Removal of As(III) from water using modified jute fibres as a hybrid adsorbent

Abstract

Many studies have concentrated on the removal of arsenic from water using granular mineral fine particles. However, very little research has focused on the preparation of materials which aim to be applied to the situations of arsenic pollution emergencies in rivers or lakes. In this study, jute fibres were modified by loading iron oxyhydroxide (which was demonstrated to be mainly α-FeOOH) to produce an effective hybrid adsorbent (Fe-JF) with the advantages of an excellent arsenic removal effect and easy retrieval from rivers or lakes. The jute fibres were firstly esterified with succinic anhydride to graft with carboxyl groups in order to enhance the loading amount of iron(III), the maximum iron(III) loading on Fe-JF reached 208.2 ± 0.2 mg g⁻¹ while the density of grafted carboxyl groups was 2.78 mmol g⁻¹. The maximum adsorption capacity for As(III) reached 12.66 mg g⁻¹ while the density of carboxyl groups was 2.21 mmol L⁻¹. Meanwhile, the iron leaching amount was 0.178 mg L⁻¹ which could meet the requirement of the standard limit of iron in drinking water (China, 0.3 mg L⁻¹). Influential factors, such as pH, contact time and coexisting anions are investigated in this study. The column experiments showed that the breakthrough point declined from 2300 BV (bed volume) to 1200 BV when EBCT (empty-bed contact time) decreased from 3.5 min to 1.8 min. The Adams–Bohart model was adopted to describe the continuous flow system.

---

1. Introduction

In China, with the rapid development of the economy and urbanization process, sudden water pollution accidents have occurred frequently in recent years. These malignant water pollution accidents have resulted in severe environmental problems.¹ Among these accidents, arsenic contamination has attracted much attention because it occurs most frequently. For instance, the groundwater of Huaihua city of Hunan province suffered arsenic contamination in 2008; the Dasha river of Henan province was severely polluted by the illegal arsenic emissions of a chemical factory; the Yangzonghai lake of Yunnan province suffered arsenic contamination because of As-containing wastewater without effective treatment; one particularly serious event was arsenic contamination in the Picang river of Shandong province in 2009, where malignant pollution emergencies relating to arsenic happened twice in a short span of three months.

Adsorption was considered as a promising method to apply to arsenic polluted water remediation among the conventional methods of chemical precipitation, flocculation, membrane separation, ion exchange and so on. The key aspect of adsorption is the appropriate adsorbent, thus seeking an effective and economic adsorbent has been pursued by many researchers. Activated carbon is the most commonly used adsorbent throughout the world. However, it is still facing some problems related to high cost, non-selectivity and its complicated regeneration requirements.^2,3

Although many mineral adsorbents such as goethite, ferrihydrite, akaganéite and lepidocrocite have been reported as having high affinities toward arsenic, they cannot be applied in arsenic contamination emergencies because of their tiny size and they are not easy to retrieve from water. Therefore, many supports including sands, activated carbon, ion exchange resins and waste natural materials have been selected to be loaded with tiny sized active components.^4,5 In recent years, there have been many studies focused on bio-composite-based materials,^6–8 which were considered to be a promising direction in the preparation of adsorbents applied to heavy metal removal from water. It was reported that a novel hybrid material was produced by combining γ-Fe₂O₃ and biochar, which exhibited excellent ferromagnetic properties, and the highest adsorption capacity for arsenate reached 3.15 mg g⁻¹.⁹ Hristovski et al.¹⁰ investigated the effect of the synthesis conditions on iron (hydr)oxide impregnated activated carbon, the iron content of the hybrid adsorbent ranged from 0.5 to 16% Fe per g of dry media.

As an adsorbent, jute would be a promising material because it is a commonly available economic crop which could be obtained in large quantities. It was reported that Shukla and Pai¹¹ used a modified jute fibre to remove Cu(II), Ni(II) and Zn(II) from aqueous solutions and achieved satisfactory results. The jute fibre can be easily manipulated and is applicable to various situations, especially for rapid environmental remediation after arsenic pollution accidents. The exhausted adsorbents could be burned, producing little arsenic-containing ash, which can be further treated through immobilization technology. In addition, the fibrous adsorbent showed lower water resistance than granular adsorbents and could be woven into various shapes.¹² This indicated that the fibrous adsorbents have the most potential market prospects.

Many studies have demonstrated that iron, including zero-valent iron, iron oxyhydroxides and iron minerals have high affinity toward arsenic.^13–15 In our study, the raw jute was firstly grafted with carboxylic groups, and then was complexed with ferric ions. NaOH solution was added dropwise to form iron oxyhydroxide, which was the major reactive center responsible for arsenic removal.¹⁶ The aim of introducing carboxylic groups was to increase the iron(III) amount complexed on the jute fibres. This study focused on As(III) removal from water because As(III) is more toxic and occurs more frequently in environmental emergencies than As(V).

2. Materials and methods

2.1. Preparation of Fe-JF

2.1.1. Preparation of carboxylated jute fibres. A total of 10 g pretreated dry jute fibres was reacted with succinic anhydride under pyridine reflux (200 mL) at 90 °C for 12 h in order to obtain the carboxylate-functionalized jute fibres. The fibres were subsequently washed with acetone, anhydrous ethanol, and deionized water to remove the residual organic solvent on the fibres.

2.1.2. Impregnation of iron. The introduction of iron(III) was accomplished by immersing the carboxylate-functionalized jute fibres in an aqueous solution of 0.05 M Fe(NO₃)₃·9H₂O (pH was adjusted to be 7.0 ± 0.2) for 12 h. After filtration the fibres were filled into a glass column of 25 × 500 mm, and then 0.5 mol L⁻¹ NaOH was slowly dropped at a rate of 100 ± 1 mL h⁻¹ through the glass column for 2 h. The fibres were then aged at room temperature for another 12 h. Deionized water was constantly dropped into the column until the color of the effluent changed from yellow to colorless. The product was finally dried at 50 °C and stored for further use. The synthesis route is shown in Fig. 1.

Fig. 1 Synthesis route used to obtain Fe-JF.

2.2. Arsenic adsorption equilibrium tests

A stock solution of As(III) at a concentration of 1000 mg L⁻¹ was prepared using sodium As(III) (NaAsO₂; AR). A series of 250 mL glasses of the solution were added to 100 mL of 1, 5, 10, 25, 50, 80, 100, 120, 150 mg L⁻¹ arsenic solutions and 0.5 g Fe-JF. The samples were shaken at 150 rpm at room temperature for 24 h with pH adjusted to 7.0 ± 0.1 with dilute HNO₃ or NaOH solutions.

2.3. Arsenic adsorption kinetics

The kinetics experiments were conducted in a round glass reactor that was placed in a thermostat water bath, where the temperature could be adjusted. A pH electrode combined with a thermometer was inserted below the surface of the solution to maintain constant pH conditions in the reactor.

2.4. Effect of pH and competing anions on arsenic removal

The pH experiment was conducted with the initial As(III_{C}) concentration of 10 mg L⁻¹. The pH level was adjusted in the range of 3–10. The samples were adjusted several times during the initial 6 h to maintain the stability of the desired pH. For the study of competing anions, the initial As(III) concentration was 0.05 mg L⁻¹, the co-existing anions were prepared including Cl⁻, SO₄²⁻, NO₃⁻, F⁻, PO₄³⁻, SiO₃²⁻, their respective concentrations are listed in Table 1.

Table 1 The list of the coexisting anions

Co-existing anions	Conc./mg L^−1
SO4^2−	50
NO3^−	2
SiO3^2−	20
Cl^−	35
F^−	10

---

2.5. Fixed bed column experiments

A glass column of 22 × 200 mm was employed with a bed depth of 50 mm. A nylon mesh (1000 μm opening size) was placed at the bottom of the column to prevent Fe-JF from being discharged into the tubes. The feed solution was prepared from 0.5 mg L⁻¹ arsenic. The aqueous solution was pumped through the packed column with a peristaltic pump (BT-200B, Shanghai Qingpu analytical instrument Co., China).

2.6. Analytical methods

The density of carboxyl groups on the jute fibres was determined by a back-titration method.¹⁷ 0.1 g carboxylated jute fibres was suspended in 100 mL of 0.01 mol L⁻¹ NaOH solution for 1 h under constant stirring. After filtration, the excess NaOH was back-titrated with 0.01 mol L⁻¹ HCl solution. Methyl orange was used as an indicator and the titration was terminated when the solution color turned from yellow to orange. The density of carboxyl groups was calculated by:

SEM images were obtained using a Quanta 200 FEG (USA), which was coupled with an EDS system. The samples were pretreated with gold spraying in order to enhance the conductivity and make it easier to observe non-conductive samples. Total dissolved As and Fe levels were determined using inductively coupled plasma emission spectrometry (ICP-OES) (PerkinElmer, Optima 2000, UK). For the experiments with initial As(III) concentration below 0.05 mg L⁻¹, the arsenic analytical method was hydride generation atomic fluorescence spectroscopy (HG-AFS) (AFS-930, Beijing Jitian Co., Ltd, China), which is capable of detecting arsenic at concentrations as low as 1.0 μg L⁻¹. All of the samples were reduced by 5% (w) thiourea–5% (w) ascorbic acid to ensure that all arsenic species were converted to detectable As(III). Iron(III) loaded on Fe-JF was extracted by HNO₃ (98%, v/v).

3. Results and discussion

3.1. Characterization of Fe-JF

The synthesized adsorbent has a fibrous shape and excellent mechanical strength. The highest content of Fe loaded on the fibers was 208 ± 2 mg of Fe per g after three cycles of the loading process.

The photographs of the raw and modified jute fibres are shown in Fig. 2a and b. The BET surface area of the raw jute is only 0.57 m² g⁻¹. As shown in Fig. 2(c), the jute surface is smooth and it is composed of very few pores. The elemental composition of the raw jute (as shown in the square of Site 1 (S1) in Fig. 2(c)) is listed in Table 2. The modified jute fibres develop a deep brown color due to the dispersion of iron oxyhydroxide. Fig. 2(d) shows that the modified jute surface is heterogeneous. Site 2 (S2) and Site 3 (S3) are selected for EDS analysis (the conducting medium of Au is deducted). The main metallic element contained in the square areas (as in Fig. 2d and f) is iron, with the percentage content varying from 25.78–34.02%. The EDS spectra of Fe has two peaks: high energy (K line: 6.3996 keV) and low energy (L line: 0.7048 keV).

Fig. 2 (a) Photograph of the raw jute fibres. (b) Photograph of the modified jute fibres. (c) SEM image of the raw jute fibres (magnitude ×5000). (d) SEM image of the modified jute fibres (magnitude ×5000). (e) EDS spectrum of the modified jute fibres. (f) SEM image of the modified jute fibres (magnitude ×10 000).

Table 2 Elemental analysis of raw jute and Fe-JF (% based on dry weight)^a

Element	Content (wt%)
	Raw jute	Fe-JF
	S1	S2	S3
a — means not detected.
C	54.62	45.20	42.90
O	42.71	29.02	23.08
Fe	—	25.78	34.02

---

The FTIR spectra of the raw jute fibres and Fe-JF are shown in Fig. 3, where a wide band is seen at 3400 cm⁻¹ due to the –OH stretching vibration, and the characteristic absorption peaks of the cellulose backbone appear at 1034 cm⁻¹ and 1059 cm⁻¹. The raw jute was pretreated with 5% (w/v) NaOH solution in order to remove the adherent impurities. Compared with the raw jute fibres, the FTIR spectrum of the carboxylated jute showed two typical bands at 1721 cm⁻¹ and 1194 cm⁻¹, which are attributed to the stretching C=O vibration and stretching C–O of COOH, respectively.¹⁸ This indicated that carboxyl groups were successfully grafted onto the cellulose of the jute fibres.

Fig. 3 FT-IR spectra of the raw jute fibres and carboxylated jute fibres.

To examine the iron mineral phase loaded on the jute fibres, the samples were detected by X-ray diffraction (XRD) as shown in Fig. 4. The peaks at 2θ = 17.8°, 21.2°, 26.7°, 33.2°, 34.7°, 36.6°, 40°, 41.2°, 47.3°, 50.6°, 53.2°, 59°, 61.4° are assigned to α-FeOOH and 2θ = 26.7°, 39.2°, 61.1° are assigned to β-FeOOH.¹⁹ Fig. S1† gives the standard XRD patterns of the four referenced kinds of iron (oxyhydr)oxide and the main characteristic reflections are summarized in Table 3. It can be considered that the iron oxyhydroxide loaded on the jute fibres is mainly α-FeOOH and β-FeOOH. It is seen that the crystallinity of the products is poor, resulting from various kinds of iron oxyhydroxide particles mixed together, which are too small to diffract.

Fig. 4 X-ray diffraction pattern of the iron oxyhydroxide loaded jute fibres.

Table 3 The XRD diffraction angles and corresponding crystal planes of five kinds of iron (oxyhydr)oxide

Iron (oxyhydr)oxides	Position [° 2θ]									Reference
	Crystal plane
α-FeOOH	21.2	26.3	33.2	34.7	36.6	40	41.2	53.2	59.0	JCPDS 29-0713
	(100)	(120)	(130)	(021)	(111)	(121)	(140)	(221)	(151)
β-FeOOH	11.8	16.8	26.7	35.2	39.2	46.4	52.0	55.9		JCPDS 34-1266
	(110)	(200)	(310)	(211)	(301)	(411)	(600)	(521)
γ-FeOOH	14.1	27.1	36.3	38.1	46.9	52.8				JCPDS 8-98
	(020)	(120)	(310)	(111)	(051)/(200)	(151)
δ-FeOOH	35.2	39.9	54.4	63.2						JCPDS 13-87
	(100)	(101)	(102)	(110)
α-Fe2O3	24.1	33.1	35.8	40.9	49.5	54	62.4	63.9	72	JCPDS 33-0664
	(012)	(104)	(110)	(113)	(024)	(116)	(214)	(300)	(1010)

---

3.2. Relationship of the density of carboxyl groups to the iron(III) amount loaded on Fe-JF

The density of carboxyl groups was determined by the back-titration method, where samples with different densities of carboxyl groups were obtained by changing the esterification reaction time to 4, 8, 12 h during the synthesis procedure. As shown in Table 4, when the density of carboxyl groups increased from 1.19 to 2.78 mmol g⁻¹, the Fe content also increased from 124 ± 2 to 208 ± 2 mg of Fe per g. However, the iron leached to the bulk solution increased simultaneously from 0.06 to 0.52 mg of Fe per g. The raw jute fibres could also adsorb iron(III) as high as 102 ± 2 mg g⁻¹. If the iron leaching amount reached 10.12 mg g⁻¹, the excess leaching of iron into water may cause secondary pollution. The results indicate that the carboxylated jute fibres could greatly reduce iron leaching from the hybrid material.

Table 4 The iron(III) loading amount and leakage associated with different carboxyl group densities

---

In order to uncover the reason for Fe leaching, iron oxyhydroxide loaded D113 (styrene-type weak acid cation exchange resin) containing a high density of fixed –COOH groups (3.26 mmol g⁻¹) was prepared following identical procedures. The results showed that the Fe content loaded on the resin reached 290 ± 3 mg of Fe per g, but the iron leached to the bulk solution was only 0.012 mg g⁻¹, which was much lower than that of the jute fibres. Therefore, we deduced that the comparatively higher leaching of iron was attributed to the swelling of cellulose due to the long time soaking in water.

3.3. Adsorption equilibria

Adsorption equilibrium tests were carried out to investigate the adsorption capacities at various initial concentrations at pH 7.0 ± 0.1. Fig. 5 shows that the carboxylated jute surface enhanced the ability to uptake ferric ions, thus greatly increasing the content of loaded iron oxyhydroxide. The iron loading amount of the raw jute was 102 ± 2 mg g⁻¹, which was significantly increased to 124 ± 2 mg g⁻¹ with a density of grafted carboxyl groups of 1.19 mmol g⁻¹. When the density of grafted carboxyl groups was enhanced from 1.19 to 2.78 mmol g⁻¹, the amount of loaded iron was accordingly increased to 208 ± 2 mg of Fe per g. The adsorption capacity increased with an increase of carboxyl group density because more carboxyl groups means more anchor locations available for combining with ferric ions, thus resulting in a higher amount of loaded iron oxyhydroxide.

Fig. 5 Adsorption isotherms for As(III) at different densities of carboxyl groups (Fe-JF dosage 0.5 g, pH 7.0 ± 0.2, temperature 25 °C).

The adsorption equilibrium data were fitted with Langmuir and Freundlich isotherm models. The Langmuir isotherm equation is expressed as below:²⁰

where q_e is the quantity of the species adsorbed at equilibrium (mg g⁻¹), K_L is a constant representing the virtual bonding strength between the target species and adsorbent, C_e is the equilibrium concentration of adsorbate in the solution, and q_max is the maximum adsorption capacity.

The Freundlich isotherm equation is expressed as follows:²¹

ln q_e = ln K_F + 1/n ln C_e (3)

where q_e is the quantity of the species adsorbed at equilibrium (mg g⁻¹), K_F is a constant which is a measure of sorption capacity, 1/n is a measure of adsorption density, and C_e is the equilibrium concentration of adsorbate in the solution.

The isotherm fitting parameters are listed in Table 5, the maximum adsorption capacity was 12.66 mg g⁻¹. Higher values of the constant k_L and n indicated the easier uptake of As(III). The results appeared to be similar to arsenic adsorption behavior of iron oxyhydroxide fine powders.¹⁶ Compared with iron oxide-coated sand (q_max = 0.04 mg g⁻¹),²² Fe-JF showed a much higher adsorption capacity, which was attributed to the high amount of loaded iron and the appropriate iron mineral phase of α-FeOOH. Comparison of the Langmuir capacities of more adsorbents for arsenic adsorption are listed in Table 6.

Table 5 Langmuir and Freundlich adsorption isotherm parameters for As(III)

As species	Langmuir			Freundlich			Fe content (mg g^−1)	Max As/Fe (mol/mol)
	kL	qmax (mg g^−1)	R^2	KF	n	R^2
As(III)	0.167	11.49	0.984	1.721	2.203	0.985	124 ± 2	0.069
	0.171	12.66	0.958	1.950	2.132	0.989	181 ± 2	0.052
	0.210	13.69	0.980	2.499	2.409	0.974	208 ± 2	0.049

---

Table 6 Comparison of the Langmuir capacities of different adsorbents for arsenic adsorption

Adsorbents	Capacity (mg g^−1)		Reference
	As(III)	As(V)
ZVI nanoparticles	3.5	—	23
Fe2O3 nanomaterial	1.25	—	24
Activated alumina	3.45	9.9	25
Iron hydroxide-coated alumina	7.65	15.9	26
Fe3O4 coated wheat straw	3.9	8.1	24
Natural laterite	0.17	—	27
Iron oxide-coated sand	0.029	0.14	22
Iron chitosan composite flake	16.15	—	28
Magnetite-reduced graphene oxide	10.20	5.27	29
Iron oxide@carbon	29.4		30
Iron modified activated carbon	—	6.57	31
Iron modified polymer	15.45	7	32
Perlite/γ-Fe2O3 composite	—	4.64	33
Biochar/γ-Fe2O3 composite	—	3.147	34
Modified siderite	9.429	—	35
Iron oxide coated cement (∼0.212 mm)	0.69	—	36
FeCl3 treated tea fungal biomass	5.4	10.26	37
ZVI nanoparticles modified starch	12.2	14	38
Iron modified jute fibres	12.66	—	This work

---

3.4. Adsorption kinetics

The time dependence of the As(III) adsorption process is shown in Fig. 6. The initial uptake rate was rapid, followed by a slower reaction rate that gradually approached equilibrium. About 70% As(III) was removed during the first 60 min. The slower adsorption occurred due to the decrease in the driving concentration difference between the bulk solution and Fe-JF surface. The adsorption achieved equilibrium within 200 min when 80% As(III) was removed.

Fig. 6 Kinetic adsorption data and pseudo-second order model fitting (initial As(III) concentration: 10 mg L⁻¹, Fe-JF dosage 1 g, Fe content of Fe-JF 175 ± 2 mg of Fe per g, pH 7.0 ± 0.2, temperature 25 °C).

Several kinetic models such as pseudo-first-order, pseudo-second-order and the Elovich equation were used to investigate the adsorption mechanism.³⁹

The pseudo-first-order kinetic equation is expressed as follows:³⁹

log(q_e − q_t) = (log q_e) − k₁t (4)

where t is the time, q_t is adsorption capacity at t and k₁ (min⁻¹) is the rate constant.

The pseudo-second-order equation can be expressed as:⁴⁰

where t is the time, q_t is adsorption capacity at t, q_e is the adsorption equilibrium capacity and k₂ (g mg⁻¹ min⁻¹) is a constant which strongly depends on the applied operating conditions such as the initial solute concentration, pH, temperature and so on.

The intraparticle diffusion equation is expressed as follows:⁴⁰

q_t = k_dt^1/2 + C (6)

where t is the time, q_t is adsorption capacity at t and k_d (mg L⁻¹ min^−1/2) is the rate constant. A plot of q_t versus ln(t) should yield a straight line if the kinetic curves follow the Elovich equation.

Among the three kinetic models, the pseudo-second-order described the experimental data best, where the determinant coefficients (R²) were all above 0.99. The arsenic removal process is presumed to involve three stages: (1) mass transfer across the external boundary layer; (2) diffusion within the pores of the adsorbent; (3) adsorption at a special site on the surface. Good conformation to the intraparticle diffusion equation suggests that the rate-determining step is the diffusion within the pores of the adsorbents. Many papers in the literature have reported that porous materials or iron loaded porous materials usually fit well to the intraparticle diffusion equation.⁴¹ In this study, diffusion is not likely to be rate-limiting due to the smooth surface of the jute fibres and the agitation conditions. The results showed a good fit to the pseudo-second-order model with the correlation coefficient above 0.999, indicating that As(III) adsorption onto Fe-JF was dominated by the chemical process.

3.5. Effect of pH

The uptake of As(III) as a function of pH ranging from 3 to 10 is shown in Fig. 7. The percentage removal did not change much in the range of pH 3–8, but was obviously reduced from ca. 85% to ca. 55% when the pH shifted from 8 to 10. The optimal adsorption condition was found between pH 3–8, which is in accordance with the pH of natural water and it is not necessary to pre-adjust the pH for As(III) contaminated natural water when Fe-JF is applied to As(III) remediation.

Fig. 7 Effect of pH on As(III) adsorption on Fe-JF (initial As(III) concentration: 10 mg L⁻¹, Fe-JF dosage 0.5 g, Fe content of Fe-JF 183 ± 2 mg of Fe per g, temperature 25 °C).

The aqueous pH is one of the most important factors that influence the adsorption efficiency. As(III) appears to be different species in the wide pH range of 3–10. From eqn (7)–(9), As(III) exists mostly as neutral H₃AsO₃ when the pH is lower than 9, while mono-valent H₂AsO₃⁻ is the stable form between pH 9–12.^3,42

H₃AsO₃ ↔ H₂AsO₃⁻ + H⁺, pK_a1 = 9.2 (7)

H₂AsO₄⁻ ↔ HAsO₄²⁻ + H⁺, pK_a2 = 12.1 (8)

HAsO₄²⁻ ↔ AsO₄³⁻ + H⁺, pK_a3 = 12.7 (9)

Cumbal and SenGupta⁴³ reported that the distribution of three surface functional groups of iron oxyhydroxide (e.g., FeOOH₂⁺ (pH < 6.5), FeOOH (6.5 < pH < 9), and FeOO⁻ (pH > 9)) changed with respect to pH. Electrostatic repulsion is not dominant since neutral H₃AsO₃ is stable at pH 3–8, which was consistent with the slight change of percentage removal at pH 3–8. The adsorption of As(III) onto the iron oxyhydroxide of the Fe-JF surface is mainly by ligand exchange (eqn (10) and (11)), while electrostatic attraction is insignificant. The remarkable decline between pH 8–10 was attributed to the electrostatic repulsion between FeOO⁻ and H₂AsO₃⁻ (eqn (12) and (13)).

FeOOH₂⁺ + H₃AsO₃ ↔ FeOH₂AsO₃ + H₂O + H⁺ (10)

FeOOH + H₃AsO₃ ↔ FeOH₂AsO₃ + H₂O (11)

FeOO⁻ + H₂AsO₃⁻ ↔ FeOHAsO₃⁻ + OH⁻ (12)

FeOO⁻ + HAsO₃²⁻ ↔ FeOHAsO₃²⁻ + OH⁻ (13)

3.6. Effect of co-existing anions

In natural water sources such as rivers and lakes, many anions might exist together. The effects of those co-existing anions, including Cl⁻, SO₄²⁻, NO₃⁻, F⁻ and SiO₃²⁻, on As(III) adsorption by Fe-JF have been studied, and their effects are shown in Fig. 8. When the dosage of Fe-JF was 2 g L⁻¹, the residual arsenic concentration decreased from 0.008 mg L⁻¹ to 0.006 mg L⁻¹ in the presence of these co-existing anions, the results showed that the presence of Cl⁻, SO₄²⁻, NO₃⁻, F⁻, SiO₃²⁻ has a slightly adverse effect on As(III) removal. Among all of the co-existing anions, sulfate is considered to be in greater competition due to the enhanced electrostatic interactions, but it was still uncertain whether an outer or inner sphere complex forms. Phosphate was often reported to have a profound competing impact on arsenic adsorption onto iron hydroxide, it is not difficult to understand that phosphorus and arsenic are in the same main group, PO₄³⁻ and AsO₄³⁻ have identical chemical structures, and both molecules are tetrahedral oxyanions with similar pK_a values. Experiments were conducted with 0.1 mg L⁻¹ phosphate in order to investigate the effect of competition from phosphates. As can be seen from Fig. 8, phosphates caused an essential drop in As(III) removal efficiency, the maximum removal efficiency decreased from 98% to 90%. The interference effects of phosphates may be attributed to the competition for active sites.

Fig. 8 Effect of competing anions on As(III) adsorption on Fe-JF (initial As(III) concentration: 0.05 mg L⁻¹, Fe content of Fe-JF 172 ± 2 mg of Fe per g, pH 7.0 ± 0.2, temperature 25 °C).

3.7. Fixed-bed column runs

The column adsorption experiments were conducted at a neutral pH level. As shown in Fig. 9, at an EBCT of 3.5 min the SLV (superficial liquid velocity) was 0.86 m h⁻¹ and the effluent concentration of As(III) was well below the breakthrough point of 0.01 mg L⁻¹ with the bed volume less than 2300 BV (bed volume), above which the effluent As(III) concentration began to increase significantly. When the effluent concentration reached 90% of the initial concentration, the adsorption column was considered to be exhausted. EBCT is an important parameter and a bigger EBCT results in a slower feed flow rate. Generally, the bed volumes of the breakthrough point are increased with a decrease in EBCT, this is also proved in our study. When the EBCT decreased from 3.5 min to 1.8 min, the breakthrough point declined from 2300 BV to 1200 BV. The large decrease of ca. 50% indicates that the uptake efficiency depends strongly on the contact time. These results are in accordance with the kinetics results, that Fe-JF needs a comparatively longer time to reach equilibrium. Many papers in the literature reported that a lower kinetic performance was more sensitive to EBCT.⁴⁴

Fig. 9 Schematic diagram of the dynamic column tests and As(III) breakthrough curves at different EBCT (initial As(III) concentration 0.5 mg L⁻¹, pH 7.0 ± 0.2, Fe content of Fe-JF 181 ± 2 mg of Fe per g, temperature 25 ± 2 °C).

The detailed numerical results are given in Table 7. The breakthrough capacities (until a breakthrough point of 0.01 mg L⁻¹ was observed) for As(III) were 9.63 and 5.02 mg g⁻¹ at an EBCT of 3.5 and 1.8 min, respectively, values which are much lower than the saturated adsorption capacity of 12.66 mg g⁻¹ because the terminal mass transfer unit was not fully adsorbed. Moreover, the compacting effect in the column and channel flow may lead to insufficient contact between Fe-JF and water. The breakthrough capacity decreased significantly from 9.63 to 5.02 mg L⁻¹, indicating that the flow rate had a significant influence on the removal efficiency. The higher flow rate lowered the full contact time between Fe-JF and the As(III) species, thus resulting in an increase of effluent arsenic concentration.

Table 7 Column test results for As(III) uptake under two different EBCT

EBCT (min)	Feed solution	Bed depth (cm)	SLV (m h^−1)	Breakthrough point (BV)	Breakthrough capacity (mg g^−1)	Adams–Bohart model
	C0 of As(III) (mg L^−1)					Ka (L mg^−1 h^−1)	N0 (mg L^−1)
3.5	0.5	5	0.86	2300	9.63	0.69	100.05
1.8	0.5	5	1.67	1200	5.02	1.82	159.21

---

Several models were widely used to simulate the column test data, i.e. Thomas, Yoon–Nelson, Wolborska and Adams–Bohart models. The Adams–Bohart model is used in this study to describe the initial part of the breakthrough curve. The model is described by the following equation:⁴⁵

where, the parameter N₀ is the saturation concentration of the column bed (mg L⁻¹), K_a is the adsorption rate constant (L mg⁻¹ h⁻¹), x is the column bed depth (m) and U₀ represents the empty bed flow linear velocity (m h⁻¹).

The parameters N₀ and K_a can be determined from a linear plot of ln(C_t/C₀) versus time at a given flow rate and bed depth. A linear relationship between ln(C_t/C₀) versus t is obtained for the initial part (before the breakthrough point) as shown in Fig. S2.† The respective values of N₀ and K_a are presented in Table 7.

4. Conclusion

The present work indicates that the modified jute fibres (Fe-JF) possess a higher efficiency for adsorbing As(III) than the raw jute fibres. The loading amount of Fe is increased with an increase in grafted carboxyl group density. XRD patterns indicate that the loaded iron oxyhydroxide was mainly α-FeOOH. The adsorption is governed by a Langmuir model giving an extremely good fit. The optimal adsorption performance is achieved at around neutral pH, and there is no obvious adverse effect of common co-existing anions on As(III) adsorption. Flow rate has a significant influence on the removal rate. The saturation concentration simulated by the Adams–Bohart model is found to be 159.21 mg L⁻¹ at an EBCT of 3.5 min. From a practical viewpoint, this novel fibrous material has advantages for practical applications because of its quite effective cost and ease of retrieval from water.

Acknowledgements

This work was supported by the Funds for Creative Research Groups of China (Grant no. 51121062) and the National Water Pollution Control and Management Technology Major Projects (2012ZX07205-005).

References

1. Q. Hu, D.-W. Gao, H.-Y. Pan, L.-L. Hao and P. Wang, RSC Adv., 2014, 4, 40071–40077.
2. A. Demirbas, J. Hazard. Mater., 2009, 167, 1–9.
3. D. Mohan and C. U. Pittman Jr, J. Hazard. Mater., 2006, 137, 762–811.
4. S. N. Kartal and Y. Imamura, Bioresour. Technol., 2005, 96, 389–392.
5. H. Xu, B. Allard and A. Grimvall, Water, Air, Soil Pollut., 1991, 57, 269–278.
6. A. Özer, J. Hazard. Mater., 2007, 141, 753–761.
7. A. Özer and H. Pirincci, J. Hazard. Mater., 2006, 137, 849–855.
8. W. Wan Ngah and M. Hanafiah, Bioresour. Technol., 2008, 99, 3935–3948.
9. M. Zhang, B. Gao, S. Varnoosfaderani, A. Hebard, Y. Yao and M. Inyang, Bioresour. Technol., 2013, 130, 457–462.
10. K. D. Hristovski, P. K. Westerhoff, T. Möllerc and P. Sylvester, Chem. Eng. J., 2009, 146, 237–243.
11. S. Shukla and R. S. Pai, Bioresour. Technol., 2005, 96, 1430–1438.
12. M. R. Awual, A. Jyo, T. Ihara, N. Seko, M. Tamada and K. T. Lim, Water Res., 2011b, 45, 4592–4600.
13. J. M. Calo, L. Madhavan, J. Kirchner and E. J. Bain, Chem. Eng. J., 2012, 189, 237–243.
14. S. Lin, D. Lu and Z. Liu, Chem. Eng. J., 2012, 211, 46–52.
15. Y. Glocheux, M. M. Pasarín, A. B. Albadarin, S. J. Allen and G. M. Walker, Chem. Eng. J., 2013, 228, 565–574.
16. M. Faria, R. S. Rosemberg, C. A. Bomfeti, D. S. Monteiro, F. Barbosa, L. C. Oliveira, M. Rodriguez, M. C. Pereira and J. L. Rodrigues, Chem. Eng. J., 2014, 237, 47–54.
17. H. P. Boehm, E. Diehl, W. Heck and R. Sappok, Angew. Chem., Int. Ed. Engl., 1964, 3, 669–677.
18. A. Keranen, T. Leiviska, B.-Y. Gao, O. Hormi and J. Tanskanen, Chem. Eng. Sci., 2013, 98, 59–68.
19. X. Wang, X. Chen, L. Gao, H. Zheng, M. Ji, C. Tang, T. Shen and Z. Zhang, J. Mater. Chem., 2004, 14, 905–907.
20. V. Vinodhini and N. Das, Am.-Eurasian J. Sci. Res., 2009, 4, 324–329.
21. Z. Akmar Zakaria, M. Suratman, N. Mohammed and W. Azlina Ahmad, Desalination, 2009, 244, 109–121.
22. O. Thirunavukkarasu, T. Viraraghavan and K. Subramanian, Water, Air, Soil Pollut., 2003, 142, 95–111.
23. C. Daikopoulos, Y. Georgiou, A. B. Bourlinos and M. Baikousi, Chem. Eng. J., 2014, 256, 347–355.
24. Y. Tian, M. Wu, X.-B. Lin and P. Huang, J. Hazard. Mater., 2011, 193, 10–16.
25. T.-F. Lin and J.-K. Wu, Water Res., 2001, 35, 2049–2057.
26. J. Hlavay and K. Polyak, J. Colloid Interface Sci., 2005, 284, 71–77.
27. A. Maiti, S. DasGupta, J. K. Basu and S. De, Sep. Purif. Technol., 2007, 55, 350–359.
28. S. R. Kanel, B. Manning, L. Charlet and H. Choi, Environ. Sci. Technol., 2005, 39, 1291–1298.
29. V. Chandra, J. Park, Y. Chun and J.-W. Lee, ACS Nano, 2010, 4, 3979–3986.
30. K. Gupta, A. Maity and U. C. Ghosh, J. Hazard. Mater., 2010, 184, 832–842.
31. J. Fang, Z. M. Gu and D. C. Gang, Environ. Sci. Technol., 2007, 41, 4748–4753.
32. D. Ociński, I. Jacukowici-Sobala, J. Raczyk and E. Kociolek-Balawejder, React. Funct. Polym., 2014, 83, 24–32.
33. D. N. Thanh, M. Singh, P. Ulbrich and N. Strnadova, Sep. Purif. Technol., 2011, 82, 93–101.
34. M. Zhang, B. Gao, S. Varnoosfaderani and A. Hebard, Bioresour. Technol., 2013, 130, 457–462.
35. K. Zhao, H. M. Guo and X. Q. Zhou, Appl. Geochem., 2014, 48, 184–192.
36. S. Kundu and A. K. Gupta, J. Hazard. Mater., 2007, 142, 97–104.
37. G. S. Murugesan, M. Sathishkumar and K. Swaminathan, Bioresour. Technol., 2006, 97, 483–487.
38. M. Mosaferi, S. Nemati, A. Khataee and S. Nasseri, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 2014, 12, 69–74.
39. F.-C. Wu, R.-L. Tseng and R.-S. Juang, Water Res., 2001a, 35, 613–618.
40. X. Xu, B.-Y. Gao, Q.-Y. Yue, Q.-Q. Zhong and X. Zhan, Carbohydr. Polym., 2010, 82, 1212–1218.
41. W. Chen, R. Parette, J. Zou, F. S. Cannon and B. A. Dempsey, Water Res., 2007, 41, 1851–1858.
42. K. Wu, R. Liu, T. Li, H. Liu, J. Peng and J. Qu, Chem. Eng. J., 2013, 226, 393–401.
43. L. Cumbal and A. K. SenGupta, Environ. Sci. Technol., 2005, 39, 6508–6515.
44. M. R. Awual, A. Jyo, S. A. El-Safty, M. Tamada and N. Seko, J. Hazard. Mater., 2011, 188, 164–171.
45. B. Preetha and T. Viruthagiri, Sep. Purif. Technol., 2007, 57, 126–133.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra11901k

---