pH >9)) increased and the two species were not preferably adsorbed by IOC-RS in comparison with H₂AsO₄⁻.

It was previously reported that the main factors governing the adsorption of As species are (i) the electrostatic interaction between iron oxyhydroxide sites of the adsorbent surface and (ii) the anionic As(V) species (Pehlivan et al. 2013). As the equilibrium pH increased from lower pH to pH_zpc, the decreased percentage removal of As(V) was attributed to the decreasing electrostatic attraction between the surface of iron oxyhydroxide loaded in IOC-RS and anionic multivalent HAsO₄²⁻ and AsO₄³⁻ species. Over the pH_zpc value, the surface sites of IOC-RS were negatively charged and inappropriate for adsorbing anionic arsenate species. The decrease in arsenate adsorption at pH > pH_zpc may have been due to the increase in the negative charge density at the surface of the adsorbent, and to the increase in the number of OH⁻ ions in the solution, in competition with the anionic arsenate species for adsorption.

To be able to estimate maximum capacities of adsorbents, it is necessary to know the quantity of adsorbed As(V) as a function of As(V) concentration in solution. The As(V) adsorption isotherms for the sorbent are shown in Figure 6. The sorption data obtained were analyzed by fitting the Freundlich and Langmuir isotherm models (Ranjan et al. 2009; Mishra & Ramaprabhu 2011; Maji et al. 2013). Table 1 gives a full overview of the Freundlich and Langmuir adsorption isotherm parameters.

A plot of C_e/q_e versus Ce is given in a straight line with its slope of 1/Q and intercept of 1/Qb and the results are shown in Table 1. According to the coefficients of correlation obtained (R² > 0.9961), the sorption of As(V) ions onto sorbent material, are fitted well to the Langmuir model. The maximum adsorption capacity of sorbent material for As(V) removal at pH 6 and 4 were found to be 19.96 mg/g and 21.739 mg/g, respectively (Table 1). Recently, several studies related to As(V) ions adsorption from water used for drinking and irrigation of crops have been carried out. Table 2 compares As(V) adsorption capacities of iron-based material adsorbents reported in previous studies. As seen, iron(III)-loaded resin has a high adsorption capacity for As(V) at pH < 2. In comparison to these approaches, the IOC-RS material (at pH 6) reported herein achieved better results than the amorphous iron oxide and ferric oxyhydroxides-modified sawdust, displaying at least five-fold better removal efficiency than the iron oxide nanoparticles and iron(III)-coated rice husk and a higher efficiency compared to the iron ores, iron oxide-coated sand (IOCS) and zeolite (ICZ). On the other hand, it was also found that the adsorption capacities of IOC-RS carried out at pH 4 and 6 are, respectively, comparable to those of iron-coated chitosan flakes (ICF) (pH 7) and iron(III)-loaded sponge (pH 9). The content of iron in the matrix is a crucial factor which impacts on the As adsorption capacity. In our case, there was a significant increase of ash content, which indicates high amounts of iron adsorbed on the surface of the RS substrate compared to the reported rice husk. This could relate to the chemical and physical properties of the feedstock. Rice husk is made of hard materials, including opaline silica and lignin, and has a more recalcitrant cell wall structure than RS, which contains a lesser amount of lignin (Boonmee 2012). In a future study, the correlation between the amount of iron-coated RS and the As adsorption capacity will be investigated. Moreover, it was found that IOC-RS could be regenerated by HCl and NaOH treatment. The highest recovery of >88% was achieved with 1 M NaOH.

The sorption of As(V) in the presence of common anions ions may be affected due to precipitation, complex formation or competition for sorption sites. Interference of anions on the sorption of As(V) onto sorbent was carried out with As(V) solution containing 50 mg/L of anion (i.e., NO₃⁻, SO₄²⁻, or PO₄³⁻) ions, at pH 4. It was observed that except for PO₄³⁻, other ions, e.g., NO₃⁻ and SO₄²⁻ have no significant interference with sorption of As(V) ions. The decrease of percentage removal of As(V) in the presence of PO₄³⁻ was 6%.

In this work, an inorganic–organic hybrid IOC-RS was used as adsorbent for the removal of As(V) from aqueous solution. The major conclusions based on the experimental study were as follows:

1. The pH, contact time, initial concentration, and dissolved Na₃PO₄ on the adsorption, significantly affect the As(V) adsorption capacity of IOC-RS.

2. The isotherm modeling revealed that the Langmuir equation could better describe the adsorption of As(V) on the IOC-RS as compared to Freundlich and D‒R models. The maximum adsorption capability of IOC-RS reached ∼22 mg As(V)/g of IOC-RS at pH 4.0 and 22 °C. In addition, the adsorption energy indicates that the adsorption process is dominated by the chemisorption.

3. The removal capacity Q of IOC-RS sorbent for As(V) ions was found to be higher than that of the majority of other iron oxide sorbents reported in the literature. Therefore, it can be stated that this sorbent has significant potential for the removal of As(V) ions from natural water.

This investigation was performed at the Guest Chair within the project ‘Exceed – Excellence Center for Development Cooperation – Sustainable Water Management in Developing Countries’ at the Technische Universitaet Braunschweig, Prof. Pehlivan being the visiting professor, and Ms Tran and Dr Ouédraogo being the international exchange staff members. The Exceed Project is granted by the German Federal Ministry for Economic Cooperation and Development (BMZ) and German Academic Exchange Service (DAAD), and their financial support we gratefully acknowledge.