Efficacy of novel Al–Zr impregnated cellulose adsorbent prepared using microwave irradiation for the facile defluoridation of water

Journal of Environmental Chemical Engineering 1 (2013) 1325–1335

Contents lists available at ScienceDirect

Journal of Environmental Chemical Engineering journal homepage: www.elsevier.com/locate/jece

Efficacy of novel Al–Zr impregnated cellulose adsorbent prepared using microwave irradiation for the facile defluoridation of water M. Barathi, A. Santhana Krishna Kumar, N. Rajesh * Department of Chemistry, Birla Institute of Technology and Science, Pilani-Hyderabad Campus, Jawahar Nagar, Shameerpet Mandal, R.R. Dist., 500078 AP, India

A R T I C L E I N F O

A B S T R A C T

Article history: Received 6 June 2013 Accepted 30 September 2013

The remediation of fluoride from water warrants effective adsorbents. The proposed method involves the impregnation of Al–Zr in cellulose matrix wherein fluoride ion from aqueous medium interacts with the cellulose hydroxyl groups as well as cationic Zr and Al hydroxides. The facile preparation of the adsorbent was accomplished by microwave-assisted synthesis. The adsorbent prior and subsequent to the adsorption of fluoride was characterized comprehensively using Fourier transform infrared spectroscopy (FT-IR), energy dispersive X-ray spectrometry (EDX) and X-ray diffraction (XRD) studies. Hydrogen bonding and electrostatic interactions support the adsorption mechanism. Various adsorption isotherm models, kinetics and thermodynamics were studied in detail. Pseudo second order kinetics supports the adsorption data. The adsorbent exhibits excellent adsorption up to 5 mg L1 fluoride and shows good potential toward practical application. ß 2013 Elsevier Ltd. All rights reserved.

Keywords: Cellulose Zr–Al cationic hydroxide Microwave assisted synthesis Fluoride Adsorption

Introduction The contamination of groundwater due to fluoride is an important problem to be addressed. The challenges are manifold and it is quite important to develop effective methodologies for fluoride remediation. The toxicity of fluoride is well known [1,2] and in India the affected states include some parts of Andhra Pradesh, Rajasthan and Tamil Nadu. Anthropogenic as well as natural sources are responsible for the ground water contamination with fluoride. According to WHO guidelines, the drinking water should contain fluoride within the limits 1.0–1.5 mg L1 [1,3]. In addition to fluorosis, high fluoride concentration can also cause other adverse effects including cancer, digestive and nervous disorders and respiratory problems [4,5]. Wastewater emanating from steel, semiconductors, electronic, toothpaste and insecticide manufacturing plants could also lead to contamination of ground water with fluoride ions [6]. There are few reports that state that excess fluoride might interfere with DNA synthesis [7] and also interfere with the carbohydrate metabolism [8]. The ingestion of excess fluoride can lead to the formation of hydrofluoric acid in the stomach leading to gastrointestinal irritation [8]. Considering the gravity of the problem, and to surmount some of the inadequacies in the existing methods it is imperative to develop effective adsorbents with removal efficiency to less than the permissible limit. A variety of adsorbents have been reviewed

* Corresponding author. Tel.: +91 40 66303503; fax: +91 40 66303998. E-mail address: [email protected] (N. Rajesh). 2213-3437/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jece.2013.09.026

for defluoridation [9]. Activated alumina is one of the versatile and well known adsorbents for removing fluoride but the pH of the treated water and the leaching of excess aluminum is one concern that needs to be addressed in this methodology. Activated alumina [10] and magnesia amended activated alumina [11] have been utilized for defluoridation with an adsorption capacity of 4.04 mg g1 and 10 mg g1, respectively. Alumina chitosan composite has been tested for defluoridation from villages with a defluoridation capacity of 3800 mg kg1 [12]. Nano hydroxyapatite/chitosan composite [13], magnesia/chitosan composite [14] are also known to exhibit good defluoridation capacity. Lanthanum impregnated carboxylated chitosan beads [15] have been studied for the removal of fluoride with an adsorption capacity of 4.7 mg g1. Similarly, zirconium tungstophosphate coated chitosan [16] has also proved to be equally effective for defluoridation. Alagamuthu and Rajan [17] have studied the removal efficiency from a fluoride endemic area using zirconium impregnated cashewnut shell carbon. Zr(IV) impregnated collagen fiber [18] is yet another novel adsorbent reported for defluoridation with an adsorption capacity of 2.29 mmol g1 at pH 5.5. Fe(III) coordinated amine modified PGMA grafted cellulose [19] and acidic alumina [20] have also proven to be effective for deflouridation. Despite the varied methods for the removal of fluoride, there is still a growing need and demand for the development of more effective and simpler adsorbents with good removal efficiency. Cellulose is a natural polysaccharide endowed with intramolecular hydrogen bonding and good stability [21]. The small size of fluoride, high electronegativity and its behavior as a hard base also makes it compatible with metal ions such as aluminum and

1326

M. Barathi et al. / Journal of Environmental Chemical Engineering 1 (2013) 1325–1335

zirconium. Taking advantage of this fact, cellulose biopolymer was explored as an effective matrix for impregnation of zirconium and aluminum through electrostatic interaction with the hydroxyl groups in cellulose. The hydrogen bonding interaction between the cellulose hydroxyl groups and fluoride also influences the adsorption of fluoride anion on the surface of the Al–Zr biopolymer. The objective of the present investigation was to develop a novel Al–Zr impregnated cellulose biopolymer adsorbent so as to bring down the concentration of fluoride to less than the stipulated toxic limit. The efforts that were channelized toward the development of this novel sorbent material by optimizing various analytical parameters are discussed at length in this work. Experimental Chemicals and reagents Analytical grade reagents were used for the fluoride adsorption studies. Milli pore water (Elix 3 Millipore unit) was used in preparation of the stock solution of fluoride. A 1 L volume of 1000 mg L1 stock solution of fluoride was prepared using sodium fluoride (Merck, India) and stored in a polypropylene bottle. Thin disposable nitrile gloves were used to handle fluoride solutions. A working solution of 5 mg L1 fluoride for the batch adsorption study was prepared by appropriate dilution. Cellulose was procured from Himedia, India and zirconium oxy chloride, activated neutral alumina as well as other common reagents used in the study was procured from s.d. fine chemicals, India. TISAB II buffer solution was obtained from Hanna instruments, USA. Preparation of Al–Zr impregnated cellulose adsorbent A known weight of cellulose biopolymer (2.0 g) was dispersed in minimum amount (1–1.5 mL) of DMF. About 2.0 g of activated alumina and 1.0 g of zirconium oxychloride were added to the dispersed cellulose and stirred for 5–10 min to get a homogeneous mixture. It was subjected to microwave irradiation at 160 W for 2 min with 30 s alternating time interval, and washed with Milli pore water. The Al–Zr impregnated biopolymer adsorbent was dried at 60 8C in a vaccum oven (Biotechnics, India) for 4 h and further characterized analytically thoroughly through various techniques. Measurement of fluoride concentration Ion selective meter (Model No.98185 Hanna Instruments, USA) equipped with a fluoride ion electrode was utilized for monitoring the levels of fluoride before and after adsorption. The instrument was calibrated with standard solutions of fluoride. The total ionic strength was adjusted using (1:100) TISAB II buffer solutions to get an optimum pH (5.0–5.5) for measurement of fluoride concentration. The fluoride ion selective electrode was calibrated using 1, 2, 10, 100 mg L1 of standard fluoride solutions to get an optimum slope value between 90 and 110%. The fluoride concentration was measured for the samples using the calibrated fluoride ion selective electrode. Instrumentation The Al–Zr impregnated cellulose biopolymer (Al–ZrIC) adsorbent was characterized using various physico-chemical characterization techniques. A SHARP model 23-GT microwave oven procured from Cole-Parmer with a maximum power range 1600 W was used in the preparation of the adsorbent. The FT-IR spectra of the adsorbent was recorded by mixing 0.01 g of the sorbent with spectroscopy grade KBr and the spectra were

recorded using a Jasco-4200 FT-IR spectrometer in the range 400–4000 cm1 at a resolution of 4 cm1. The X-ray diffraction (XRD) patterns of cellulose, Al–Zr impregnated cellulose sorbent before and after the adsorption of fluoride ion were recorded using a Philips PAN analytical X’pert PRO diffractometer with Cu Ka radiation (l = 0.154 nm) source operating at 40 kV and 30 mA with step size of 0.0178. The samples were scanned at the rate 2.08 min1 in the 2u range 78–608. The surface morphology of Al–Zr adsorbent before and after fluoride adsorption was obtained using a Hitachi S-520 scanning electron microscope. A JEOL JSM6390 model electron microscope was used to observe the morphology changes and record the energy dispersive X-ray spectrum in native cellulose. The adsorbent was coated with a thin layer of gold prior to recording the spectrum and the pH adjustments were done with HCl/NaOH using an LI-127 pH meter (Elico, India). Batch adsorption study The batch adsorption studies were conducted by equilibrating 0.3 g of the Al–ZrIC adsorbent material with 50 mL of 5 mg L1 fluoride ion solution at pH 5.5 in an orbital incubator shaker (Biotechnics, India) for varying time intervals and the concentration of fluoride in the solution phase was estimated by the ion selective electrode method. The amount of fluoride adsorbed (mg g1) at equilibrium (qe) was calculated using the expression qe ¼

C0  CeV W

(1)

where C0 and Ce are the initial and final liquid phase concentrations (mg L1) of fluoride, V is the volume of aqueous solution (L) and W is the weight of the Al–Zr impregnated cellulose adsorbent (g) used in the adsorption study. The percentage of fluoride adsorbed increased with time with good removal efficiency at 60 min thereby bringing down the concentration to less than 1.0 mg L1 level. Results and discussion FT-IR characterization of the adsorbent The FT-IR spectrum (Fig. 1) shows characteristic peaks pertaining to the different functional groups in cellulose [22,23] and Al–ZrIC adsorbent. A strong peak at 3328 cm1 is associated with O–H stretching and that at 2901 cm1 is due to C–H stretching in the biopolymer. A strong peak at 901 cm1 emanates from the C–O–C pyranose ring vibration [22]. The peak at 1643 cm1 could be ascribed to the H–OH hydrogen bonding. The peaks characteristic of Al–ZrIC were observed at 787 cm1, 841 cm1 (Al–O stretching vibration and Zr–O–C vibration) and 1701 cm 1 (formation of C5 5O by partial oxidation of cellulose) respectively [24–28]. The interaction of zirconium and aluminum with glycosidic linkage of the cellulose biopolymer is evident from the decrease in the peak intensity at 901 cm1. The O–H peak also shows some broadening after the adsorption of fluoride. The FT-IR spectrum (Fig. 1) shows some distinct peaks for the various functional groups in Al–ZrIC bioadsorbent before and after fluoride adsorption. Significant changes occur after the adsorption of fluoride ion with the appearance of deformation peaks at 584 cm1 and 567 cm1 attributed to the adsorption of fluoride on the adsorbent surface [29]. Furthermore, a new peak appears at 763 cm1 and the peak intensity increases in the range 500– 900 cm1. The hydrogen bonding interaction (–OH  F) [22,23] is evident from the peak at 763 cm1 and the metal-fluoride interactions are dominant in the region 500–900 cm1.

[(Fig._1)TD$IG]

M. Barathi et al. / Journal of Environmental Chemical Engineering 1 (2013) 1325–1335

1327

impregnation of metal ions onto the cellulose matrix (Fig. 3b), the above peaks are shifted corresponding to 2u values 37.738, 37.708, 42.818, 46.088, and 67.538. The average crystallite size was calculated from the full width at half-maximum of the peak using Debye Scherrer’s equation [31]

D¼

0:9l bcosu

where D is the average crystallite size, l is the characteristic wavelength of X-ray used (1.5406 A˚), u is the diffraction angle, and b is the angular width at intensity equal to half of the maximum peak intensity. The average crystallite size of cellulose biopolymer decreased from 8.048 to 1.06 mm due to strong interaction of the interaction of metal ions with the glycosidic linkage of cellulose biopolymer which was also confirmed by FTIR spectra. The new sharp peaks corresponding to 2u value 37.708 and 67.538 could be ascribed to aluminum oxide (JCPDS04-0877) and illustrates that the metal ions have been successfully impregnated onto cellulose biopolymer. Moreover, the peaks corresponding to 37.738 and 46.088 are assigned to the g alumina phase [32]. After the fluoride adsorption, XRD pattern of Al–ZrIC biopolymer (Fig. 3c) adsorbent shows peaks corresponding to 2u values 37.658, 39.638, 42.648, 45.858, and 67.158. Herein, the average crystallite size of the cellulose biopolymer increased from 1.06 to 8.38 mm and this could be attributed to the hydrogen bonding interaction between the cellulose hydroxyl groups and fluoride. The new sharp peaks corresponding to 2u value 37.658 and 39.618 indicates the presence of aluminum fluoride (JCPDS–pdf no.47-1659). The peaks at 42.648 and 45.858 is due to the formation of zirconium oxy fluoride (JCPDS–pdf no. 39-1216) and zirconium fluoride (JCPDS–pdf no. 39-1217) on Al–Zr impregnated cellulose biopolymer adsorbent. Mechanism of interaction of Al–Zr impregnated cellulose with fluoride

Fig. 1. (a) FT-IR spectrum of cellulose biopolymer (A) and Al–Zr impregnated cellulose adsorbent (B). (b) FT-IR spectrum of Al–Zr impregnated cellulose adsorbent (A) and the fluoride adsorbed (B).

Energy dispersive X-ray spectrum (EDX) and scanning electron microscopy analysis of Al–ZrIC adsorbent The SEM images (Fig. 2) clearly shows the structural and morphological changes in cellulose, Al–ZrIC adsorbent on and after fluoride adsorption. Cellulose is linear polymer of b-(1,4)-Dglucopyranose units and each glucose unit has three hydroxyl groups that could bond with Zr and Al metal ions [30]. The EDX spectrum (Fig. 2) confirms the presence of Al and Zr elemental peaks on the surface of the cellulose biopolymer and this indicates that Al and Zr were successfully loaded onto the cellulose matrix. The adsorption of fluoride on the surface of the Al–ZrIC adsorbent was ascertained from the EDX spectrum which shows the presence of fluorine along with the other major peaks such as C, O, Al and Zr respectively. Powder X-ray diffraction (XRD) analysis of the Al–ZrIC adsorbent As shown in Fig. 3a, the observed diffraction peaks for native cellulose [31] at 2u = 14.788, 16.428, 22.88 and 34.578 correspond to the planes 1¯ 0 1; 1 0 1; 0 0 2 and 2¯ 3 1 respectively. These planes are associated with the d-spacing values 6.04, 5.39, 3.88, and 2.65 A˚ respectively (JCPDS No. 03-0289). After the

Microwave assisted preparation of the adsorbent ensures efficient dielectric heating and under microwave (MW) irradiation there is rapid energy transfer from the cellulose hydroxyl groups to neighboring molecules [33]. Further MW radiation also results in lowering of Gibbs energy of activation [34] thereby promoting the effective interaction of the cationic Zr(OH)2+ and Al(OH)2+ hydroxides with the hydroxyl groups of cellulose. The mechanism of fluoride ion interaction with Al–Zr impregnated cellulose is shown in Fig. 4. The metal ion interacts with the glycosidic linkage of cellulose in the form of a strong electrostatic attraction. In aqueous solution, aluminum and zirconium ions could exist as cationic hydroxides such as Al(OH)2+, Zr(OH)22+, etc. and these species also interact with fluoride through electrostatic attractive forces. Furthermore, the hydrogen bonding interaction between the cellulose hydroxyl groups and fluoride would further reinforce the adsorption of fluoride anion on the surface of the Al–Zr biopolymer adsorbent. Optimization of pH The effect of pH is important in the adsorption of fluoride. Due to small size and high electronegativity, fluoride is solvated and the undissociated HF at very low pH can lead to reduction in the removal efficiency of fluoride [35]. The pH at the point of zero charge (pHPZC) of the adsorbent was determined by batch equilibration technique [36,37]. About 0.2 g of Al–Zr adsorbent added to 50 mL of 0.1 M KNO3 solution and the initial pH was adjusted from 3.0 to 9.0 by addition of 0.1 mol L1 HNO3 or NaOH. The experiments were carried out in an incubator shaker at 25 8C

[(Fig._2)TD$IG]

1328

M. Barathi et al. / Journal of Environmental Chemical Engineering 1 (2013) 1325–1335

Fig. 2. SEM images and EDX spectrum of cellulose biopolymer (A) Al–Zr impregnated cellulose biopolymer adsorbent (B) and fluoride ion adsorbed (C) on the adsorbent.

for 24 h. After a time period of 24 h, the final pH values of supernatant solutions were measured. The pHPZC of the Al–Zr impregnated cellulose biopolymer adsorbent was determined from the plot of DpH [pHinitial  pHfinal] versus pHinitial (Fig. 4c). The pHPZC was found to be 4.7. When pH < pHpzc the surface charge would be positive, and at pH > pHPZC, it would be negative. Therefore, the protonated adsorbent surface favors the adsorption of fluoride, whereas above pHpzc, more of surface sites are negatively charged and the fluoride would be adsorbed to a lesser extent due to the repulsive interaction between F ions and negative adsorbent surface. The effect of pH onto Al–Zr impregnated cellulose biopolymer adsorbent toward the fluoride ion adsorption was given in Fig. 4d. The adsorption of fluoride decreases to 46.37% at high pH values. In the pH range 2.5–3.5, Al and Zr cationic complexes interact with fluoride effectively. At pH < 4.0, the extent of fluoride adsorption increased but turbidity was observed in the aqueous phase attributed to the formation of HF and HF2 which interacts with

Al and Zr metal ions effectively to form soluble metal fluorides. It is well-known that Al(OH)3 dissolves to form soluble species of aluminum at low or high pH conditions [38]. It is possible that Al(OH)3 would yield soluble aluminates ([Al(OH)4]) at sufficiently alkaline pH. The formation of these soluble Al species affects the effective interaction of the cationic hydroxides of aluminum and zirconium with fluoride. However, in the pH range 4.5–5.5, there is no turbidity in the aqueous solution that emerges after adsorption and this shows that Al and Zr cationic complexes strongly bind with cellulose biopolymer and effectively interacts with fluoride. Indeed, in this pH range the concentration of fluoride remaining in the solution was found to be less than the permissible value of 1.0 mg L1. In weakly acidic medium, the fluoride ion interacts with the positively charged hydroxyl groups present on the metal ions and the cellulose biopolymer composite surface as evidenced by the FT-IR study. According to Pearson’s classification, fluoride is grouped as a hard base [39] while aluminum and zirconium are

[(Fig._3)TD$IG]

M. Barathi et al. / Journal of Environmental Chemical Engineering 1 (2013) 1325–1335

1329

Fig. 3. XRD pattern of cellulose biopolymer (A) Al–Zr Impregnated Cellulose biopolymer adsorbent (B) after fluoride ion adsorption on the adsorbent (C).

[(Fig._4)TD$IG]

Fig. 4. (a) Schematic diagram showing the interaction of metal ions with cellulose biopolymer surface. (b) Schematic representation of interaction of fluoride with Al–Zr impregnated cellulose biopolymer adsorbent surface. (c) Point of zero charge for the Al–Zr impregnated cellulose biopolymer adsorbent. (4) Effect of pH on the adsorption of fluoride.

1.1236 0.97 0.9479 0.1228 1.1316 0.84 0.351

r2 r2

3.153 0.032 0.95

x2

qm (mg g1)

0.9146 Elovich

0.0252 0.93

)

B

1.213

KT

1.859

0.1643 Temkin

)

b (L mg

1 1

5.7696

(3)

qo (mg g

1 1 þ bC 0

Langmuir

RL ¼

Table 1 Adsorption isotherm parameters.

where Ce is the equilibrium concentration of the fluoride ion in mg L1 qe is the amount of fluoride ion adsorbed at equilibrium in mg g1, qo is the maximum adsorption capacity in mg g1, and b is a constant (L mg1) related to the energy of adsorption. The maximum adsorption capacity, qo and the constant b are evaluated from the slope and intercept of the plot of Ce/qe against Ce (Fig. 5a). The regression coefficient obtained from this plot was found to be 0.93 and the respective isotherm parameters given in Table 1 indicate good affinity of the fluoride anion toward the Al–ZrIC adsorbent. The favorable nature of adsorption can be evaluated through a dimensionless parameter RL, given as

RL

r

2

(2)

0.4594

KF (mg

x

2

Freundlich

Langmuir adsorption isotherm The Langmuir isotherm model [40] is used to calculate the maximum adsorption capacity which is a measure of amount of the fluoride ion adsorbed per unit weight of the adsorbent. The main assumptions in this model include monolayer coverage and site equivalency with surface uniformity. The maximum adsorption capacity was obtained by fitting the experimental data to the linearized Langmuir expression given as

1–1/n

g

1

L

KE

1/n

Adsorption isotherm studies

0.1092

x2 x2

0.96 1.7135

)

n

r

2

The amount of adsorbent used in the batch study with 5 mg L1 fluoride was varied in the range 0.1–1.0 g. The removal of fluoride was effective in the range 0.3–0.5 g in 50 mL sample volume. The initial increase in adsorption is attributed to the strong electrostatic attraction between the fluoride anion and the Al–ZrIC biopolymer adsorbent. Beyond 0.3 g, there was no appreciable increase in the percentage adsorption, which indicates the saturation of the active adsorption sites.

Ce 1 Ce þ ¼ qe q0 b q0

g

B

0.4459 3.222

Redlich-Peterson

0.0371

x

2

Amount of adsorbent

qm (mg g1)

Dubinin–Radushkevich

Beyond pH 5.5, there is a decrease in the percentage adsorption of fluoride. This could be attributed to the deprotonation of the surface hydroxyl groups in the Al–ZrIC adsorbent and similarly at higher pH, the competition of the hydroxide anion with the fluoride anion for the active adsorption sites could also lead to a reduction in the percentage adsorption of fluoride. Three replicate analyses yielded a reduction of fluoride in the aqueous solution to 0.6  0.02 mg L1, which is appreciably less than the permissible limit.

B (mol2 kJ2)

A

[TD$INLE]-

1.0589

E (kJ mol1)

r2

r2

x2

classified as hard acids respectively. Hence, there could be an effective interaction between the positively charged metal ion and negatively charged fluoride anion as shown below.

1.0923

M. Barathi et al. / Journal of Environmental Chemical Engineering 1 (2013) 1325–1335

0.85

1330

[(Fig._5)TD$IG]

M. Barathi et al. / Journal of Environmental Chemical Engineering 1 (2013) 1325–1335

1331

Fig. 5. (a) Langmuir isotherm (b) Freundlich isotherm (c) D–R isotherm. (d) Temkin isotherm (e) Elovich isotherm (f) R–P isotherm (g) Plot of qe vs Ce.

Adsorption is irreversible if RL is zero, while the value higher than 1 signifies unfavorable adsorption. Constructive adsorption is reflected in the value of RL between 0 and 1 [41]. The value of RL for the adsorption of fluoride ion on the Al–ZrIC biopolymer adsorbent was found to be 0.4594 and this indicates the effectiveness of affinity between F ion and the Al–Zr impregnated cellulose adsorbent surface at the optimum experimental conditions.

Freundlich isotherm The study of adsorption from aqueous dilute solutions can also be examined through the linearized Freundlich isotherm [42] logqe ¼ logK F þ

1 logC e n

(4)

In this expression, the constants KF and n are the adsorption capacity and the intensity respectively. The KF and n values were

1332

M. Barathi et al. / Journal of Environmental Chemical Engineering 1 (2013) 1325–1335

obtained from the slope and intercept of the logarithmic plot of qe vs Ce (Fig. 5b) and the results presented in Table 1 are an indication of the applicability of this model also with a regression coefficient of 0.96. The Freundlich constant n lies between 1 and 10 (0.1 < 1/ n < 1.0) and this shows the favorable [43] adsorption of fluoride onto Al–ZrIC biopolymer adsorbent surface.

The above equation shows correspondence with the Langmuir model when g is unity. The corresponding R–P isotherm parameters g and B (Fig. 5f and Table 1) can be obtained from the linear expression.   Ce lnA (10)  1 ¼ glnC e þ lnB qe

Dubinin–Radushkevich (D–R) isotherm The Dubinin–Radushkevich isotherm (D–R) [44] has similarity to Langmuir isotherm and it gives the adsorption energy and the nature of the adsorption mechanism involved in the interaction between fluoride ion and the Al–Zr impregnated cellulose adsorbent surface. The D–R isotherm is expressed as

The exponent g was found to be 1.13 which illustrates the fact the adsorption of fluoride can well be explained through the Langmuir isotherm model as well.

lnqe ¼ lnqm  be2

(5)

where, qm is the maximum adsorption capacity and the parameter b is a constant related related as   1 e ¼ RTln 1 þ (6) Ce In the above equation e is called as Polanyi potential and the values of b and qm for this adsorption system are acquired from the slope and intercept of the plot of ln qe against e2 (Fig. 5c) and the respective isotherm parameters are shown in Table 1. The regression coefficient value was found to be 0.85. The adsorption energy, E can also be expressed as (2b)0.5and the negative value of E (1.059 kJ mol1) indicates that the interaction between the fluoride anion and metal ion impregnated biopolymer adsorbent is exothermic and hence adsorption is favored at low temperature. Temkin isotherm The Temkin and Pyzhev [45] model is represented by the equation qe ¼ B1 lnK T þ B1 lnC e

(7)

Values of B1 and KT were calculated from the plot of qe against ln Ce (Fig. 5d and Table 1) where RT/b = B1. In this isotherm the heat of adsorption of all the molecules in the layer decreases linearly with coverage [46] due to adsorbent – adsorbate interactions with uniform distribution of binding energies until it reaches a maximum value [45]. As shown in Table 1 for the Temkin isotherm, R2 > 0.94, which is close to the value obtained in Langmuir model and this indicates the vital interaction between fluoride ion and Al–Zr biopolymer adsorbent surface. The value of b (kJ mol1) was found to 2.076 which illustrates the electrostatic interaction between fluoride ion and Al–ZrIC biopolymer adsorbent [46]. Elovich isotherm The multilayer adsorption governed by the Elovich model is given by the expression [47] q q ln e ¼ lnK E qm  e Ce qm

Redlich and Peterson (R–P) isotherm The R–P isotherm yields an exponent g which could be correlated with the Langmuir model and in addition, other constants A and B are also obtained. The R–P isotherm (nonlinear) can be expressed as [48] AC e 1 þ BCeg

x2 ¼

X ðqe exp  qe cal Þ2 qe cal

(11)

where qe cal (mg g1) is the amount of fluoride adsorbed at equilibrium obtained by calculation from the different isotherm models and qe exp (mg g1) is the experimentally obtained value. The experimental equilibrium adsorption capacity (qe) was found at varying initial fluoride ion concentrations. The value of x2 would be lower when the data obtained from an isotherm model correlates with the experimental values [50]. The lower x2 values for Langmuir isotherm model (Table 1) indicates that the experimental data can be described through this model satisfactorily and the L shaped adsorption isotherm [51] data plots of qe against Ce (Fig. 5g) also show proximity to this model. The plot also indicates that the experimental equilibrium adsorption capacity data shows good fit to the Freundlich isotherm. Considerable deviations were reflected in the calculated equilibrium adsorption capacity values for Redlich and D–R isotherm model isotherms and hence are not very ideal to describe the adsorption of fluoride ion into Al–ZrIC biopolymer adsorbent surface. Adsorption kinetics The kinetics of adsorption of fluoride onto the metal ion impregnated cellulose adsorbent can be well described through pseudo first order [52] and the second order rate equations [53] logðqe  qt Þ ¼ logqe 

k1 t 2:303

(12)

The intraparticle diffusion [54] equation can be expressed as t q t ¼ þ qt k2 q2e qe

(13)

qt ¼ kint t 0:5

(14)

(8)

where KE and qm are characteristic equilibrium constant and the adsorption capacity. The Elovich isotherm parameters are obtained from the plot of ln (qe/Ce) against qe (Fig. 5e as given in Table 1).

qe ¼

Chi square test and equilibrium adsorption capacity comparison for the adsorption isotherms Although a higher regression coefficient indicates the applicability of a specific isotherm model, chi square test [49] was adopted in order to find the suitability of an isotherm that correlates the experimental data. The sum of the squares of the differences between the experimental and calculated adsorption capacity values divided by the calculated value is expressed as

(9)

where qe and qt refers to the amount of fluoride absorbed at equilibrium and time t. The slope and intercept of the kinetic plots (Fig. 6a and b) gives the corresponding parameters (Table 2). In addition, Weber–Morris [54] intraparticle diffusion is used to ascertain whether intraparticle diffusion is the rate-determining step. In this model, a plot of qt versus t0.5 would be linear if and if the plot passes through the origin then intraparticle diffusion is the only rate-limiting step. In the present adsorption process, the plot is linear, the slope gives the intraparticle rate constant kint, and the non-zero intercept (Fig. 6c) shows that diffusion is not the only phenomenon that controls the adsorption of the fluoride ion on the

[(Fig._6)TD$IG]

M. Barathi et al. / Journal of Environmental Chemical Engineering 1 (2013) 1325–1335

1333

Fig. 6. (A) Pseudo first order kinetic plot, (B) pseudo second order kinetic plot, (C) Plot of qt versus square root of time and (D) variation of ln K with temperature. Table 2 Kinetic parameters and intra-particle rate constant for fluoride ion adsorption. Concentration of F ion (mg L1)

qe (mg g1)

3.81 4.91 6.92

0.6182 0.7246 0.7895

Pseudo first order kinetic model 1

k1 (min 0.0371 0.0202 0.0217

)

1

q1 (mg g 0.0323 0.6000 0.0796

)

Pseudo second order kinetic model 2

1

R

k2 (g mg min

0.741 0.815 0.912

4.7894 3.0974 0.4728

Al–ZrIC biopolymer adsorbent [55]. The overall rate of adsorption of fluoride ion on the surface of the metal ion impregnated cellulose adsorbent could be influenced [56] by (a) external mass transfer when the fluoride ion is transported from the bulk solution to the external surface of the Al–Zr impregnated biopolymer adsorbent. (b) intraparticle or pore diffusion, in which the adsorbate molecules permeate the interior of the adsorbent particles and (c) adsorption at the interior sites of the Al–Zr impregnated cellulose adsorbent. Among these factors, the adsorption step happens relatively fast and hence it is assumed that it does not have considerable influence on the overall kinetics of adsorption. The overall adsorption rate could be controlled by surface or intraparticle diffusion. The kinetic and intra particle diffusion plots for varying fluoride concentrations are shown in Fig. 6, and Table 2 lists the rate constants for different initial fluoride concentrations using the pseudo-first-order, pseudosecond-order and intra-particle diffusion models. The value of correlation coefficient R2 for the pseudo-second-order adsorption model is very high (0.998) and therefore this model (Fig. 6b) is more suitable to describe the adsorption kinetics of fluoride onto Al–ZrIC biopolymer adsorbent.

)

q2 (mg g 0.6171 0.7088 0.999

1

)

Intraparticular diffusion model R

2

0.998 0.998 0.998

kint (g mg1 (min0.5)1

R2

0.0109 0.0086 0.0105

0.733 0.809 0.926

Thermodynamics of adsorption The novel adsorbent material was tested at varying temperatures to ascertain the spontaneity of the adsorption process and the following equations give the thermodynamic parameters. ln K ¼

DH0 DS0 þ RT R

DG0 ¼ RTlnK c 0

(15)

(16) 0

DH and DS are enthalpy and entropy changes while DG0 corresponds to the Gibb’s free energy change. The equilibrium constant K is obtained from the ratio of concentration of fluoride ion adsorbed on the Al–ZrIC biopolymer adsorbent to that in the solution. The values of DH0 and DS0 were calculated from the ln K against 1/T plot (Fig. 6d). The spontaneity of adsorption process is ascertained from the Gibb’s free energy values (Table 3) which are negative at the temperatures studied for 5 mg L1 of fluoride [57]. The activation energy of adsorption (Ea) at various temperatures

M. Barathi et al. / Journal of Environmental Chemical Engineering 1 (2013) 1325–1335

1334

Table 3 Thermodynamic parameters DH0, DS0 and DG0 for adsorption of fluoride ion onto Al–ZrIC adsorbent surface at different temperatures. Temperature (Kelvin)

DGo (kJ mol1)

DHo (kJ mol1)

DSo (J K1 mol1)

Ea (kJ mol1)

303 313 323 333

5.53 5.53 5.34 5.25

8.52

9.73

5.88

Table 4 Adsorption capacity comparison against other sorbents. Sl. No.

Adsorbent

Adsorption capacity (mg g1)

Reference

1. 2. 3. 4. 5. 6. 7. 8. 9.

Activated alumina Aluminum impregnated activated carbon Lanthanum impregnated silica gel Manganese-oxide-coated alumina Meso-structured zirconium phosphate Hydroxy apatite Lanthanum incorporated chitosan beads Unmodified cellulose Al–Zr impregnated cellulose biopolymer adsorbent

2.40 1.07 3.80 2.85 4.27 4.54 4.70 1.11 5.76

[29] [60] [61] [62] [63] [64] [65] Present work Present work

0 can be obtained from the relation [58] Ea ¼ DHads þ RT and for the present adsorption system, the average energy of activation was found to be 5.88 kJ mol1 in accordance with the exothermic physical adsorption process. This fact could also be corroborated from the negative value of calculated DH0 (8.52 kJ mol1) and the magnitude of DH0 would reflect the adsorption mechanism. If physical adsorption occurs, then DH0 is usually lower than 80 kJ mol1, while for chemical adsorption the value lies in the range 80–400 kJ mol1 [59]. The DGtotal involved in the overall adsorption process can be expressed through the Gibb’s equation relating enthalpy and entropy as

DGtotal ¼ DHtotal  T DStotal

(17)

The DStotal also depends on the entropy changes occurring due to the Al–Zr-cellulose as well as Al–Zr-cellulose  F interaction. This can be expressed as DStotal = f(DSAl–Zr-cellulose, DSAl–Zr-cellulose  F). Since, entropy is categorized as an extensive thermodynamic property; the overall entropy change is hence reflected in the negative value obtained from the summation of these entropy changes. The negative DS0 values also suggest decreased randomness at the Al–Zr cellulose adsorbent–solution interface. Laboratory scale-up by column study The applicability of the Al–Zr impregnated cellulose adsorbent material on a laboratory scale was examined for the defluoridation of water from a larger sample volume using a glass column (2.5 cm diameter, 30 cm length) by packing 2.0 g of the adsorbent. A known volume (300 mL) of tap water (original fluoride conc. 0.6 mg L1) spiked with 3 mg L1 fluoride ion was transferred to the column at a flow rate of 8 mL min1 and the concentration of fluoride was measured in the emerging solution. The concentration of fluoride was found to be than 0.72 mg L1 signifying a fluoride removal efficiency of greater than 80%. The column could be re-used by desorption with 5 mL of 1 N sodium hydroxide as sodium fluoride in the eluate. Comparison with other adsorbents The efficacy of this novel metal ion impregnated cellulose biopolymer adsorbent was compared with other adsorbents [29,60–65]. The adsorption capacity values in Table 4 show that the developed adsorbent is efficient for the removal of fluoride.

Conclusions This work has demonstrated the effective interaction between the Al–Zr impregnated cellulose biopolymer and fluoride ion. The novel Al–Zr impregnated cellulose adsorbent exhibits an adsorption capacity of 5.76 mg g1 and the experimental data showed a good fit to the Freundlich and Langmuir isotherm models. The adsorption of fluoride is favored by the interaction of cationic aluminum and zirconium hydroxides through electrostatic, hydrogen bonding and complexation mechanism. The spontaneity of adsorption and second order kinetic model describes the adsorption process. A sample volume of 300 mL on a laboratory scale column containing 3.6 mg L1 of fluoride could be brought down to less than the permissible limit of 1.0 mg L1. This is attainable at the natural pH prevailing in the water and hence the method could be tested very well for the field applications. Further studies are ongoing in our laboratory to accordingly modify the cellulose biopolymer by functionalizing and test the efficacy to a still higher sample volume. Acknowledgments We acknowledge the financial support from Department of Science and Technology (DST), New Delhi, India (Project No: DST/ TM/WTI/2K11/350(G). Thanks to Central Electrochemical Research Institute (CECRI) Karaikudi, India, SAIF, Cochin, India, and Indian Institute of Chemical Technology Hyderabad, India, for their valuable assistance in characterization of adsorbent. References [1] V.T. Yadugiri, Fluorosis: a persistent problem, Curr. Sci. 100 (2011) 1475–1477. [2] G.M. Whiteford, The physiological and toxicological characteristics of fluoride, J. Dent. Res. 69 (1990) 539–549. [3] C.K. Na, H.J. Park, Defluoridation from aqueous solution by lanthanum hydroxide, J. Hazard. Mater. 183 (2010) 512–520. [4] J. Fawell, K. Bailey, E. Chilton, E. Dahi, L. Fewtrell, Y. Magara, Fluoride in Drinking Water, World Health Organization, IWA Publishing, UK, 2006. [5] R.C. Meenakshi, Maheshwari, Fluoride in drinking water and its removal, J. Hazard. Mater. B137 (2006) 456–463. [6] S. Jagtap, M.K. Yenkie, N. Labhsetwar, S. Rayalu, Fluoride in drinking water and defluoridation of water, Chem. Rev. 112 (2012) 2454–2466. [7] Y. Zhou, C. Yu, Y. Shan, Adsorption of fluoride from aqueous solution on La3+ impregnated crosslinked gelatin, Sep. Purif. Technol. 36 (2004) 89–94. [8] M. Islam, R.K. Patel, Thermal activation of basic oxygen furnace slag and evaluation of its fluoride removal efficiency, Chem. Eng. J. 169 (2011) 68–77. [9] P. Loganathan, S. Vigneswaran, J. Kandasamy, R. Naidu, Defluoridation of drinking water using adsorption process, J. Hazard. Mater. 248–249 (2013) 1–19.

M. Barathi et al. / Journal of Environmental Chemical Engineering 1 (2013) 1325–1335 [10] S. Ghorai, K.K. Pant, Investigations on the column performance of fluoride adsorption by activated alumina in a fixed-bed, Chem. Eng. J. 98 (2004) 165–173. [11] S.M. Maliyekkal, S. Shukla, L. Philip, I.M. Nambi, Enhanced fluoride removal from drinking water by magnesia-amended activated alumina granules, Chem. Eng. J. 140 (2008) 183–192. [12] N. Viswanathan, S. Meenakshi, Enriched fluoride sorption using alumina/chitosan composite, J. Hazard. Mater. 178 (2010) 226–232. [13] C.S. Sundaram, N. Viswanathan, S. Meenakshi, Uptake of fluoride by nano-hydroxyapatite/chitosan, a bioinorganic composite, Bioresour. Technol. 99 (2008) 8226–8230. [14] C.S. Sundaram, N. Viswanathan, S. Meenakshi, Defluoridation of water using magnesia/chitosan composite, J. Hazard. Mater. 163 (2009) 618–624. [15] N. Viswanathan, S. Meenakshi, Enhanced fluoride sorption using La(III) incorporated carboxylated chitosan beads, J. Colloid Interface Sci. 322 (2008) 375–383. [16] N. Viswanathan, S. Meenakshi, Development of chitosan supported zirconium (IV) tungstophosphate composite for fluoride removal, J. Hazard. Mater. 176 (2010) 459–465. [17] G. Alagamuthu, M. Rajan, Equilibrium and kinetics of adsorption of fluoride onto zirconium impregnated cashew nut shell carbon, Chem. Eng. J. 158 (2010) 451– 457. [18] X.P. Liao, B.I. Shi, Adsorption of fluoride on zirconium(IV)-impregnated collagen fiber, Environ. Sci. Technol. 39 (2005) 4628–4632. [19] T.S. Anirudhan, P.S. Suchitra, Synthesis and characterization of iron(III)-coordinated amine-modified poly(glycidylmethacrylate)-grafted densified cellulose and its applicability in defluoridation from industry effluents, Ind. Eng. Chem. Res. 49 (2010) 12254–12262. [20] A. Goswami, M.K. Purkait, The defluoridation of water by acidic alumina, Chem. Eng. Res. Des. 90 (2012) 2316–2324. [21] D. Klemm, B. Heublein, H.P. Fink, A. Bohn, Cellulose: fascinating biopolymer and sustainable raw material, Angew. Chem. Int. Ed. 44 (2005) 3358–3393. [22] T.S. Anirudhan, T.A. Rauf, S.R. Rejeena, Removal and recovery of phosphate ions from aqueous solutions by amine functionalized epichlorohydrin-grafted cellulose, Desalination 285 (2012) 277–284. [23] Y. Tian, M. Wu, R. Liu, D. Wang, X. Lin, W. Liu, L. Ma, Y. Lie, Y. Huang, Modified native cellulose fibers – a novel efficient adsorbent for both fluoride and arsenic, J. Hazard. Mater. 185 (2011) 93–100. [24] A.R. Choudhuri, C.G. Takoudis, Investigation of the aluminium oxide/Si(1 0 0) interface formed by chemical vapor deposition, Thin Solid Films 446 (2004) 155– 159. [25] S. Jagtapa, M.K.N. Yenkie, N. Labhsetwar, S. Rayalu, Defluoridation of drinking water using chitosan based mesoporous alumina, Microporous Mesoporous Mater. 142 (2011) 454–463. [26] S.C. Sundaram, N. Viswanathan, S. Meenakshi, Defluoridation chemistry of synthetic hydroxyapatite at nano scale: equilibrium and kinetic studies, J. Hazard. Mater. 155 (2008) 206–215. [27] J. Lojewska, P. Miskowiec, T. Lojewski, L.M. Proniewicz, Cellulose oxidative and hydrolytic degradation: in situ FTIR approach, Polym. Degrad. Stab. 88 (2005) 512–520. [28] R. Dongre, D.N. Ghugal, J.S. Meshram, D.S. Ramteke, Fluoride removal from water by zirconium (IV) doped chitosan bio-composite, Afr. J. Environ. Sci. Technol. 6 (2012) 130–141. [29] S. Ghorai, K.K. Pant, Equilibrium, kinetics and breakthrough studies for adsorption of fluoride on activated alumina, Sep. Purif. Technol. 42 (2005) 265–271. [30] X. Guo, F. Chen, Removal of arsenic by bead cellulose loaded with iron oxyhydroxide from groundwater, Environ. Sci. Technol. 39 (2005) 6808–6818. [31] A. Santhana Krishna Kumar, S. Kalidhasan, V. Rajesh, N. Rajesh, Adsorptive demercuration by virtue of an appealing interaction involving biopolymer cellulose and mercaptobenzothiazole, Ind. Eng. Chem. Res. 52 (34) (2013) 11838– 11849. [32] C. Ma, Y. Chang, W. Ye, W. Shang, C. Wang, Supercritical preparation of hexagonal g-alumina nanosheets and its electrocatalytic properties, J. Colloid. Interface Sci. 317 (2008) 148–154. [33] C.O. Kappe, Controlled microwave heating in modern organic synthesis, Angew. Chem. Int. Ed. 43 (2004) 6250–6284. [34] S.A. Galema, Microwave chemistry, Chem. Soc. Rev. 26 (1997) 233–238. [35] A.M. Raichur, M.J. Basu, Adsorption of fluoride onto mixed rare earth oxides, Sep. Purif. Technol. 24 (2001) 121–127. [36] N.A. Oladoja, I.A. Ololade, O.A. Alimi, T.A. Akinnifesi, G.A. Olaremu, Iron incorporated rice husk silica as a sorbent for hexavalent chromium attenuation in aqueous system, Chem. Eng. Res. Des. (2013), http://dx.doi.org/10.1016/ j.cherd.2013.03.001.

1335

[37] N.A. Oladoja, Y.D. Aliu, Snail shell as coagulant aid in the alum precipitation of malachite green from aqua system, J. Hazard. Mater. 164 (2009) 1496–1502. [38] R. Liu, W. Gong, H. Lan, T. Yang, H. Liu, J. Qu, Simultaneous removal of arsenate and fluoride by iron and aluminium binary oxide: comparative adsorption effects, Sep. Purif. Technol. 92 (2012) 100–105. [39] R.G. Pearson, Hard and soft acids and bases, J. Am. Chem. Soc. 85 (1963) 3533–3539. [40] I. Langmuir, The adsorption of gases on plane surface of glass, mica and platinum, J. Am. Chem. Soc. 40 (1918) 1361–1403. [41] S.C. Sairam, N. Viswanathan, S. Meenakshi, Defluoridation chemistry of synthetic hydroxyapatite at nano scale: equilibrium and kinetic studies, J. Hazard. Mater. 155 (2008) 206–215. [42] H.M.F. Freundlich, Over the adsorption in solution, Z. Phys. Chem. 57 (1906) 385– 470. [43] T. Ru-Ling, W. Feng-Chin, Inferring the favorable adsorption level and the concurrent multi-stage process with the Freundlich constant, J. Hazard. Mater. 155 (2008) 277–287. [44] K.Y. Foo, B.H. Hameed, Insights into the modeling of adsorption isotherm systems, Chem. Eng. J. 156 (2010) 2–10. [45] M.J. Tempkin, V. Pyzhev, Kinetics of ammonia synthesis on promoted iron catalysis, Acta Physiochim. URSS 12 (1940) 327–356. [46] W. Nigussie, F. Zewgeb, B.S. Chandravanshib, Removal of excess fluoride from water using waste residue from alum manufacturing process, J. Hazard. Mater. 147 (2007) 954–963. [47] M.C. Ncibi, Applicability of some statistical tools to predict optimum adsorption isotherm after linear and non-linear regression analysis, J. Hazard. Mater. 153 (2008) 207–212. [48] O. Redlich, D.L. Peterson, A useful adsorption isotherm, J. Phys. Chem. 63 (1959) 1024–1026. [49] J. Kang, B. Li, J. Song, D. Li, J. Yang, W. Zhan, D. Liu, Defluoridation of water using calcined magnesia/pullulan composite, Chem. Eng. J. 166 (2011) 765–771. [50] M. Auta, B.H. Hameed, Coalesced chitosan activated carbon composite for batch and fixed-bed adsorption of cationic and anionic dyes, Colloids Surf. B Biointerfaces 105 (2013) 199–206. [51] C.H. Giles, R.B. Mckay, Adsorption of cationic basic dyes by fixed yeast cells, J. Bacteriol. 89 (1965) 390–397. [52] S. Lagergren, Zur theorie der sogennanten adsorption geloster stoffe, K. Sven. Vetenskapsakad. Handl. 24 (1898) 1–39. [53] L.M. Camacho, A. Torres, D. Saha, S. Deng, Adsorption equilibrium and kinetics of fluoride on sol–gel-derived activated alumina adsorbents, J. Colloid Interface Sci. 349 (2010) 307–331. [54] W.J. Weber, J.C. Morris, Kinetics of adsorption on carbon from solution, J. Sanit. Eng. Div. Am. Soc. Civ. Eng. 89 (1963) 3–60. [55] Z. Yu, T. Qi, J. Qu, L. Wang, J. Chu, Removal of Ca(II) and Mg(II) from potassium chromate solution on Amberlite IRC 748 synthetic resin by ion exchange, J. Hazard. Mater. 167 (2009) 406–412. [56] A. Mittal, L. Kurup, J. Mittal, Freundlich and Langmuir adsorption isotherms and kinetics for the removal of tartrazine from aqueous solutions using hen feathers, J. Hazard. Mater. 146 (2007) 243–248. [57] X. Tang, Z. Li, Y. Chen, Adsorption behavior of Zn(II) on calcinated chinese loess, J. Hazard. Mater. 161 (2009) 824–834. [58] A.S. Krishna Kumar, S. Kalidhasan, V. Rajesh, N. Rajesh, Application of celluloseclay composite biosorbent toward the effective adsorption and removal of chromium from industrial wastewater, Ind. Eng. Chem. Res. 51 (2012) 58–69. [59] I.H. Gubbuk, Isotherms and thermodynamics for the sorption of heavy metal ions onto functionalized sporopollenin, J. Hazard. Mater. 186 (2011) 416–422. [60] R.L. Ramos, J. Ovalle-Turrubiartes, M.A. Sanchez-Castillo, Adsorption of fluoride from aqueous solution on aluminium-impregnated carbon, Carbon 37 (1999) 609–617. [61] S.A. Wasay, M.J. Haron, S. Tokunaga, Adsorption of fluoride, phosphate, and arsenate ions on lanthanum-impregnated silica gel, Water Environ. Res. 68 (1996) 295–300. [62] S.M. Maliyekkal, A.K. Sharma, L. Philip, Manganese-oxide-coated alumina: a promising sorbent for defluoridation of water, Water Res. 40 (2006) 3497–3506. [63] S.K. Swain, T. Patnaik, V.K. Singh, U. Jha, R.K. Patel, R.K. Dey, Kinetics, equilibrium and thermodynamic aspects of removal of fluoride from drinking water using meso structured zirconium phosphate, Chem. Eng. J. 171 (2011) 1218–1226. [64] X. Fan, D.J. Parker, M.D. Smith, Adsorption kinetics of fluoride on low cost materials, Water Res. 37 (2003) 4929–4937. [65] A. Bansiwal, D. Thakre, N. Labhshetwar, S. Meshram, S. Rayalua, Fluoride removal using lanthanum incorporated chitosan beads, Colloids Surf. B Biointerfaces 74 (2009) 216–224.